Seeing that R&D costs are the lion's share, I wonder if we are at a point where the focus can shift to improving the cost of inference.
Unless we are genuinely pushing to find AGI, at which point nothing matters, LLMs in their current form don't replace knowledge workers but are an effective force multiplier. How good is enough?
For instance, I pay about $1-2 a month for DeepSeek. It's not as sophisticated as Claude, but it still doubles my productivity as a SWE.
If Fable comes out and demands 50x the price of DeepSeek in order for Anthropic to make a profit on it, how much more productive would I be compared to my personal experience + DeepSeek? 3x? 50x?
Is it cost effective for a business to hire someone without SWE experience + Fable verses hiring someone with SWE experience and DeepSeek? When does R&D hit diminishing returns?
Even if you discount superhuman AI (which I would emphasize that frontier researchers do not discount and expect to see soon) think it’s still hard to have enough confidence that the ground is solid. Someone in 2024 trying to go down this route would have invested a lot of now-pointless effort into prompt engineering.
For clarity, inference is typically a COGS and therefore hits Gross Margin vs model training which would typically be in OpEx (where R&D lives) and would hit operating margin.
> I wonder if we are at a point where the focus can shift to improving the cost of inference.
There's always working on improving the cost of inference, but I don't think this is an area of R&D that will slow down. The reason is:
1. A better competitor model risks eating away at how much they can charge for inference (i.e. revenue)
2. Whoever unlocks AGI will unlock even more growth
3. Even when you unlock AGI, you'll want to throw gobs of money at it to improve itself and all sorts of things.
> If Fable comes out and demands 50x the price of DeepSeek in order for Anthropic to make a profit on it, how much more productive would I be compared to my personal experience + DeepSeek? 3x? 50x?
You're pricing it wrong and looking at it wrong. First, the per token price doesn't consider that a smarter model can end up using fewer tokens overall to achieve a result. Secondly, if the difference is between failing to accomplish the task and accomplishing the task, suddenly that 50x can seem like a bargain.
> Is it cost effective for a business to hire someone without SWE experience + Fable verses hiring someone with SWE experience and DeepSeek? When does R&D hit diminishing returns?
At this time, someone without SWE experience + <name AI model> vs someone good with SWE experience and <name another AI model> is a no-brainer. The AI model is an accelerant but the "no SWE experience" will be accelerated into a wall. Now maybe that doesn't matter for prototyping and certain other things, but anything in production the lack of experience will hurt them with things they won't even know about or even know how to look for it (e.g. slow, insecure, etc).
Let's put it this way, how much is 5% productivity bump worth to you?
If you're in the US and you're making 100k a year, that's worth 5k or $416/m. So you can buy two of the most expensive plans on the frontier models.
This focus on cost optimization is insane. Just use the frontier models. Even a marginal bump is worth whatever the hell they're charging, at least for now.
You really think there's zero correlation between productivity and wages? Sure, it's noisy and you might stay at $100k or even get fired. But I'd say the expected wage value of 5% higher productivity on a large sample is at least 3.5% or so.
Everyone else will be 5% more productive. Then no one is "more" productive. So everyone has a higher output, but the same wages and hours worked. There was only a gap when usable AI first came out, some contractors could do the same quantity of work in less time and enjoy time off or do more jobs. Now the gap has closed or is closing. And using AI now is more about not being less productive than peers who do use it.
That's not how productivity works. It's not a zero-sum game.
If all construction workers can build houses 5% more efficiently, that's not the same as nothing changing. Depending on supply and demand, it means 5% more houses are built, or houses are 5% cheaper, or maybe 5% bigger, or some combination. Whether or not the construction workers all get a raise or 5% get fired (or both) depends on that supply and demand, but historically they often get a piece of the growing economic pie.
Why would the company pay more when they can just not pay more? The only things I can see happening is they might lower prices as competition ramps up, or in general as there is more supply for the same cost.
If there's sufficient demand, that's just what happens.
To try and explain one path: Company A doesn't raise wages but makes 5% more money. Company B pivots from Industry B into construction (because suddenly construction is having 5% fatter margins), and hires workers at more competitive wages to poach them from Company A. Company A forces to raise wages.
If there's a demand ceiling on housing it's a different story though.
More like company B purchases a construction company and changes nothing but number go up for shareholders while wages stay stagnant for people producing actual value, as they have for decades.
Or, with current US policy, the top 10% will get 6% richer and you'll get 1% poorer. Sure, the pie can grow, but you won't (unless you are in the top 10%) get any of it.
large companies aren’t buying subscription plans. my org has a 2k per month token budget per person and starting to explore optimizations like automatic model routing.
There is no evidence that these tools provide a 5% bump, if anything they are providing a 20% liability (pulling random numbers is fun).
Also where is the evidence that the workers have ever benefited from productivity bumps? The only thing that happens is surplus gets captured by the owners while workers are forced to do more.
You're saying this like I would see that 5k in my bank account. If I'm 5% more productive that probably wouldn't even make it into annual review, let alone pay.
> Unless we are genuinely pushing to find AGI, at which point nothing matters, LLMs in their current form don't replace knowledge workers but are an effective force multiplier. How good is enough?
There's a non-negligible percentage of the industry who have a pseudo-religious belief in AGI, so I wouldn't be surprised if that was, in fact, the goal.
Who knows, maybe they'll stop once the money dries up.
If a model comes and makes developer + Deepseek even a little more productive, from employers perspective, it'd still make sense to pay a lot of money for that.
Deepseek shines for personal usage because it's possible to use it however you want and whenever you want with no session/weekly limits stress because you use the API and it's priced very reasonably.
> Unless we are genuinely pushing to find AGI, at which point nothing matters
I think the third coming out Jesus Christ in closer than AGI. Seriously, I dread how much of Silicon Valley is wrapped in this narrative of AGI and Singularity.
How can all these "rationalists" fail to see that this is what religion looks like: Faith and promises of heaven and hell.
These "rationalists" understand that beliefs should be evaluated on whether they match what we observe, rather than preconceptions about which buckets certain ideas fall into. If the Southern Baptist Convention had announced a theological breakthrough in 2022 that lets them map out the precise calendar dates of the events in Revelations, and using this map they made a series of specific predictions that ended up coming true, it would be rational to start studying End Times Christianity in more detail and irrational to say that it's religious so you're not going to worry too much about it.
That's interesting on Deepseek. But I think as long as the models are still making noticeable gains with each iteration it's hard to say "good enough."
Similarly, some bosses might believe that they can hire 100 cheap, unmotivated SWEs to replace Linus Torvalds or Fabrice Bellard and achieve something slightly worse. But in certain areas, it doesn't work like that.
on the other hand, the sad reality is many swe are working on dumb crud apps and the code-quality is already very low ime, and the jury is still out if ai tools can long term replace those; but what i have seen first-hand isnt super promising do far...
I'm a little confused here. Cost of revenue is lower than revenue. That's good. R&D is the main contributor to losses here and this seems normal in an industry like this. For OpenAI specifically, I think this is problematic. They were the first movers but despite the large R&D they've lost so much ground to Anthropic despite Anthropic seemingly gifting them with weird PR self owns. But if we were to extrapolate this to the industry as a whole, this seems more positive than negative. Am I reading this incorrectly? Unless there's an assumption that R&D costs have to forever go up in order to increase revenue, I feel like this shows that the AI industry is actually on a path to profitability in the long term.
Whether it can physically be as all encompassing as it makes itself out to be or whether it will just be healthily profitable remains to be seen. Kind of like how Uber went from "We'll autonomously drive the world" to "Look, we deliver food, goods, and people to locations and we figured out how to do that in a way that makes profits. Also, ads".
How in the world could you read that article and think there is anything positive about OpenAI's prospects? We've been hearing for months that these companies need to make trillions of dollars in a handful of years, growing at record rates in order to break even and justify their massive outlay.
I tend not to focus on that future too much. I used to do so long ago. For example, how could Facebook possibly justify their losses while asking for such a big valuation? Same for Uber. Same for any number of big companies. And it turns out that growth in the future is impossible to predict accurately. Shopify is a good example where at the time the addressable market of online stores was tiny. But it turned out that Shopify created its own market which is huge today. Technology improvements have a way of creating new markets which far surpass today's total addressable market. Factor in currency depreciation and whatnot and sometimes, futures that looked impossible turn out to be possible.
Not saying anyone is wrong in pointing at the buildouts for AI and questioning its feasibility. Just making the argument for why I personally only look at operational costs and revenue because it's the only real-ish value I can look at and judge if a business can grow sustainably.
As a counter point, the red flag to all of this is R&D costs growing for each model release. If that continues and revenue cannot outstrip it, then these companies have a problem and it'll probably be that just 1-2 frontier labs can survive this once the dust settles.
I don't think Uber was a great ROI for investors though. It lost almost all the money they gave it in return for a business with entirely average profit margins (average across all industries, far lower than average for a SaaS app).
Since Uber's never paid dividends, ROI is easy to calculate.
At the end of its first day of trading (in May 2019), Uber's price was $41.57 per share, and it is currently $72–73 per share for a compound annual growth rate (CAGR) of roughly 8.2% per year.
In comparison, the BVP Nasdaq Emerging Cloud Index earned roughly 22–25% CAGR over the same time interval.
The Uber comparison makes no sense. This is the opposite situation. Uber lost money on rides, OpenAI is (possibly) making money on inference. Uber used an R+D moonshot to autonomous driving to justify capturing an established industry without reducing costs meaningfully. OpenAI has a core product that risks becoming a commodity with open source models only 6 months behind.
The vast, vast amounts of money they spent on driver incentives city by city would seem to support the OPs claim (source: I was familiar with their spend on ads in the US approximately 10 years ago).
This is an impossible ask unless one works at Uber. I can tell you that i saw how much they were spending on ads back in 2016, and how long it continued and can assure you that they were 100% losing money back then.
Like, even now their margin is around 10% (they made 5bn on 50bn of revenue). Other software companies make a much, much, much better margin because Uber is basically not a real software business, it's an app attached to a low-margin delivery business.
Uber kept fares artificially low while simultaneously paying high bonuses to drivers to build a massive network. After burning through roughly $30+ billion over its first decade, Uber then pivoted its business model by raising rider fares, increasing restaurant fees on Uber Eats, and cutting driver pay.
Do you have a source for the claim that Uber was making money on rides during its decade of enormous unprofitability?
Its public stance was that growth was more important than profit. Why wouldn't they be subsidizing rides to fuel growth if that is their publicly stated goal?
And anyway, we got the Uber Files some years ago which made it explicit:
"In October 2014 in Madrid, the presentation shows, the hourly subsidy to drivers of $17.50 was almost twice the hourly fare it charged, which was only $9.10. In Berlin, the gross hourly fare Uber charged was $2.20, while the subsidy it paid out to drivers was $10.20 an hour. Uber burned through cash to “buy revenue”, in the words of the presentation."
"Cost of revenue" isn't the entire cost of running the company, (ie R&D, operations, sales, marketing, etc). It's just a cost they've associated with revenue IN ADDITION to the other costs I mentioned.
HSBC say they need to turn a 13b revenue to 200b by 2030 AND also find another 204b, in order to become profitable.
> It's just a cost they've associated with revenue
Its a little less arbitrary than that. Cost of Revenue/Cost of Sales/Cost of Goods Sold are clear, if you're following GAAP. To label these expenses as cost of revenue, they must meet the matching principle in that the expenses must be directly tied to the generation of specific revenue. If you didn't make that "sale" then that specific cost would not exist.
Other operating expenses come later on the income statement.
Total Revenue - Cost of Revenue = Gross Profit first, then you subtract OpEx from there for EBIT.
For OpenAI, I'd assume cost of revenue is almost directly inference costs + customer support & app dev.
I’m not sure how people are looking at numbers that show, even if we wipe off the enormous R&D expenditures, they are still in the red for inference + sales/marketing + admin and responding “this seems positive”.
It’s like being a sold a car and being told “well if you ignore the fact it has no engine it’s a good buy” yet it also has no wheels.
> Unless there's an assumption that R&D costs have to forever go up in order to increase revenue, I feel like this shows that the AI industry is actually on a path to profitability in the long term.
There are three futures right, I’ll rank them in order of fantasy -
1. Someone achieves AGI. At that point the economics of an individual company don’t even matter.
2. R&D costs do have to forever continue, because LLMs can be continually iteratively improved. Much like chip development, there is no end in sight, at least not on a near term timescale. If you are not continually at the frontier, customers will use a competitor or open/local alternatives.
3. LLMs reach a plateau of functionality. Further gains are minimal, quality reaches the apex of what the technology permits. In this scenario the hyperscalers have no business because open/local models will rapidly reach that same plateau as well.
The leaked numbers completely ignore how much of their compute is subsidized.
It also ignores how much of "R&D" is actually needed for the thing they offer to keep working. Looking at the thread everyone seems to be presuming "R&D" is all "training new models", but that is uncertain.
So... 50% operational costs and about $100 spent on sales for each paying customer.
If they manage to keep those customers for several years without more sales, that bit looks like a normal "high-touch" business.
They shouldn't look like a "high-touch" business, but their unitary numbers look way better than I expected. They just need to grow some 10 times to star making a profit... Maybe 100 to cover the opportunity cost of their capital.
It's just a matter of finding 5 billion people willing to pay US prices :)
I'd be cool with that. YouTube premium is one of the best value subscriptions I have. Steering people toward paying instead of ads-by-default is a net good imo
the AI providers will experiment with sustainable ad models and users will demand transparency and responsibility. An equilibrium will be reached, I'm sure.
What is the mental model folks have that “just do ads” is easy, like only two companies have figured out how to make money from ads and I imagine they won’t take the competition lying down.
Wouldn't the real story be to get government contracts? Those are more immune to public fickleness and market competition and usually have truly ludicrous margins.
If they're the only ones who ̶a̶r̶e̶ ̶w̶i̶l̶l̶i̶n̶g̶ ̶t̶o̶ ̶b̶e̶ ̶t̶h̶e̶ ̶e̶n̶g̶i̶n̶e̶ ̶f̶o̶r̶ ̶a̶u̶t̶o̶n̶o̶m̶o̶u̶s̶ ̶k̶i̶l̶l̶b̶o̶t̶s̶ can draw a reciprocation dingle-arm to reduce soinosoidal repleneration, then "I'm sure the government will buy it" [0]
These numbers seem insufficiently detailed to really evaluate anything. They’re had $13bn in gross revenue in 2025, and they cost of that revenue was $7.5bn. Both are growing fast (we assume) and the ratio ought to stay roughly constant.
But: how are they calculating the cost of revenue? Do they have rapidly depreciating assets that are also needed to produce that revenue? (Starlink has this issue.) Will their cost per arithmetic operation for inference rise or fall? (Anthropic is paying xAI an absolutely insane amount to lease GPUs. They must be betting that they will not need to repeat that.) Is a large portion of the cost allocated to R&D actually being used to support their revenue?
I certainly believe that the cost of inference can be plenty low for them to make a profit, but a more granular breakdown would make it easier to evaluate.
I am also curious what fraction of that revenue is government contracts. If they got say $N billion in government contracts, that is not going to have meaningful growth in the future.
To be honest I almost think the numbers are irrelevant. In 2024/25 there was a lot going on - will AI replace authors, film makers etc. Will it replace social media (anyone remember Sora?). A tonne of that stuff didn't work out. At the tail end of 2025 a real product market fit emerged. Coding agents. They work. They do a job that you can actually profit from.
So everything else is kind of academic. Of course they were losing money in 2025, they had a technology that was kind of cool - clearly eventually going to deliver something great, but they didn't actually have anything somebody should pay for. Now they have a thing that people will pay for. So who cares what they lost in 2025?
So what's important today is - how competitive are they with Anthropic in delivering that product. How do the economics of companies using AI agents for coding work. That's all. I don't think there's really an argument about them losing money on inference any more.
My key realisation from playing with open weights models on my own laptop is that at least where text is concerned, the vast majority of what an average non-programmer consumer thinks AI does, my laptop can now do with the wifi disabled. And arguably where speech and audio is concerned, too.
There is, put simply, a huge, huge information gap about the uniqueness of these commercial services.
There's an open question about how open weights models will be funded when they can't be used in a war between these companies, but the reality is that the amount Apple is paying Google for the right to distill Gemini, for example, is strongly indicative of the total size of the consumer market. Because pretty soon everyone's phones will be doing what local models can do.
Global markets will ultimately learn that coding agents are, at a first approximation, the only source of revenue for this stuff over the medium term at least, and the value proposition for consumer AI in the long term (beyond being a feature of a phone) hasn't yet been invented, and any that might exist depend on micropayments architectures that don't exist.
The free fast-follow situation with open models seems to be a big "if" here. It's not particularly hard to set up OpenCode for your company and plug in one of the myriad inference providers running free models. All of these stacks are one release away from being a dramatically different value proposition, as proved by, well, Anthropic themselves.
I guess we'll see if people will pay a premium for Anthropic in ~6 months, 12 months, etc. If not, well, it's a race to commodity.
I think there's an assumption that current costs, which are high, will remain high.
Most of that seems to be going to model development, rather than model operation, which is the product we all use.
To me the most likely option will be, that development slows and prices rise somewhat, as they can't keep burning cash forever. But I'm willing to pay a little bit more, considering how powerful the tool is.
I'm guessing that might be so in certain professions, but I would expect the employer to pay for that. For the rest of us, it seems unlikely. At least for me, I don't have a need of a device to generate text for me. And I bet most people are are in the same boat as me.
They won't try to. ChatGPT is already starting with ads, which is potentially far more profitable (as evidenced by the fact that the most profitable company of all time makes 90%+ of their revenue through ads).
>as evidenced by the fact that the most profitable company of all time makes 90%+ of their revenue through ads
the biggest reason for this is that the digital ad market is a duopoly (charitably a triopoly if you count Amazon in), if all of the LLM companies start to go into ads that's going to be a much more competitive market for ad buyers. It's not going to be so straight forward when both customers and merchants have ten different places to go.
Also not to forget that ChatGPT has zero moat, unlike social Facebook and Google.
That’s why I don’t understand why Google’s stock has gone up so much recently. They already have maximum market share of digital ads; they can only lose share to competitors like OpenAI. The only way they can make more money is through paid subscriptions.
My takeaway from this is that it's incredibly validating as a business model. Inference is _highly_ profitable. Of course, like any company that has ever tried to grow at breakneck pace, you run at a loss until you "win."
Yes, it is like a new era - the startups have huge direct revenue on real products instead of "users" which yet to be monetized.
And the network effect which ruled for the last 20 years seems to have relaxed its death grip just a bit (of course it is still there as having more customers using your tools and models provides more training data, etc., yet the current network effect doesn't seem to have that high exponential value like before)
So far as I'm aware, we don't know that inference for free users is counted as cost of revenue as opposed to sales and marketing.
With sales and marketing at 5.7b, which seems unusually high compared to cost of revenue, I think it's deeply irresponsible to ignore that when considering how profitable OpenAI may be. Even taking a charitable interpretation, OpenAI is having to spend considerable budget (more than their gross profit!) to keep people paying for using their services. The less charitable interpretation is that they're deeply unprofitable and they're pulling every legal accounting trick to hide the sources of unprofitability.
> My takeaway from this is that it's incredibly validating as a business model. Inference is _highly_ profitable.
The problem is you can't just separate training costs from inference costs. If OpenAI just didn't train a new model for the next five years, sure, they'd do OK. Assuming all those dirt cheap Chinese models nipping at their heels don't make up the gap while OpenAI is resting on their laurels.
Without being a frontier model (read: continuous, incredibly expensive training), they effectively don't have much to sell. So inference and training costs are intertwined to some extent.
Isn't this what all of the big companies that spend a lot on R&D and engineers promise?
And then the reality turns out not to be the case - you have to continuously spend on R&D to avoid getting your lunch eaten by someone else.
This isn't a social media network with lockin either. People can and will just switch to whatever whenever they feel like it. Maybe it becomes a defacto standard like google but if someone is much better than you, well...
Watch them flare out like a star… but there is lots of questions re the the return on RnD. Is it worth spending another order of magnitude for only marginal frontier gains?
People keep overlooking the fact that costs for these providers scale along with customer acquisition. Most startups don't have that linear expense. Also, training costs are accelerating to get new models out faster. One doesn't simply "get rid of R&D" costs as a comment upstream mentioned. I can't actually imagine R&D goes down anytime soon unless you're willing to play third fiddle.
Unless these frontier providers feel some type of squeeze or constraint the Chinese are well positioned to leave the US bag holders of an NVidia bound system. And if anyone has to wonder how one provider for a critical piece of infrastructure will go, well...
Even if they keep the R&D costs, more efficient inference and 0 Marketing spend also gets you there. Inference is honestly super inefficient at this point, we can do far better than GPUs, push utilisation up, build more efficient datacentres.
It's more like once you figure out how to make a really good lamp then producing lots of lamps will be profitable. But the lamps are currently suboptimal so we'll be in the red until that time.
And then someone will come up with lamp pro max and you’ll be out of business. You realize why R&D exists in tech companies even though it’s a cost center right?
OpenAI won't be able to cut R&D spend and collect rent on their existing models as long as the Chinese models keep up the pace of being ~6 months behind them for a fraction of the price.
It's more like you have a business making engines, each generation of engine has eventually turned out to be profitable over its lifespan, but each generation has an exponentially increasing R&D cost and your customers will switch from the old engines to a competitor if they don't like the newest generation.
You're stuck racing against your competitors with the distinct possibility that your R&D costs will outgrow the market demand, and you can't stop because otherwise your customers will stop investing in your dead end tech and switch.
Except that your ability to sell that lamp is conditioned on other people not giving away the same lamp for free. GLM 5.2 is free.. which leaves OpenAI with what? Harness layers that competitors do better? Seems like OpenAI needs to keep training, but distillers will always be able to distill cheaper than you can train.
> R&D costs are hurting profit side and while you can cut that one just becomes irrelevant overnight in this space if you do, hence the problem.
I see where you're coming from but I still make a distinction between "R&D costs hurt profits but are necessary to stay relevant" and "R&D costs money but is the source of future profits". It's the difference between a necessary expense and an investment.
Not to mention they will need to research how to make their models faster and cheaper to run in order to fit some margin within what people are actually willing to pay.
> As OpenAI’s worth rose, the increased value of those investor rights created a roughly $30bn charge, added the person. The charge is not expected to recur following the restructuring, they said.
> Stripping out the charge and other non-cash expenses, such as stock-based compensation of staff and computing credits from Microsoft, OpenAI’s losses were $8bn, according to the person.
Ceteris paribus, those figures imply a $45bn loss this year, $90bn loss next year and $110bn loss in 2028 before breakeven in 2029.
That's $250bn of losses to be financed from 2026 onwards. (They raised ~$120bn, $25bn up front and the rest based on milestones. So Another ~$125bn uncovered.) That only works if OpenAI stays a fundraising darling. So not a doomsday sceanario. But perilous, and dependent on short-term trends extending into long-term curves.
just for completeness, I think the closer analogue is probably total expenses: $12.48 billion to $34 billion -- roughly 2.7x. But this is still pretty close to what you said, so I don't particularly disagree with the numbers.
I do wonder if this comparison is really meaningful. It looks like if they can grow infinitely, then at some point they should be profitable. However, that's already a somewhat sad story ("in the limit as x->inf, we'll actually _make_ money!"). And there are of course limitations. Anthropic, Google, open models etc are all real competitors, and it seems to me that there will only be one winner. If openAI is losing money faster than the others, then it may not survive long enough to reach that eventual profitability. And finally, the human population is limited. There isn't a true infinity that the pattern can extend to. If we've only reached 10% of the TAM that's fine, but if we're at like 70% (which personally I suspect is about right), then this looks bad.
The AI companies also have a lot of space to grow their income (more ads, price hikes, ...). It seems realistic for them to turn profitable. But the market expected much more from these companies.
> The AI companies also have a lot of space to grow their income (more ads, price hikes, ...).
Ads, maybe, but not only are they already walking back recent price hikes, the paying customers were hitting the brakes even on the original price.
Note that this data you see (their increased revenue) came from a period where they were onboarding customers who were competing to see who used the most tokens.
IOW, this is the best-case scenario for them - customers with no cap on token spend.
But... the caps from customers came in before they hiked prices. Then they hiked prices. That resulted in a short-term boost to revenue to compensate for the caps. Now they are talking about walking back those hikes. That means they are going to find an equilibrium lower than their best-case scenario.
I like this read. Eventually, management did collectively realize that tokens spent leaderboards were a bad idea. That is going to massively reduce the waste that was needlessly being generated to hit work quotas.
> Revenue went from $3.7B to $13.07B — roughly 3.5x.
> Operating loss went from ~$8.8B to ~$20.9B — roughly 2.4x.
> Doesn't seem like a domesday scenario.
Those two lines are moving up and to the right, but are not parallel.
It all depends on where those two lines meet (the break-even point): too far in the future and the company will be dead anyway. Almost all companies will eventually be profitable; the problem is that the majority of them will need constant cash injections to keep the lights on.
Like the old aviation saying: even a brick will fly if it has enough thrust. doesn't make the brick a plane, though.
Compounding revenue & operating loss at those same rates (3.5x and 2.4x respectfully) puts those two lines meeting at around 2031. That'd be about 9-10 years to profitability, that seems pretty normal. Amazon took 9 years, Uber took 14 years before its first profitable year.
Neither Amazon nor Uber have monopolies nor much of a network effect. Amazon retail is or was famously low or near zero margin with their profits driven by AWS. Uber's margins are not much better than any average business.
>Amazon took 9 years, Uber took 14 years before its first profitable year
Both had a path to profitability in an environment of falling interest rates. OpenAI is going public in an environment of higher for longer interest rates. The discounting math is nowhere near as attractive for investors.
I think it depends on a lot of things, not the least of wish is, this could be the worst their financials get, or depending how competitive this whole thing is, it could be the best:
The scale of the numbers is exceptional, but the shape is pretty typical for a high-growth, scale startup with a big TAM where a winner can take most. And compute, supply constrained as it is for the foreseeable future, is absolutely a moat. I come away from this thinking OpenAI is actually in very good shape given that revenue is growing fast enough that break-even has a clear path without doing anything draconian.
I feel like I have a different $20 plan than everyone else. I have no problem hitting my 5 hour and weekly limits. Don’t get me wrong, it’s a great deal compared to API pricing, but it’s a far cry from “unlimited”.
I get about 20 minutes of work from my 5h limit with the $20 plan. It wouldn't bother me as much if codex would continue after the token bucket refills instead of waiting for me to show up and tell it to continue. I don't jump to the $100 plan because I would be in the exact same situation.
Harness matters in this. Using the Codex sub with Hermes eats tokens like nothing. Using it with Pi is much less but you don’t get the long term memory. When you were able to use the Claude subscription with Pi, I barely hit the 5hr limit. When they stopped allowing that, CC harness just chews thru tokens.
Interesting. I'm mostly using Claude, so perhaps I'm not nearing the limits, but I do use Codex (for coding and reviews occasionally) and use chatgpt for second opinion many times, including "pro" research. Never got to my limits. But again, not my main go to tool.
> Interesting. I'm mostly using Claude, so perhaps I'm not nearing the limits, but I do use Codex (for coding and reviews occasionally) and use chatgpt for second opinion many times, including "pro" research. Never got to my limits. But again, not my main go to tool.
I’ve hit the limit on Codex/ChatGPT one time on the $20 subscription since it came out, when I accidentally left the thinking limit on very high all day. I still could have resumed using it just 2 hours later.
AI is growing much faster than the other components of MS and Alphabet's business, and OAI is 100% dedicated to AI while the other two only have small portions on AI
Let's say Company 1 has $1B revenue and has grown 5x in the last year, and 20x the last 2 years.. Let's say Company 2 has $1B revenue and that's the same as it was last year and the year before.
> why is Amazon, a company that has never paid a dividend, such a valuable company?
You'd hope every publicly-traded long term minded company operates the same way Amazon does. Reinvestment of money they themselves earn in "growth" and still retain a trickle in profits.
Public money is looking for profits. Dividends are one way to get there. A better way to get to larger profits is selling the stock after it gets much more valuable. In the example above the growing company has a good chance of being worth much more a few years later, and that increases the value.
6bn seems excessive but despite GPT 5.5 arguably being better than Claude I don't see a lot of adoption of Codex yet.
Some of my coworkers even use Sonnet (the default in Claude Code for the 20 USD subscription) and see no reason to change even though that model is definitely "outdated" compared to current SOTA.
Marketing might help at some workplaces, presumably that are dedicated to Microsoft, for example our network blocks Claude (and DeepSeek) and is slowly rolling out Codex team by team. They should encourage Amazon/AWS to market for them.
Most people have no need for a SOTA model, and even a model from a year ago would be fine for their needs (a little bit of research, small bits of writing prose, etc).
At the end of his previous article (https://www.wheresyoured.at/ai-is-slowing-down/), Ed hyped this news as "a story that will possibly burst the AI bubble" and "imagine what the worst possible thing for me to get would be and you’re probably close." This news doesn't fit either criteria: OpenAI losing billions of dollars isn't shocking news and both AI boosters and AI skeptics have likely assumed that. If anything, the news that OpenAI has $25B on hand in cash as reported here, plus the $122B raised in March, show that OpenAI won't implode for another year or two if it does...and that doesn't say anything about the AI bubble. There's also the confounder that Codex wasn't released until this year which turbocharged revenue with an uncertain increase in operating costs, so it will be difficult to extrapolate 2025 finances to 2026 and beyond.
When I read "the worst possible thing for me to get" I had assumed it would be evidence that inference/Codex is fundamentally unprofitable (as Ed often blogs about) but there isn't enough information here to support that argument either: revenue is still greater than cost of revenue, and the major losses are clearly delineated.
Yeah, this pretty much seals it for me that Ed has basically nothing. Sure OpenAI isn’t currently profitable, but this doesn’t say to me that they can’t become so soon(ish).
> I had assumed it would be evidence that inference/Codex is fundamentally unprofitable (as Ed often blogs about)
I'm not sure where they'd get that idea from? If inference was fundamentally unprofitable, I don't think we'd have seen the massive CapEx spend & VC cash flooding into AI, it'd be a negative gross margin trap if that were the case.
It looks unprofitable because of the massive CapEx spend right now to build data centers.
People that think inference is not profitable are mistaking the total compute cost as inference cost, when really it needs separated into training compute vs. inference compute.
The bigger question is, is when does training slow down, if at all? If we hit plateaus with LLMs, at that point inference becomes nearly pure profit once you own the compute (and a hardware refresh cycle every 3-5 years).
LLMs eventually hitting a dead end for more advanced capabilities is what would spell trouble for the labs. Any existing hyperscaler cloud can run inference all day long, as long as they have access to a model. They don't need OpenAI or Anthropic for that. The frontier labs entire valuations rely purely on them staying ahead of the commodity curve. The moment they can't do that, they're done.
> I had assumed it would be evidence that inference/Codex is fundamentally unprofitable (as Ed often blogs about)
Ed's claim is that they haven't shown inference to be profitable. Which is true? And that he personally believe it is unprofitable (his personal opinion, not what his data report).
I think that's a meaningful distinction with your statement
I think Ed would argue that if 90% of your customers are only using your product because their usage is subsidised and the money to cover that subsidisation is coming from unsustainable customers tokenmaxxing then you are “fundamentally unprofitable”.
The question is, can OpenAI survive if customers start tokenminning? A pure inference business could be profitable but that’s not the business OpenAI are in. OpenAI has a billion users that OpenAI loses money on.
During the internet bubble collapse in the 00s quite some companies went bankrupt. But that's actually a good thing. It doesn't stop progress. It creates new opportunities and new baselines. Same will happen here. AI will not be less or gone or reduced to useless. It will become better , bigger and faster.
Elon must be licking his chops right now, hoping that the "OpenAI problem" will just solve itself which bumps up X.ai as a competitor to Anthropic but under the guise and financial manipulation of all of SpaceX and it's subsidiaries to fool the public into thinking it is a long term player.
I'm really curious about something: how far will you go to support AI? Clearly they'll need to monetize things further, would you still use [whatever AI you are paying for] if the price was doubled? Tripled? Where would you stop and would you stop using AI altogether or would you look at competitors?
Stretching the analogy, something that gets you from point A to point B for a fraction of the price without the same level of comfort is totally fine for me. For some of my tasks, that means using local models. For others it might mean a frontier-last-year kind of model. That's totally acceptable most of the time. For anything else I guess it's like renting a truck to move; just get the right vehicle as needed and pay the premium.
A $50k car used 1,000 miles per month probably costs close to a thousand per month, assuming 200k miles of life. I imagine this is not unusual in the US.
Is it smarter to totally cheap out and have an unreliable car that breaks down, stranding you, causing you to miss work though? If you're in a line of work that's customer facing, where having a beater of a car is going to hurt your job prospects? Without knowing the rest of the context, absolute statements based on absolute numbers is also dumb.
How many months is this car loan? What was the down payment? What's the interest rate? We haven't even asked what kind of car it is yet.
Agreed. For personal use it's already easily worth $100 a month (to me personally). More probably. For work, it's entirely based on its financial impact for a given role, and for some people/companies it will be worth the cost even at $X thousand per month per seat.
I've spent a grand total of $25 on AI ever, so apparently my answer is $25. But I'm not a big time software dev like the rest of you.
When I bought my last GPU, running AI models locally was a consideration though not the only one, and I have it set up but haven't used it much yet. I mostly use the free tiers of ChatGPT or Google to write the occasional script for me. I guess they're going to have to inject a truly unfathomable number of ads to get their money's worth.
I have a feeling my experience is closer to an average persons' than a dev, but it doesn't seem like they'll be able to monetize just from devs even if each one is spending thousands a month.
I'm not a coder but now work way faster than the coder I pay, stuff breaks but it's tenable and it's easier to get things to completion as the harnesses get better.
Don't give up just keep trying you can truly build personally life changing things. Don't look at it purely from a how do I sell this lense, just empower yourself with these tools while the getting is good
I hear you there.... too many years on my side paying over priced tools for the agencies, I'm well normalized on monthly subs and you can't compare these things to what we used to pay for. I've dropped 90% of the tools I used to sub to. But like others downstream have said no ones ever going to pay 500 to 1000 month open source models will just eat that margin up and flip the economics so ads it is lol.
For work, it depends, but if I have to spend more than a few hundreds bucks probably I'll start looking for alternatives (local models, Chinese providers, ecc)
PS: I'm in Italy, I guess in several parts of the world these figures are even smaller.
I will do nothing to “support” AI. Either it has utility or it doesn’t. I feel no loyalty or duty to help make it work if it doesn’t.
Anyway: Zero, as of right now.
I fully expect to be able to run useful LLMs on a machine I can justify buying for other reasons. I already can on the secondhand kit I own, and I don’t expect the cost-benefit analysis of local LLMs to ever really get worse.
If I ever need to pay for it, it will likely be to shift some of the capacity into the cloud for either business or pragmatic personal reasons (so I can just carry an iPad etc.)
I fully intend my expenditure to be negligible. Because once one realises that outspending others is impossible, only spending minimisation makes sense.
I foresee it potentially making sense for me to move some mature tools off a local LLM to openrouter, maybe. But probably to the same or similar models.
I don’t and won’t support AI. For a while I paid 200€ a month and would have been happy to pay up to maybe 600€. However I don’t want to participate anymore in using such an anti-human technology and industry
I make an important distinction between cloud services and local AI. My lifetime spending on cloud AI is probably less than $500, and I don't intend to spend any more. But I've already dropped $2.5k on new hardware for local inference, and could easily see myself spending more in the future. In fact, I'm regularly browsing for deals. I would also be open to paying for local models, if there was a way to make that compatible with fully open models.
AI is so important, I want to have it under my control. Even if I have to pay a penalty in terms of capabilities.
It depends on how much time it saves me and how much I make per hour generally, right?
If AI allows me to cut my time to do something in half on average or allows me to do 2x more it would be worth it to pay up to what my monthly income was before assuming my income scaled with my output.
Well probably your income will decrease slightly as your field gets automated. So you'll output twice the much for the same effort and make a bit less money.
I don’t think there’s a single person out there that will ‘support’ AI
Maybe it’s just your phrasing but people will only pay for what works, no one is loony enough to support a trillion dollar industry out of the kindness of their heart or spirit of innovation
Businesses know exactly how much they make or save thanks to AI. Take your hourly wage and count how many hours you save, and you know what it's worth to you. People who use AI for real world tasks would probably mostly accept double or even triple.
A few things to note - the financial literacy here is... sometimes lacking?
1. Revenue GROWTH is 3.5x; Expense GROWTH -> Slightly less than 3x. There's a path to profitability
2. However, the COSTS probably assume a 5 (or longer) year depreciation on GPUs. If that assumption dies, the whole thing goes down.
If R&D costs don't go up - where does the moat come from? Cheaper players catch up with 'good enough' and will erode their revenue. Most of human tasks just don't require that much intelligence.
They're racing toward 'superintelligence' that recursively self-improves.
No indication we're anywhere close to reaching it.
I think something people are missing in the headlines. The actual losses were 60b with 17b removed from the bottom line figure. To quote a reddit post "removing $17.87 billion in costs via that “net loss attributable to noncontrolling members capital”"
Look, for coding and a lot of other things, AI is awesome.
But the here's the killer. I have a dinky 16gb VRAM card, and that's kind of the sweet spot for the level of AI I actually want. I don't want it doing too much, I'd rather create slowly than have it one shot something that I have to then pore over later.
Feels like a company investing kazillions in, i don't know, air-conditioning or building wi-fi. Yes, it's going to be around, and also no one's gonna need THAT MUCH.
I'm a simple guy and I don't understand the "sales and marketing" cost.
I don't like these products. I have several negative opinions on them. To the extent they work and there is a customer base what marketing could you /possibly/ be engaged in? Doesn't the product sort of market itself? Or another way is this a product that you can market to expand your MAUs?
It's so polarizing I can't imagine how that $5.7B is being spent.
I've seen physical billboards in the Portland, OR area for OpenAI, so I guess that accounts for at least part of it. Not really sure what kind of return they're getting on those but apparently they can just do whatever they want, even if they're losing money.
I didn't look at the financials but the subscription product is heavily discounted relative to the API pricing and that difference could well be booked as a marketing expense. They also have a string of grant and similar initiatives (like $50M each) that could be marketing. There's a lot of stuff they could assign at least partially to marketing, and it sounds like they spend money pretty freely.
I've seen lots of ads saying I should use chatgpt to plan a workout or give me recipes. Thats apparently the killer app for 95% of the population at this point.
They have a large and rapidly growing enterprise sales organization. If you want to sell to enterprises you need account executives, solutions engineers, forward deployed engineers, etc.
I cannot consume any content anywhere without being slapped in the face with an unending stream of OpenAI ads and paid plugs. I'd guess most of that money is going directly to Google and Facebook.
They need marketing because they have competition that essentially offers an identical product. Why should a consumer choose openai over anthropic or whatever else there is? The answer is not obvious.
Half of the comments on this site at any given moment are from bots or shills shilling OpenAI and Anthropic. Now include Reddit, Twitter and everywhere else with a tech audience, paying for all that "organic" marketing doesn't come cheap.
>It's so polarizing I can't imagine how that $5.7B is being spent.
In every way imaginable and then more, looks like beyond the imagination :)
>I don't like these products. I have several negative opinions on them.
You're not alone, and the crowd seems to be building at the same time enthusiasts are proliferating too.
So much widespread negativity I would guess that's about what it's expected to cost to fully overcome resistance and objections. Which must be bigger than we think, they sure have more information than us.
I wonder how effective the marketing is (not much it seems).
I was watching a World Cup match last week and one of the TV ads during half time was something to the tune of ChatGPT being used by kids to improve their street soccer skills. This was Brazilian TV. Anyone even remotely familiar with Brazil would find this ad deeply, thoroughly out of touch. I can't think of a worse chatbot pitch than that.
You mean the lack of pro-Anthropic/OpenAI comments, who are gambling tokens at their casinos and won't admit that they are very expensive.
This is because people here are quietly realizing that they fell for the "token-maxxing" marketing drive which was complete BS for you to gamble more money on tokens as the big AI labs gave heavily subsidized token prices they cannot afford.
Jevon's paradox does not exist at those companies, but it certainly exists at the Chinese AI Labs at Deepseek, Alibaba, z.AI and Xiaomi.
>This is because people here are quietly realizing that they fell for the "token-maxxing" marketing drive which was complete BS for you to gamble more money on tokens as the big AI labs gave heavily subsidized token prices they cannot afford.
Good callout. All these "trends" in AI were definitely from the AI companies themselves in order to push the sales of more tokens. What's after agent orchestration? Whatever it is, it will involve a big spend.
It's the thing to do in HN comments. Downvote anything AI related and armchair diagnosing AI coding as psychosis. :/
Luckily, we've seen this before. Doom and gloom when smartphones came out. And then the same again when mobile development was preferred and there was an outcry from the web dev crowd and constant downvoting of phone apps.
Good analysis. But who cares? It takes a long time for companies to figure out how to become profitable. And I honestly believe that OpenAI/Anthropic etc. have done humanity a huge favor. The money they're burning is not yours or mine. They're institutional investor money. So, again, who cares?
It will become profitable. Local models and local on-laptop inference will get good enough. This argument has been made for decades. It's not like everyone is walking around hosting email and photos on their personal machines. Sometimes it takes a large investment to make servers and clouds for this stuff possible.
We need to get away from this idea that in order for one thing to succeed, the other must fail. We also need to stop thinking in binary and accept that all these things (profitability, local models, powerful laptops, etc.) can all happily coexist.
Buyers of consumer storage, RAM, and GPUs care. People affected by the data center buildouts care. Workers losing jobs due to underpriced tokens care. People on the receiving end of AI slop care.
People aren't affected by PC component prices. The MacBook Neo was introduced and selling extraordinarily well for $500 during this component price crisis. 99% of the population isn't building their own PCs or smartphones.
They haven't done humanity a favor at all. The innovation that these LLMs have produced has been small. A few fun math theorems where the answer was gleaned from a pattern in the training data. Great,.but it doesn't change the world one bit.
That latest drug for pancreatic cancer? Yeah, all human. After the trillions already spent, AI hasn't come up with any new medications, no new inventions to save lives... Nothing
We're only a few years into it, and yet all generations of folks are using it for all kinds of things and getting joy out of it. That's positive impact. Even folks outside of tech are having fun with it. That is a positive change for humanity. Similar to Radio, TV, smartphones, internet, microwaves and PCs.
It's already being used in the medical field in many different ways, and I believe it will be able to fold new proteins to help make new drugs. It's coming.
Just because people are using it doesn't mean it's a net good. Lots of people use social media and that's just rotting brains and making people far more polarized than they ever were before.
It's use in medicine hasn't resulted in anything meaningful. Nobody's medical bill has gotten cheaper and nobody has lived longer or healthier because of anything AI did.
Did you read it? It just said AI assisted with the process. So did the Internet. AI didn't come up with the solution or even the idea. It simply helped prune the search space, which is valuable, but not worth the trillions of dollars invested and all the added CO2.
> The money they're burning is not yours or mine. They're institutional investor money. So, again, who cares?
This is not happening in a vacuum. A lot of index funds and retirement accounts have bought into AI and AI adjacent companies, many with stakes in OpenAI. If OpenAI keels over, even when private, it will affect a lot of americans. If they IPO, it's even worse.
Index funds are based on a variety of tech stock. This whole "if they keel over" has been beaten to death ever since Tesla is surpassed Ford in market cap. And then Twitter was bought. No market crash. There will be some market corrections, but nothing be alarmed about.
This headline is not what I would read from this. The numbers are more favorable than the general tone of rumors, and point towards the expected shape of a fast-growing R&D heavy business.
This title is not how I'd actually interpret the results.
Glad to see more sane takes in the comments. All these articles on their current financials are missing the point.
OpenAI is doing pretty well.
Capital expenditure is required to deliver on 1) better models 2) better infra and 3) better products. Insane CapEx is required to do all the above + compete with Google, Meta, Microsoft, Apple, Anthropic, etc. etc. etc. who are all trying to do the same. These financials are sane, considering the scenario.
What is the right way to deal with Ed Zitron articles because he’s historically extremely inaccurate and makes wild claims.
People ignore all his horrendous takes from last year and still eat this years “analyses” like it’s Gods words.
He has been predicting the doom for years and years now and it is strange to see HN still putting credence here.
This is what he said around a week back
“ One of my sources has come forward and brought me a story that will possibly burst the AI bubble. The reason they brought this to me is that I’ve shown — and will continue to show — that I actually give a shit about this industry and the people in it.
If you’re wondering what the story is, know that it’s the information I’ve wanted for years, delivered as I have always wanted it, and I will treat it with the reverence it deserves. Imagine what the worst possible thing for me to get would be and you’re probably close.
I expect it to be out in the next two weeks, and you’ll know exactly when it runs. There’ll be a podcast and a newsletter, and very likely follow-on coverage elsewhere.
I can guarantee you it’ll be worth it, and you’ll be stunned by what I report.”
This is qanon tier stuff. He’s been pulling this shtick for a while and people still haven’t caught on.
Yeah he has zero credentials and authority and an agenda to push. Not to mention most of his articles are financially and technically illiterate and full of mistakes and inaccuracies.
I think there's some fundamental thing in his writing that speaks to people -- they want AI to fail and they want a prophet to give them reasons to think so.
It's simpler than that, some people just like sardonic writing. I don't know if I believe Ed any more than some AI cheerleader. But his writing is proper relaxing compared to hype rants that I wouldn't blame someone for suspecting to be coke-fueled.
Genuine question, do you have examples of inaccuracies and mistakes? If you ignore his caustic tone and predictions what I’ve seen reported by Ed Zitron has been accurate.
I concur. His tone greatly undermines the value of the facts he reports. Sometimes / oftentimes his analysis is off the mark, but I have not found him to be reporting falsehoods or inaccuracies.
I don't think reasoning and agentic is an improvement over what came before. It's like going back to batch processing once you've had access to an interactive terminal.
> I don't think reasoning and agentic is an improvement over what came before.
Respectfully, it shows that you haven't been using agentic models or reasoning models. I would advice you to go and use them and make an opinion afterwards. If you have come to this conclusion after extensively using these models then I don't know what to say. I guess you are the audience for Ed Zitron.
That's really not convincing. I personally don't care much about predictions, nobody is an oracle who can predict the future. If someone looks at SpaceX, talk about the financial engineering that went in the stock, then predict incorrectly that the price will go down, the meaningful information isn't the prediction, it's the analysis. The price might never go down due to how insane and detached from reality the stock market is when it comes to Elon Musk.
The facts I've seen in his reports seem to reflect reality as far as I can tell, he is correct that software companies switched from very low Capex to be extremely Capex heavy. And that announced datacenter aren't getting built. And that AI labs do not have a business model. And we've been since a few years in a financial bubble. And companies shifting to full agentic didn't take pricing into account until the switch from subscriptions to API pricing. And that nobody can say how much the use of agents cost beforehand (because both output tokens and the amount of tokens required for a given task cannot be predicted in advance). Etc.
Are you aware he's not an investor or stock analyst? His short term predictions of the market and the industry are irrelevant and doesn't invalidate the general thesis or his reported data. Have you seen mistakes and inaccuracies that aren't related to predictions?
As a side note I do personally have a thing for caustic writing, even if I wouldn't agree with his analysis I would still be happy reading some of his articles. Reminds me of blog writers from 2010
Hang on; You said financially and technically illiterate. You didn't just say the predictions were inaccurate. GP is asking for evidence of illiteracy.
I don't think people ignore anything. Every single Ed Zitron post on HN has dozens of top-level comment exactly like yours, "No no no don't listen to Ed, he's a hack and AI is great".
It's possible that I'm just not up to date with current news, but I'm having trouble connecting this quote to the article. Or really even understanding the quote at all. Can you elaborate?
The commenter above seems to be describing late stage capitalism, where businesses exist mainly to milk investors, as told by bad boy tech executive Dick Jones in the 1980's action movie RoboCop.
Remember when Nvidia gave us HBM for the 1080 ti and then took it away because it was "too expensive for consumer products"? I remember.
I feel like the 1080 ti is like a prophet of the current crisis, these companies are buying $10k paperweights per user to MAYBE... LUCKILY... charge what... $200 a year? and that is for every 1/100 users.
this same 10k hardware will be outdated in a couple of years...
It just doesn't make financial sense, if you couldn't sell standalone GPUs that people PAID for with HBM in them, what makes you think that you can sell a POSSIBLE subscription utilizing a $10k+ GPU?
The fact that people here are looking at these numbers and saying "this is fine" is absolutely bonkers.
Basically, it's a company that's not sustainable for two separate reasons. The first one is that they have an extremely high overhead. SG&A of 55% is really bad. The seconds reason is that their R&D costs are truly astronomical. They could probably cut those costs to some extent, but they're not going to cut them to nothing. They're already losing ground to Anthropic even with this much R&D.
To put it differently, even if OpenAI cut its R&D and inference costs by half, they would still be leaking money like a sieve.
These companies are clearly calling things that are R&D that aren't R&D.
If you're building a model that lasts a few months before it's no longer the most current one, and maybe a year before it's completely unusable by anybody, then that should just be COGS.
Doing that, however, would betray the real problem with this business model.
They are also likely overestimating the useful lifespan of the hardware. They keep extending the number of years on the GPUs to make the accounting look better.
Presumably when the power consumption costs more than the cost of replacement.
It’s not so much that these GPUs stop working after 3 years, but that newer GPUs can handle more requests with less power for the same purchase price. So the useful value of the GPU degrades until eventually it’s cheaper to replace than to keep running.
If the supply side constraints remains the same, I doubt they'll be releasing their GPUs as they could be considered strategic assets. Their current moat is largely hardware right.
In few years open weight models may be good enough for anything but advanced usecases. With right hardware, competition may grab the lower end of market using open models. There's also potential loss of interesting training data from real conversations.
This is the venture model now though. Spend until profitable. Uber did it. It seems OpenAI could do it as well given we seem to be in a 2 horse race for foundation models and having capital to get better pushes them further ahead.
Before Uber did it, Amazon had been doing it for almost two decades. It's nothing new. There is a difference between 1 billion and 20 billion in losses, though. Amazon in, I forget, 2014? Ran a profitable quarter with I think $1 in profits, simply to prove they were in control of their finances, and "we can stop any time we want". Sam gets a lot of shade, but he's been around the YC block once or twice, I suspect whatever risk they're taking on is at least somewhat measured.
Amazon structured their entire operation to look like this but as you indicated, could have switched to a porfit-making, dividen-paying company more than a decade ago, that just wasn't their strategy. The same can not be said for OpenAI. Even if they slashed their R&D, their marketing and sales costs are extremely high for a tech company. On paper they look more like a utility and those are not worth double-digit multiples; they compete with t-bills and GICs
Looking at the fact that third parties are making a profit offering XYZ third party open models on OpenRouter, it stands to reason that OpenAI could turn off their R&D, marketing, hype jedi, legal departments and just sell GPT9.999 and turn a profit.
Again like in the Amazon analogy, I don't think they're done growing, and unfortunately, I think they've positioned themselves (perhaps intentionally) as too big to fail, and need to continue growth at all costs.
I'm glad I'm not OAI's CFO sounds like a stressful job trying to justify/account for whatever Sam says to the board, or whatever the board demands. Sam hasn't said hardly anything since about February so I'm guessing the CFO simply bends to the will of the board these days. But that's speculation.
it stands to reason that OpenAI could turn off their R&D, marketing, hype jedi, legal departments and just sell GPT9.999 and turn a profit.
That rests on 2 assumptions:
1) That inference on OpenAI's frontier models is actually cost competitive with open models. Their high SG&A suggests otherwise.
2) That slashing R&D won't lead to a marketshare collapse when everyone (remaining) moves to Anthropic to get on their frontier models. All evidence suggests otherwise again, with Anthropic already exerting enormous competitive pressure on OpenAI's marketshare.
I think OpenAI is in a terribly tenuous position: they're getting squeezed from Anthropic (on the high end) and open models on the low end. A lot of companies in a lot of industries suffered this fate. Getting stuck in the middle is not a good thing!
It wouldn't surprise me if they have an unadvertised model router. Like, it's extremely clear you're chatting with a lobotomized model if you use voice mode in their app. Wild speculation here but I'm reasonably confident that as a $20/mo user I am not getting the same level of max thinking model that my enterprise account gets for the same question, despite both being labeled as the same model. Nowhere in their literature does it say $20/mo users get the exact same model/thinking effort, either
I’m not sure Sam is actually well-regarded around the YC block? Didn’t he lie about being the chairman of the board of YC? That’s what it says on his Wikipedia page.
Let’s not forget the whole fiasco where he was already almost fired from OpenAI.
His business track record is basically one failed location-based social media company.
I think it’s starting to become clear that OpenAI is going to be the first casualty of the AI race, and I think these undisciplined operations are a big part of the reason.
A major tell is how Apple Intelligence is seemingly steering away from OpenAI and is embracing Google instead.
Anthropic has the most useful B2B tooling and found their product niche, and they have the model leads in that niche.
xAI gets financially shielded by being a part of a gigantic financial instrument and the Elon Musk reality distortion field. Cursor has a similar product market fit as Anthropic and gets to consolidate with xAI.
Google and Microsoft get to use AI within their highly profitable ecosystems.
Apple gets to mostly sit out but act as one of the biggest toll booths for everyone else.
You've seen Sam Altman's interviews, yet you still think him a competent man? I think he's rather the embodiment of the death of meritocracy as an idea.
Uber’s situation was different, though. The reason Uber were bleeding money is because they purposefully made all their rides cheap to undercut the taxi businesses. People used Uber because it was cheaper than renting a taxi.
Now you can’t really find taxis anywhere, even at airports it’s a lot more difficult than it used to be.
Once the taxi business was disrupted enough, Uber’s pricing skyrocketed and customers had basically no other options for competition on pricing.
OpenAI basically created a new market. There is no AI chatbot incumbent to disrupt and swallow.
Some humans will need to interpret the thinking and apply it somewhere and take some responsibility for those decisions. If you think AI can do all that end to end it’s a different question but we’re nowhere near that right now.
Definitely, I’m not saying that AI can entirely replace humans. But AI is definitely replacing parts of many jobs. If AI companies raise their rates to be profitable, and it turns out that paying for profitable AI is not worth it vs paying for humans, that might be a sticky situation.
There will always be a competitor that can undercut the inference market. There is no "moat" given that you can self host decently capable LLM agents like Qwen3.6 on not super expensive hardware, like an AMD R9700, and still get competitive speeds to most cloud interfaces.
If you can self host it that easily, any Joe can scale it out much like shared web hosting, and shared web hosting or even dedicated rented boxes has always been cheaper than the big cloud providers.
I don't think OpenAI or Anthropic can reasonable compete in the long term if they can't achieve "AGI", and they won't, no matter what shareholders desire.
Actually the point is total cost wise outside of subsidy it is not cheaper than humans. the bigger problem is as the parent said open AI created a market. It is selling a commodity service with investor funds. There is no moat. your second sentence soon you won't be able to find human thinkers is on its face absurd, assuming the human race continues. Thinking is the human ecological niche.
Japan too. never thought I'd see it here but a taxi driver took the long way after a work drinking party. I guess he thought we were too drunk to notice. Well my boss sure did and lost his mind at the guy.
Likely the continued existence of taxis are keeping Uber's prices in check in the Australian market.
Uber will be running an optimisation model and be charging the maximum market can sustain, with additional goals such as eliminating competition and not being shut down by regulators.
My family and I have gone back to using car services for rides to the airport b/c "Uber XL" seems to include a WIDE variety of vehicles in terms of size and cleanliness.
A car service is about the same cost, the car looks brand new and clean and the driver is helpful.
Uber/Lyft takeover had little to do with price (though, yes, they were cheaper) and everything to do with reliability and overall quality of service. Even though ride sharing industry lost money in subsidy arms race and side bets it was fundamentally sound in major metros since early on (similar to how Amazon was fundamentally sound from early on, despite not recognizing profit for a long time). Popular "analyses" kept equating Uber/Lyft with firms losing money on every sale with no path to fix it but the demand was always there as riders had already left taxis and transit on reliability and convenience grounds.
For now, businesses are getting addicted to cheap tokens. As the screws get turned, business will debate whether they should spend budget on humans or tokens. What's further devastating is that humans are also becoming addicted to cheap tokens. Much human output is nowadays a token slopfest. People are becoming dumber too. So the real business question will be spending budget on token monkeys or tokens.
Which doesn't work the same way at all. With taxis, making them unprofitable leads to a long-lasting lack of taxis. When lots of jobs are lost, it actually becomes easier to hire someone with the right experience.
But when lots of jobs are lost, consumer spending is lost, and it becomes harder to sustain a business (whether B2C or B2B) and afford to hire someone...
It depends on how long you can keep those people un- or underemployed. I think engineers are rapidly bleeding experience even while being employed if all they do is prompting.
It might work very much the same. Discourage a cohort of CS grads into following another career path. Give businesses enough time to fully commit to “agentic workflows” such that they don’t have the expertise for in-house engineering anymore. Completely spaghettify every code base such that only AI would be willing and able to implement new features in it. Let customers lower their expectations of quality to meet what AI can product. By the time they crank up the token price, it may be hard or impossible for businesses just to switch back to human engineers.
If you knew about how much man power it takes to maintain, evaluate and improve agentic workflows, I don’t think you would write such a thing. In this context, AI is a jobs program for permanent employment.
Uber's situation is exactly the same. OpenAI is offering inference for a bunch of industries at prices that make it more competitive than hiring humans to do the same work.
If the break-even price to actually provide the service wasn't actually economic compared to humans, would there be nearly as much of a market? That's the real question. OpenAI is basically betting that they can live long enough that AI systems get built around them, which creates enough of a lock-in that they still have customers when prices increase by a lot.
I think you underestimate the price by a few orders of magnitude where it makes sense to pay a model instead of a human. If someone earning 200,000 a year gets replaced by paying 500 a day to Anthropic or OpenAI their employer comes out ahead.
There's likely always going to be value in limiting the number of $200k+ SWEs you have to pay. But that's not the interesting case.
What about the $10k/year offshored employees that are getting replaced by AI call centers? If that were the break even, then once you close down the whole building and develop the systems to not need them, then how much would inference costs have to go up before all that gets unwound and handed back to humans? It's more than you think - there's real margin there.
The uber situation was even more insidious than that. It wasn't like college students were calling cabs to go to bars in 2013. Uber created a market. It was essentially a mind virus. Gee now I can go to this place all for $7. Chum the water, establish the new pattern of living that people won't ever back away from, then twist the knife and raise prices knowing they won't revert back to whatever Old Way now long forgotten or not even engaged with by the upcoming generation.
The problem here is that open weight models are already good enough for a majority of process automation and intelligence tasks and that is where a good chunk of efficiency corporate dollars are. So there's an ever shrinking slice of inference that will hit the frontier models and inference is where the insane margins are. Now to be fair, Claude co-work and Claude code/Codex do seem magical today and these potentially will continue to be high margin/leverage plays. Frontier models are also likely to push towards decision making - so we'll have to see how it shakes out but the bottom end is already commoditized and it is getting bigger and bigger.
Yeah insane that people think it'll be okay in the long run but wondering how much different the financial status of other such company would be? Not much I guess.
I can't imagine there's a large variation in costs per token for inference and training costs between companies, and since they're all basically doing exactly the same thing, and competing on price... yeah.
I'd agree for OpenAI and Anthropic, but Google is a different story, because they're renting TPUs to both OpenAI and Anthropic, and presumably making reasonable margins on them given how supply constrained the entire industry is.
In other words, Codex likely has a lower demand:capacity ratio compared to Claude. Which can mean either OpenAI did a better job at building out capacity, or the demand for Claude outstripped Anthropic's projections. Or both.
Or Claude is better at getting people to move to more expensive membership tiers. From reading here, it seems like Claude still has a lot of users. If Claude has lower limits for their $20 plan, it stands to reason that people are paying for more expensive plans to get similar levels of usage. This assumes they aren’t reducing demand through the throttling, which is a big assumption.
I’d love to know what Anthropic’s comparable numbers look like.
huh, I felt like (gpt5.5 < opus4.8 << fable) in terms of code quality, and (g ~= o < f) in terms of pure pass rate; which one did you mean? curious about your typical workflow/tasks
I want to like gpt5.5 but it's like an evil genie: bugs are fixed, features are implemented, you are now a proud owner of a 2kloc file with a single function that makes you wish you had keybindings for horizontal scrolling
GPT 5.5 regularly got things correct where opus 4.8 got things wrong. Just today, it was launching an executable and putting the environments after the launch. It should have been
Env1=val env2=val ./executable ....
But instead it did
./executable
Then set the envvars. Chatgpt 5.5 made no such mistake.
Not here to entirely disagree its bonkers, but Tesla lost about 6b until 2022 when it got profitable and has since returned a healthy multiple of its prior losses as profits.
Amazon and Uber are other examples of businesses that looked like total basket cases for years. I remember reading, and at the time being persuaded by, articles arguing that Uber was doomed because there are no real economies of scale in livery services, and so the minute they began to achieve a dominant position and hike prices, countless competitors would easily swoop in and undercut them. Didn't turn out that way.
We are watching an experiment: how high the Tower of Babel can stand of it's built with AI slop.
"According to the narrative story in Genesis 11, the city received the name "Babel" from the Hebrew verb bālal,[e] meaning to jumble or to confuse, after Yahweh distorted the common language of humankind.[11] According to Encyclopædia Britannica, this reflects word play due to the Hebrew terms for Babylon and "to confuse" having similar pronunciation.[7]" (Wikipedia)
If anything this is MORE evidence that the infinite money printer will be coming online any second now! Yep aaaaany second now... OH THERE IT- awww one of you guys wasn't praying hard enough.
AI is a huge bubble, just as dot-com was a huge bubble (I remember when people spent huge amounts of money to have key .com domains; as much as people paid for linux.com back then, the domain essentially went nowhere), and just as buying houses for too much money with loaned money was a huge bubble.
Just as with the two previous bubble, we’re seeing companies hemorrhaging huge amounts of money, and when the dust settles the market is going to crash big time like it did with the two previous bubbles.
Unlike previous bubbles, this bubble isn’t giving people high paying jobs until everything crashes (programmers with the dot-com bubble; construction people during the real estate bubble), but it very annoyingly is making memory and SSD storage cost far too much causing computers to cost about 150% as the cost two years ago before the AI bubble was in full force, forcing Apple to make a “MacBook Neo” model with the absolute minimum of ram and SSD storage space.
Like the dot-com bubble, we will have very few winners left (with dot-com, the big winners were Amazon and Google) but unlike the previous bubble, it’s incredible how political this particular bubble is (i.e. the controversy around Grok).
Anyone remember how immensely incorrect most of HN commenters were on Uber's eventual profitability? For years we heard endless admonishment of Uber being an unsound business model. They made $10b in profit last year, $150b company at 18 P/E ratio. I would take the average HN opinion of business profitability with a grain of salt.
Everyone's financial literacy seems to evaporate when discussing AI companies. They assume that companies need to be profitable or they're a bubble waiting to burst.
The whole point of the company is that they are investing a huge amount of money upfront in order to make models that are better and better, and thus have a higher productivity multiplier.
They are very profitable on inference, they just know that the race to AGI requires a huge amount of investment, compute, getting the best researchers, etc.
I think that the issue most people have is that the degree to which they would need to be profitable in order to pay back their debt is not realistic. It is unlikely that they would be able to get that large a portion of US GDP and if they did then there will likely be riots in the streets.
It can fail, but the cost will be pushed on small retail investors, pension funds, index funds etc. The investors and managers that made it fail and waste money will be rewarded and will remain rich. It will be the "socialize losses" situation.
Seeing that R&D costs are the lion's share, I wonder if we are at a point where the focus can shift to improving the cost of inference.
Unless we are genuinely pushing to find AGI, at which point nothing matters, LLMs in their current form don't replace knowledge workers but are an effective force multiplier. How good is enough?
For instance, I pay about $1-2 a month for DeepSeek. It's not as sophisticated as Claude, but it still doubles my productivity as a SWE.
If Fable comes out and demands 50x the price of DeepSeek in order for Anthropic to make a profit on it, how much more productive would I be compared to my personal experience + DeepSeek? 3x? 50x?
Is it cost effective for a business to hire someone without SWE experience + Fable verses hiring someone with SWE experience and DeepSeek? When does R&D hit diminishing returns?
Even if you discount superhuman AI (which I would emphasize that frontier researchers do not discount and expect to see soon) think it’s still hard to have enough confidence that the ground is solid. Someone in 2024 trying to go down this route would have invested a lot of now-pointless effort into prompt engineering.
For clarity, inference is typically a COGS and therefore hits Gross Margin vs model training which would typically be in OpEx (where R&D lives) and would hit operating margin.
> I wonder if we are at a point where the focus can shift to improving the cost of inference.
There's always working on improving the cost of inference, but I don't think this is an area of R&D that will slow down. The reason is:
1. A better competitor model risks eating away at how much they can charge for inference (i.e. revenue) 2. Whoever unlocks AGI will unlock even more growth 3. Even when you unlock AGI, you'll want to throw gobs of money at it to improve itself and all sorts of things.
> If Fable comes out and demands 50x the price of DeepSeek in order for Anthropic to make a profit on it, how much more productive would I be compared to my personal experience + DeepSeek? 3x? 50x?
You're pricing it wrong and looking at it wrong. First, the per token price doesn't consider that a smarter model can end up using fewer tokens overall to achieve a result. Secondly, if the difference is between failing to accomplish the task and accomplishing the task, suddenly that 50x can seem like a bargain.
> Is it cost effective for a business to hire someone without SWE experience + Fable verses hiring someone with SWE experience and DeepSeek? When does R&D hit diminishing returns?
At this time, someone without SWE experience + <name AI model> vs someone good with SWE experience and <name another AI model> is a no-brainer. The AI model is an accelerant but the "no SWE experience" will be accelerated into a wall. Now maybe that doesn't matter for prototyping and certain other things, but anything in production the lack of experience will hurt them with things they won't even know about or even know how to look for it (e.g. slow, insecure, etc).
> Even when you unlock AGI
Massive assumption there.
Let's put it this way, how much is 5% productivity bump worth to you?
If you're in the US and you're making 100k a year, that's worth 5k or $416/m. So you can buy two of the most expensive plans on the frontier models.
This focus on cost optimization is insane. Just use the frontier models. Even a marginal bump is worth whatever the hell they're charging, at least for now.
The problem is it might be worth it to the company, but likely not to you - a 5% productivity bump likely results in $100k a year.
You really think there's zero correlation between productivity and wages? Sure, it's noisy and you might stay at $100k or even get fired. But I'd say the expected wage value of 5% higher productivity on a large sample is at least 3.5% or so.
Companies aren't paying for tokens so that their employees can capture the gains.
The one man and the dog at the Carolina spinning factory earn far more than their equivalents in Birmingham.
Everyone else will be 5% more productive. Then no one is "more" productive. So everyone has a higher output, but the same wages and hours worked. There was only a gap when usable AI first came out, some contractors could do the same quantity of work in less time and enjoy time off or do more jobs. Now the gap has closed or is closing. And using AI now is more about not being less productive than peers who do use it.
That's not how productivity works. It's not a zero-sum game.
If all construction workers can build houses 5% more efficiently, that's not the same as nothing changing. Depending on supply and demand, it means 5% more houses are built, or houses are 5% cheaper, or maybe 5% bigger, or some combination. Whether or not the construction workers all get a raise or 5% get fired (or both) depends on that supply and demand, but historically they often get a piece of the growing economic pie.
Why would the company pay more when they can just not pay more? The only things I can see happening is they might lower prices as competition ramps up, or in general as there is more supply for the same cost.
If there's sufficient demand, that's just what happens.
To try and explain one path: Company A doesn't raise wages but makes 5% more money. Company B pivots from Industry B into construction (because suddenly construction is having 5% fatter margins), and hires workers at more competitive wages to poach them from Company A. Company A forces to raise wages.
If there's a demand ceiling on housing it's a different story though.
More like company B purchases a construction company and changes nothing but number go up for shareholders while wages stay stagnant for people producing actual value, as they have for decades.
If labour supply is fixed and productivity goes up then the value and demand for labour goes up, driving up wages
Why would demand for labor rise if productivity rises?
See the increase in CEO wages vs the increase in worker wages over the last 20 years of you want to know where that 5% will almost always go.
"Historically" is doing a lot of work here. It's a well-known fact that while productivity kept increasing, wages have stagnated since the 70s.
So you're right in an academic sense, but not for any practical purpose whatsoever, in terms of how it'll affect economy now.
Or, with current US policy, the top 10% will get 6% richer and you'll get 1% poorer. Sure, the pie can grow, but you won't (unless you are in the top 10%) get any of it.
large companies aren’t buying subscription plans. my org has a 2k per month token budget per person and starting to explore optimizations like automatic model routing.
There is no evidence that these tools provide a 5% bump, if anything they are providing a 20% liability (pulling random numbers is fun).
Also where is the evidence that the workers have ever benefited from productivity bumps? The only thing that happens is surplus gets captured by the owners while workers are forced to do more.
Bad deal all around.
You're saying this like I would see that 5k in my bank account. If I'm 5% more productive that probably wouldn't even make it into annual review, let alone pay.
> Unless we are genuinely pushing to find AGI, at which point nothing matters, LLMs in their current form don't replace knowledge workers but are an effective force multiplier. How good is enough?
There's a non-negligible percentage of the industry who have a pseudo-religious belief in AGI, so I wouldn't be surprised if that was, in fact, the goal.
Who knows, maybe they'll stop once the money dries up.
If a model comes and makes developer + Deepseek even a little more productive, from employers perspective, it'd still make sense to pay a lot of money for that.
Deepseek shines for personal usage because it's possible to use it however you want and whenever you want with no session/weekly limits stress because you use the API and it's priced very reasonably.
> Unless we are genuinely pushing to find AGI, at which point nothing matters
I think the third coming out Jesus Christ in closer than AGI. Seriously, I dread how much of Silicon Valley is wrapped in this narrative of AGI and Singularity.
How can all these "rationalists" fail to see that this is what religion looks like: Faith and promises of heaven and hell.
These "rationalists" understand that beliefs should be evaluated on whether they match what we observe, rather than preconceptions about which buckets certain ideas fall into. If the Southern Baptist Convention had announced a theological breakthrough in 2022 that lets them map out the precise calendar dates of the events in Revelations, and using this map they made a series of specific predictions that ended up coming true, it would be rational to start studying End Times Christianity in more detail and irrational to say that it's religious so you're not going to worry too much about it.
That's interesting on Deepseek. But I think as long as the models are still making noticeable gains with each iteration it's hard to say "good enough."
A thousand 90 IQ cannot do what a 145 IQ can do.
Similarly, some bosses might believe that they can hire 100 cheap, unmotivated SWEs to replace Linus Torvalds or Fabrice Bellard and achieve something slightly worse. But in certain areas, it doesn't work like that.
i agree with this
on the other hand, the sad reality is many swe are working on dumb crud apps and the code-quality is already very low ime, and the jury is still out if ai tools can long term replace those; but what i have seen first-hand isnt super promising do far...
I'm a little confused here. Cost of revenue is lower than revenue. That's good. R&D is the main contributor to losses here and this seems normal in an industry like this. For OpenAI specifically, I think this is problematic. They were the first movers but despite the large R&D they've lost so much ground to Anthropic despite Anthropic seemingly gifting them with weird PR self owns. But if we were to extrapolate this to the industry as a whole, this seems more positive than negative. Am I reading this incorrectly? Unless there's an assumption that R&D costs have to forever go up in order to increase revenue, I feel like this shows that the AI industry is actually on a path to profitability in the long term.
Whether it can physically be as all encompassing as it makes itself out to be or whether it will just be healthily profitable remains to be seen. Kind of like how Uber went from "We'll autonomously drive the world" to "Look, we deliver food, goods, and people to locations and we figured out how to do that in a way that makes profits. Also, ads".
How in the world could you read that article and think there is anything positive about OpenAI's prospects? We've been hearing for months that these companies need to make trillions of dollars in a handful of years, growing at record rates in order to break even and justify their massive outlay.
It's not going to happen.
I tend not to focus on that future too much. I used to do so long ago. For example, how could Facebook possibly justify their losses while asking for such a big valuation? Same for Uber. Same for any number of big companies. And it turns out that growth in the future is impossible to predict accurately. Shopify is a good example where at the time the addressable market of online stores was tiny. But it turned out that Shopify created its own market which is huge today. Technology improvements have a way of creating new markets which far surpass today's total addressable market. Factor in currency depreciation and whatnot and sometimes, futures that looked impossible turn out to be possible.
Not saying anyone is wrong in pointing at the buildouts for AI and questioning its feasibility. Just making the argument for why I personally only look at operational costs and revenue because it's the only real-ish value I can look at and judge if a business can grow sustainably.
As a counter point, the red flag to all of this is R&D costs growing for each model release. If that continues and revenue cannot outstrip it, then these companies have a problem and it'll probably be that just 1-2 frontier labs can survive this once the dust settles.
I don't think Uber was a great ROI for investors though. It lost almost all the money they gave it in return for a business with entirely average profit margins (average across all industries, far lower than average for a SaaS app).
Since Uber's never paid dividends, ROI is easy to calculate.
At the end of its first day of trading (in May 2019), Uber's price was $41.57 per share, and it is currently $72–73 per share for a compound annual growth rate (CAGR) of roughly 8.2% per year.
In comparison, the BVP Nasdaq Emerging Cloud Index earned roughly 22–25% CAGR over the same time interval.
So (as usual) Mike Hearn is correct.
Cost of revenue is lower than revenue. That's good. R&D is the main contributor to losses here
What is counted as R&D is completely arbitrary. These figures are just playing accounting games to attempt to hide the massive ongoing costs.
We’ll see a little better when they IPO and are forced to attempt to make money but I wouldn’t invest in this business.
The Uber comparison makes no sense. This is the opposite situation. Uber lost money on rides, OpenAI is (possibly) making money on inference. Uber used an R+D moonshot to autonomous driving to justify capturing an established industry without reducing costs meaningfully. OpenAI has a core product that risks becoming a commodity with open source models only 6 months behind.
Uber didn’t lose money on rides other than some edge cases. What’s your source for this claim?
The vast, vast amounts of money they spent on driver incentives city by city would seem to support the OPs claim (source: I was familiar with their spend on ads in the US approximately 10 years ago).
There is no evidence that Uber was systemically losing money per ride instead of at edge cases. Share your evidence please.
> Share your evidence please.
This is an impossible ask unless one works at Uber. I can tell you that i saw how much they were spending on ads back in 2016, and how long it continued and can assure you that they were 100% losing money back then.
Like, even now their margin is around 10% (they made 5bn on 50bn of revenue). Other software companies make a much, much, much better margin because Uber is basically not a real software business, it's an app attached to a low-margin delivery business.
ads =/= rides
Yeah totally. In some ways Google and Facebook being so wildly profitable was very bad for future tech startups.
Nonetheless, that's the bar from a financial perspective, and I honestly don't think Uber has (or will) hit that bar.
Uber kept fares artificially low while simultaneously paying high bonuses to drivers to build a massive network. After burning through roughly $30+ billion over its first decade, Uber then pivoted its business model by raising rider fares, increasing restaurant fees on Uber Eats, and cutting driver pay.
Basically, win market through subsidy -> establish monopoly -> increase price -> profit.
Do you have a source for the claim that Uber was making money on rides during its decade of enormous unprofitability?
Its public stance was that growth was more important than profit. Why wouldn't they be subsidizing rides to fuel growth if that is their publicly stated goal?
And anyway, we got the Uber Files some years ago which made it explicit:
"In October 2014 in Madrid, the presentation shows, the hourly subsidy to drivers of $17.50 was almost twice the hourly fare it charged, which was only $9.10. In Berlin, the gross hourly fare Uber charged was $2.20, while the subsidy it paid out to drivers was $10.20 an hour. Uber burned through cash to “buy revenue”, in the words of the presentation."
https://www.theguardian.com/news/2022/jul/12/they-were-takin...
"Cost of revenue" isn't the entire cost of running the company, (ie R&D, operations, sales, marketing, etc). It's just a cost they've associated with revenue IN ADDITION to the other costs I mentioned.
HSBC say they need to turn a 13b revenue to 200b by 2030 AND also find another 204b, in order to become profitable.
> It's just a cost they've associated with revenue
Its a little less arbitrary than that. Cost of Revenue/Cost of Sales/Cost of Goods Sold are clear, if you're following GAAP. To label these expenses as cost of revenue, they must meet the matching principle in that the expenses must be directly tied to the generation of specific revenue. If you didn't make that "sale" then that specific cost would not exist.
Other operating expenses come later on the income statement.
Total Revenue - Cost of Revenue = Gross Profit first, then you subtract OpEx from there for EBIT.
For OpenAI, I'd assume cost of revenue is almost directly inference costs + customer support & app dev.
> Cost of revenue is lower than revenue.
I’m not sure how people are looking at numbers that show, even if we wipe off the enormous R&D expenditures, they are still in the red for inference + sales/marketing + admin and responding “this seems positive”.
It’s like being a sold a car and being told “well if you ignore the fact it has no engine it’s a good buy” yet it also has no wheels.
> Unless there's an assumption that R&D costs have to forever go up in order to increase revenue, I feel like this shows that the AI industry is actually on a path to profitability in the long term.
There are three futures right, I’ll rank them in order of fantasy -
1. Someone achieves AGI. At that point the economics of an individual company don’t even matter.
2. R&D costs do have to forever continue, because LLMs can be continually iteratively improved. Much like chip development, there is no end in sight, at least not on a near term timescale. If you are not continually at the frontier, customers will use a competitor or open/local alternatives.
3. LLMs reach a plateau of functionality. Further gains are minimal, quality reaches the apex of what the technology permits. In this scenario the hyperscalers have no business because open/local models will rapidly reach that same plateau as well.
The leaked numbers completely ignore how much of their compute is subsidized.
It also ignores how much of "R&D" is actually needed for the thing they offer to keep working. Looking at the thread everyone seems to be presuming "R&D" is all "training new models", but that is uncertain.
> Revenue: $13.07 billion
> Cost of Revenue: $7.5 billion
It's almost too good to be true. Did OpenAI intentionally leak this? It singlehanded eliminate the biggest concern: that tokens are sold at loss.
I think it does look like an intentional leak, but I disagree that it even shows with any clarity that inference is profitable.
So... 50% operational costs and about $100 spent on sales for each paying customer.
If they manage to keep those customers for several years without more sales, that bit looks like a normal "high-touch" business.
They shouldn't look like a "high-touch" business, but their unitary numbers look way better than I expected. They just need to grow some 10 times to star making a profit... Maybe 100 to cover the opportunity cost of their capital.
It's just a matter of finding 5 billion people willing to pay US prices :)
But it is still better than I expected.
> just a matter of finding 5 billion people willing to pay US prices
This is how you know ads are inevitable. YouTube is probably a good indicator of how BigLabs will operate for free users.
I'd be cool with that. YouTube premium is one of the best value subscriptions I have. Steering people toward paying instead of ads-by-default is a net good imo
I think we can all understand the ways in which embedded advertisement in LLMs will be fundamentally different than view-based advertisement.
the AI providers will experiment with sustainable ad models and users will demand transparency and responsibility. An equilibrium will be reached, I'm sure.
The AI providers will experiment on their users, and keep going until they lose users
It'll be like Facebook; they're not losing money but it's awful to use
Maybe this will fly in the US, but I don’t think it will in the EU and other places, with regard to laws making covert advertising illegal.
There’s also the difficulty of proving to the advertising clients that the advertising actually takes place (and how much of it), if it is covert.
You do realize that adblockers work, right?
What is the mental model folks have that “just do ads” is easy, like only two companies have figured out how to make money from ads and I imagine they won’t take the competition lying down.
My coworker is already seeing ads while using ChatGPT
Some people aren’t?
What kind of ads would you expect to see? "Hey, you're asking a lot of questions about Postgres lately. Have you considered moving to Oracle?"
Yes, their numbers can even work out into a sustainable business with ads only.
They just have to become the world dominant LLM provider by a large margin, and become a high-value ad provider with world scale.
I mean... They just have to repeat a Google. That would make them sustainable and a reasonable value for the investment.
The win for something like OpenAI isn't getting a ton of customers to pay $10-100/mo.
It's getting businesses to pay $2k/mo or more per professional employee, like a lot of Anthropic customers.
Anthropic is ahead of them there, but that is how they win.
> Anthropic is ahead of them there, but that is how they win.
Isn't Anthropic currently killing that market though? I've been hearing about a lot of businesses pulling back after having experienced the reality.
Wouldn't the real story be to get government contracts? Those are more immune to public fickleness and market competition and usually have truly ludicrous margins.
If they're the only ones who ̶a̶r̶e̶ ̶w̶i̶l̶l̶i̶n̶g̶ ̶t̶o̶ ̶b̶e̶ ̶t̶h̶e̶ ̶e̶n̶g̶i̶n̶e̶ ̶f̶o̶r̶ ̶a̶u̶t̶o̶n̶o̶m̶o̶u̶s̶ ̶k̶i̶l̶l̶b̶o̶t̶s̶ can draw a reciprocation dingle-arm to reduce soinosoidal repleneration, then "I'm sure the government will buy it" [0]
[0] https://youtu.be/Ac7G7xOG2Ag?t=89
These numbers seem insufficiently detailed to really evaluate anything. They’re had $13bn in gross revenue in 2025, and they cost of that revenue was $7.5bn. Both are growing fast (we assume) and the ratio ought to stay roughly constant.
But: how are they calculating the cost of revenue? Do they have rapidly depreciating assets that are also needed to produce that revenue? (Starlink has this issue.) Will their cost per arithmetic operation for inference rise or fall? (Anthropic is paying xAI an absolutely insane amount to lease GPUs. They must be betting that they will not need to repeat that.) Is a large portion of the cost allocated to R&D actually being used to support their revenue?
I certainly believe that the cost of inference can be plenty low for them to make a profit, but a more granular breakdown would make it easier to evaluate.
I am also curious what fraction of that revenue is government contracts. If they got say $N billion in government contracts, that is not going to have meaningful growth in the future.
To be honest I almost think the numbers are irrelevant. In 2024/25 there was a lot going on - will AI replace authors, film makers etc. Will it replace social media (anyone remember Sora?). A tonne of that stuff didn't work out. At the tail end of 2025 a real product market fit emerged. Coding agents. They work. They do a job that you can actually profit from.
So everything else is kind of academic. Of course they were losing money in 2025, they had a technology that was kind of cool - clearly eventually going to deliver something great, but they didn't actually have anything somebody should pay for. Now they have a thing that people will pay for. So who cares what they lost in 2025?
So what's important today is - how competitive are they with Anthropic in delivering that product. How do the economics of companies using AI agents for coding work. That's all. I don't think there's really an argument about them losing money on inference any more.
coding agents aren't enough to justify the amount of capital invested
My key realisation from playing with open weights models on my own laptop is that at least where text is concerned, the vast majority of what an average non-programmer consumer thinks AI does, my laptop can now do with the wifi disabled. And arguably where speech and audio is concerned, too.
There is, put simply, a huge, huge information gap about the uniqueness of these commercial services.
There's an open question about how open weights models will be funded when they can't be used in a war between these companies, but the reality is that the amount Apple is paying Google for the right to distill Gemini, for example, is strongly indicative of the total size of the consumer market. Because pretty soon everyone's phones will be doing what local models can do.
Global markets will ultimately learn that coding agents are, at a first approximation, the only source of revenue for this stuff over the medium term at least, and the value proposition for consumer AI in the long term (beyond being a feature of a phone) hasn't yet been invented, and any that might exist depend on micropayments architectures that don't exist.
The free fast-follow situation with open models seems to be a big "if" here. It's not particularly hard to set up OpenCode for your company and plug in one of the myriad inference providers running free models. All of these stacks are one release away from being a dramatically different value proposition, as proved by, well, Anthropic themselves.
I guess we'll see if people will pay a premium for Anthropic in ~6 months, 12 months, etc. If not, well, it's a race to commodity.
”The company reports over 900 million weekly active users of ChatGPT, though only about 50 million of those are paid subscribers.”
With so many free models available the ai companies are going to struggle to convert active free users to paid.
None of the free models offer anything even remotely close to the output you can get on a relatively inexpensive model.
I think that AI is going to become just another utility people pay to stay relevant. Same as their internet, electricity or gas.
Will they do it at utility / commodity prices though, or the inflated costs we see now?
That's a good question.
I think there's an assumption that current costs, which are high, will remain high.
Most of that seems to be going to model development, rather than model operation, which is the product we all use.
To me the most likely option will be, that development slows and prices rise somewhat, as they can't keep burning cash forever. But I'm willing to pay a little bit more, considering how powerful the tool is.
Sonnet 4.6 is free and works really well for coding for me from just pasting `tree` and `cat` output directly in the chat window on claude.ai
> another utility people pay to stay relevant
I'm guessing that might be so in certain professions, but I would expect the employer to pay for that. For the rest of us, it seems unlikely. At least for me, I don't have a need of a device to generate text for me. And I bet most people are are in the same boat as me.
The free models are good enough for any work (very little) that benefits from LLMs.
They won't try to. ChatGPT is already starting with ads, which is potentially far more profitable (as evidenced by the fact that the most profitable company of all time makes 90%+ of their revenue through ads).
>as evidenced by the fact that the most profitable company of all time makes 90%+ of their revenue through ads
the biggest reason for this is that the digital ad market is a duopoly (charitably a triopoly if you count Amazon in), if all of the LLM companies start to go into ads that's going to be a much more competitive market for ad buyers. It's not going to be so straight forward when both customers and merchants have ten different places to go.
Also not to forget that ChatGPT has zero moat, unlike social Facebook and Google.
That’s why I don’t understand why Google’s stock has gone up so much recently. They already have maximum market share of digital ads; they can only lose share to competitors like OpenAI. The only way they can make more money is through paid subscriptions.
When buying something thru an ad they share the commission. With A.I, they keep the commision.
Crazy that they have 50 _million_ paying subscribers and are still losing money.
Altman bough a podcast and that Ives company
They don’t care about making money at present
My takeaway from this is that it's incredibly validating as a business model. Inference is _highly_ profitable. Of course, like any company that has ever tried to grow at breakneck pace, you run at a loss until you "win."
Yes, it is like a new era - the startups have huge direct revenue on real products instead of "users" which yet to be monetized.
And the network effect which ruled for the last 20 years seems to have relaxed its death grip just a bit (of course it is still there as having more customers using your tools and models provides more training data, etc., yet the current network effect doesn't seem to have that high exponential value like before)
>> Inference is _highly_ profitable.
Totally untrue.
How much does 1 million tokens cost OpenAI?
"Revenue: 13.07b; cost of revenue: 7.5b."
This includes running inference for ~1b free users.
What is untrue about this?
So far as I'm aware, we don't know that inference for free users is counted as cost of revenue as opposed to sales and marketing.
With sales and marketing at 5.7b, which seems unusually high compared to cost of revenue, I think it's deeply irresponsible to ignore that when considering how profitable OpenAI may be. Even taking a charitable interpretation, OpenAI is having to spend considerable budget (more than their gross profit!) to keep people paying for using their services. The less charitable interpretation is that they're deeply unprofitable and they're pulling every legal accounting trick to hide the sources of unprofitability.
> My takeaway from this is that it's incredibly validating as a business model. Inference is _highly_ profitable.
The problem is you can't just separate training costs from inference costs. If OpenAI just didn't train a new model for the next five years, sure, they'd do OK. Assuming all those dirt cheap Chinese models nipping at their heels don't make up the gap while OpenAI is resting on their laurels.
Without being a frontier model (read: continuous, incredibly expensive training), they effectively don't have much to sell. So inference and training costs are intertwined to some extent.
Gemini's existence disproves this
Isn't this what all of the big companies that spend a lot on R&D and engineers promise?
And then the reality turns out not to be the case - you have to continuously spend on R&D to avoid getting your lunch eaten by someone else.
This isn't a social media network with lockin either. People can and will just switch to whatever whenever they feel like it. Maybe it becomes a defacto standard like google but if someone is much better than you, well...
If these numbers are right, it's actually not that bad. Cut r&d costs and they are mostly profitable.
Cut down on the one thing they need to keep themselves relevant in this space?
Watch them flare out like a star… but there is lots of questions re the the return on RnD. Is it worth spending another order of magnitude for only marginal frontier gains?
I bet any FAANG spend is mostly R&D.
If it's not materials, not energy or taxes, not manufacturing, not licensing or rental fees, then I can only think of R&D.
People keep overlooking the fact that costs for these providers scale along with customer acquisition. Most startups don't have that linear expense. Also, training costs are accelerating to get new models out faster. One doesn't simply "get rid of R&D" costs as a comment upstream mentioned. I can't actually imagine R&D goes down anytime soon unless you're willing to play third fiddle.
Unless these frontier providers feel some type of squeeze or constraint the Chinese are well positioned to leave the US bag holders of an NVidia bound system. And if anyone has to wonder how one provider for a critical piece of infrastructure will go, well...
Even if they keep the R&D costs, more efficient inference and 0 Marketing spend also gets you there. Inference is honestly super inefficient at this point, we can do far better than GPUs, push utilisation up, build more efficient datacentres.
So you’re saying if you cut all the cost centers a company would only have profit centers? If you ignore all the losses you’ll only have profits?
It's more like once you figure out how to make a really good lamp then producing lots of lamps will be profitable. But the lamps are currently suboptimal so we'll be in the red until that time.
And then someone will come up with lamp pro max and you’ll be out of business. You realize why R&D exists in tech companies even though it’s a cost center right?
OpenAI won't be able to cut R&D spend and collect rent on their existing models as long as the Chinese models keep up the pace of being ~6 months behind them for a fraction of the price.
And if you wait 12 months, someone will be giving away lamps for free that work just as well.
It's more like you have a business making engines, each generation of engine has eventually turned out to be profitable over its lifespan, but each generation has an exponentially increasing R&D cost and your customers will switch from the old engines to a competitor if they don't like the newest generation.
You're stuck racing against your competitors with the distinct possibility that your R&D costs will outgrow the market demand, and you can't stop because otherwise your customers will stop investing in your dead end tech and switch.
Except this is the first generation of engine manufacturers and nobody knows if it will actually be profitable yet.
We do know that Anthropic claims earlier models eventually turned a profit, and OpenAI is presumably the same.
What is in doubt is whether past performance is an indicator of future results. How long will the ever-increasing R&D expenditure keep paying off?
And there are just tons of free engines sitting around that are basically almost as good as the newest ones...
Except that your ability to sell that lamp is conditioned on other people not giving away the same lamp for free. GLM 5.2 is free.. which leaves OpenAI with what? Harness layers that competitors do better? Seems like OpenAI needs to keep training, but distillers will always be able to distill cheaper than you can train.
High risk high reward I guess
This is private equity 101 no?
If they cut down on R&D they will be no better than the open source models you can run at cost yourself.
While you cant discount 100% R&D they are close, agreed
Yes if you ignore all the reasons why they’re horribly unprofitable, they’re profitable.
R&D costs are hurting profit side and while you can cut that one just becomes irrelevant overnight in this space if you do, hence the problem.
> R&D costs are hurting profit
That’s quite the hot take, considering it’s literally an R&D company that got to where it is by doing R&D.
Isn’t the post above saying the same thing after the part where you cut it off…?
> R&D costs are hurting profit side and while you can cut that one just becomes irrelevant overnight in this space if you do, hence the problem.
I see where you're coming from but I still make a distinction between "R&D costs hurt profits but are necessary to stay relevant" and "R&D costs money but is the source of future profits". It's the difference between a necessary expense and an investment.
Not to mention they will need to research how to make their models faster and cheaper to run in order to fit some margin within what people are actually willing to pay.
OpenAI can easily cut R&D costs by replacing engineers with Claude Code
I am having difficulty parsing this sentence ... :-)
Numbers are probably not right as classifying everything aa r&d is going to the temptation
Actually reduce R&D to ZERO and they are still losing money.
Relevant: https://www.ft.com/content/e15b0d7e-ff6b-4f16-ba7a-4068feddb... this uses the same sources and answers more honestly and Ed Zitron doesn't touch on this.
> As OpenAI’s worth rose, the increased value of those investor rights created a roughly $30bn charge, added the person. The charge is not expected to recur following the restructuring, they said.
> Stripping out the charge and other non-cash expenses, such as stock-based compensation of staff and computing credits from Microsoft, OpenAI’s losses were $8bn, according to the person.
Revenue went from $3.7B to $13.07B — roughly 3.5x.
Operating loss went from ~$8.8B to ~$20.9B — roughly 2.4x.
Doesn't seem like a domesday scenario.
> Doesn't seem like a domesday scenario
Ceteris paribus, those figures imply a $45bn loss this year, $90bn loss next year and $110bn loss in 2028 before breakeven in 2029.
That's $250bn of losses to be financed from 2026 onwards. (They raised ~$120bn, $25bn up front and the rest based on milestones. So Another ~$125bn uncovered.) That only works if OpenAI stays a fundraising darling. So not a doomsday sceanario. But perilous, and dependent on short-term trends extending into long-term curves.
You're adding absolute dollars rather than using percentages - that usually isn't how that works.
Hahaha what a bozo.
Of course you don’t use percentages when the magnitude of the numbers are so high.
> rather than using percentages
Not really.
Fractions (7/2), ratios (3.5x) and percentages (+250%) are fundamentally mathematically identical.
There are a lot of problems with this back-of-the-envelope estimate, but I’m not sure the one I understand you presenting is one of them.
? https://www.theinformation.com/articles/openai-burned-3-7-bi... ?
https://www.theinformation.com/briefings/index-startup-ornn-...
Any of those three would be fine. They did not use any of them. They simply used absolute dollars.
This news matters because investors should prefer safer investments than: well at least it's not a "doomsday scenario" grade.
Tell that to the SpaceX investors.
Challenge accepted:
Facebook: https://www.facebook.com/share/v/1DC1GotK2F/
just for completeness, I think the closer analogue is probably total expenses: $12.48 billion to $34 billion -- roughly 2.7x. But this is still pretty close to what you said, so I don't particularly disagree with the numbers.
I do wonder if this comparison is really meaningful. It looks like if they can grow infinitely, then at some point they should be profitable. However, that's already a somewhat sad story ("in the limit as x->inf, we'll actually _make_ money!"). And there are of course limitations. Anthropic, Google, open models etc are all real competitors, and it seems to me that there will only be one winner. If openAI is losing money faster than the others, then it may not survive long enough to reach that eventual profitability. And finally, the human population is limited. There isn't a true infinity that the pattern can extend to. If we've only reached 10% of the TAM that's fine, but if we're at like 70% (which personally I suspect is about right), then this looks bad.
The AI companies also have a lot of space to grow their income (more ads, price hikes, ...). It seems realistic for them to turn profitable. But the market expected much more from these companies.
> The AI companies also have a lot of space to grow their income (more ads, price hikes, ...).
Ads, maybe, but not only are they already walking back recent price hikes, the paying customers were hitting the brakes even on the original price.
Note that this data you see (their increased revenue) came from a period where they were onboarding customers who were competing to see who used the most tokens.
IOW, this is the best-case scenario for them - customers with no cap on token spend.
But... the caps from customers came in before they hiked prices. Then they hiked prices. That resulted in a short-term boost to revenue to compensate for the caps. Now they are talking about walking back those hikes. That means they are going to find an equilibrium lower than their best-case scenario.
I like this read. Eventually, management did collectively realize that tokens spent leaderboards were a bad idea. That is going to massively reduce the waste that was needlessly being generated to hit work quotas.
> Revenue went from $3.7B to $13.07B — roughly 3.5x.
> Operating loss went from ~$8.8B to ~$20.9B — roughly 2.4x.
> Doesn't seem like a domesday scenario.
Those two lines are moving up and to the right, but are not parallel.
It all depends on where those two lines meet (the break-even point): too far in the future and the company will be dead anyway. Almost all companies will eventually be profitable; the problem is that the majority of them will need constant cash injections to keep the lights on.
Like the old aviation saying: even a brick will fly if it has enough thrust. doesn't make the brick a plane, though.
and again, there are good models racing right behind.
the brick has a lot of thrust but there is a airplane behind it, and it's moving on its own
Compounding revenue & operating loss at those same rates (3.5x and 2.4x respectfully) puts those two lines meeting at around 2031. That'd be about 9-10 years to profitability, that seems pretty normal. Amazon took 9 years, Uber took 14 years before its first profitable year.
both amazon and uber used that spending to deliver a network effect moat/almost monopoly.
But openai's chance of a moat on model quality is dropping as we go, not increasing
Neither Amazon nor Uber have monopolies nor much of a network effect. Amazon retail is or was famously low or near zero margin with their profits driven by AWS. Uber's margins are not much better than any average business.
>Amazon took 9 years, Uber took 14 years before its first profitable year
Both had a path to profitability in an environment of falling interest rates. OpenAI is going public in an environment of higher for longer interest rates. The discounting math is nowhere near as attractive for investors.
I think it depends on a lot of things, not the least of wish is, this could be the worst their financials get, or depending how competitive this whole thing is, it could be the best:
https://www.reuters.com/technology/openai-considers-drastic-...
The scale of the numbers is exceptional, but the shape is pretty typical for a high-growth, scale startup with a big TAM where a winner can take most. And compute, supply constrained as it is for the foreseeable future, is absolutely a moat. I come away from this thinking OpenAI is actually in very good shape given that revenue is growing fast enough that break-even has a clear path without doing anything draconian.
Is this surprising to anyone? I thought that was a given. I'm getting de-facto unlimited use of a model more expensive than Opus 4.8 for $20 a month.
I feel like I have a different $20 plan than everyone else. I have no problem hitting my 5 hour and weekly limits. Don’t get me wrong, it’s a great deal compared to API pricing, but it’s a far cry from “unlimited”.
I get about 20 minutes of work from my 5h limit with the $20 plan. It wouldn't bother me as much if codex would continue after the token bucket refills instead of waiting for me to show up and tell it to continue. I don't jump to the $100 plan because I would be in the exact same situation.
Harness matters in this. Using the Codex sub with Hermes eats tokens like nothing. Using it with Pi is much less but you don’t get the long term memory. When you were able to use the Claude subscription with Pi, I barely hit the 5hr limit. When they stopped allowing that, CC harness just chews thru tokens.
Interesting. I'm mostly using Claude, so perhaps I'm not nearing the limits, but I do use Codex (for coding and reviews occasionally) and use chatgpt for second opinion many times, including "pro" research. Never got to my limits. But again, not my main go to tool.
> de-facto unlimited (...) for $20 a month
Would love to hear some details on that one...
Or was that a typo and you meant the $200/mo plan instead maybe? That one I could believe, assuming no or frugal subagent use that is.
Copying my other reply.
> Interesting. I'm mostly using Claude, so perhaps I'm not nearing the limits, but I do use Codex (for coding and reviews occasionally) and use chatgpt for second opinion many times, including "pro" research. Never got to my limits. But again, not my main go to tool.
I’ve hit the limit on Codex/ChatGPT one time on the $20 subscription since it came out, when I accidentally left the thinking limit on very high all day. I still could have resumed using it just 2 hours later.
Napkin maths.
Alphabet: ~$4.5T value / ~$403B revenue ≈ 11× revenue
Microsoft: ~$2.9T value / ~$282B revenue ≈ 10× revenue
OpenAI: ~$850B value / ~$13B revenue ≈ 65× revenue
Can someone explains that logic?
AI is growing much faster than the other components of MS and Alphabet's business, and OAI is 100% dedicated to AI while the other two only have small portions on AI
Let's say Company 1 has $1B revenue and has grown 5x in the last year, and 20x the last 2 years.. Let's say Company 2 has $1B revenue and that's the same as it was last year and the year before.
Should these companies be valued the same?
> Should these companies be valued the same
By who? Public money is looking for dividends (profits) not growth?
If indeed the public is looking for dividends, why is Amazon, a company that has never paid a dividend, such a valuable company?
Amazon has ~10 Billion outstanding shares and the current market price for one of those shares is ~$240.
If folks only care about dividends, why would anyone buy an Amazon share at that price?
> why is Amazon, a company that has never paid a dividend, such a valuable company?
You'd hope every publicly-traded long term minded company operates the same way Amazon does. Reinvestment of money they themselves earn in "growth" and still retain a trickle in profits.
Public money is looking for profits. Dividends are one way to get there. A better way to get to larger profits is selling the stock after it gets much more valuable. In the example above the growing company has a good chance of being worth much more a few years later, and that increases the value.
SpaceX: ~$2.5T value / ~$18.7B revenue ≈ 133× revenue
Growth versus blue chip (do we even use that term anymore?)
Almost 6 bln in sales in marketing? It looks an enormous amount given that they used to have the best models and used to give-aways tokens.
6bn seems excessive but despite GPT 5.5 arguably being better than Claude I don't see a lot of adoption of Codex yet.
Some of my coworkers even use Sonnet (the default in Claude Code for the 20 USD subscription) and see no reason to change even though that model is definitely "outdated" compared to current SOTA.
Marketing might help at some workplaces, presumably that are dedicated to Microsoft, for example our network blocks Claude (and DeepSeek) and is slowly rolling out Codex team by team. They should encourage Amazon/AWS to market for them.
most people are behind the curve
Most people have no need for a SOTA model, and even a model from a year ago would be fine for their needs (a little bit of research, small bits of writing prose, etc).
I meant software engineers, not most of all living people
That is a great username right there.
At the end of his previous article (https://www.wheresyoured.at/ai-is-slowing-down/), Ed hyped this news as "a story that will possibly burst the AI bubble" and "imagine what the worst possible thing for me to get would be and you’re probably close." This news doesn't fit either criteria: OpenAI losing billions of dollars isn't shocking news and both AI boosters and AI skeptics have likely assumed that. If anything, the news that OpenAI has $25B on hand in cash as reported here, plus the $122B raised in March, show that OpenAI won't implode for another year or two if it does...and that doesn't say anything about the AI bubble. There's also the confounder that Codex wasn't released until this year which turbocharged revenue with an uncertain increase in operating costs, so it will be difficult to extrapolate 2025 finances to 2026 and beyond.
When I read "the worst possible thing for me to get" I had assumed it would be evidence that inference/Codex is fundamentally unprofitable (as Ed often blogs about) but there isn't enough information here to support that argument either: revenue is still greater than cost of revenue, and the major losses are clearly delineated.
Yeah, this pretty much seals it for me that Ed has basically nothing. Sure OpenAI isn’t currently profitable, but this doesn’t say to me that they can’t become so soon(ish).
> I had assumed it would be evidence that inference/Codex is fundamentally unprofitable (as Ed often blogs about)
I'm not sure where they'd get that idea from? If inference was fundamentally unprofitable, I don't think we'd have seen the massive CapEx spend & VC cash flooding into AI, it'd be a negative gross margin trap if that were the case.
It looks unprofitable because of the massive CapEx spend right now to build data centers.
People that think inference is not profitable are mistaking the total compute cost as inference cost, when really it needs separated into training compute vs. inference compute.
The bigger question is, is when does training slow down, if at all? If we hit plateaus with LLMs, at that point inference becomes nearly pure profit once you own the compute (and a hardware refresh cycle every 3-5 years).
LLMs eventually hitting a dead end for more advanced capabilities is what would spell trouble for the labs. Any existing hyperscaler cloud can run inference all day long, as long as they have access to a model. They don't need OpenAI or Anthropic for that. The frontier labs entire valuations rely purely on them staying ahead of the commodity curve. The moment they can't do that, they're done.
> I had assumed it would be evidence that inference/Codex is fundamentally unprofitable (as Ed often blogs about)
Ed's claim is that they haven't shown inference to be profitable. Which is true? And that he personally believe it is unprofitable (his personal opinion, not what his data report).
I think that's a meaningful distinction with your statement
I think Ed would argue that if 90% of your customers are only using your product because their usage is subsidised and the money to cover that subsidisation is coming from unsustainable customers tokenmaxxing then you are “fundamentally unprofitable”.
The question is, can OpenAI survive if customers start tokenminning? A pure inference business could be profitable but that’s not the business OpenAI are in. OpenAI has a billion users that OpenAI loses money on.
During the internet bubble collapse in the 00s quite some companies went bankrupt. But that's actually a good thing. It doesn't stop progress. It creates new opportunities and new baselines. Same will happen here. AI will not be less or gone or reduced to useless. It will become better , bigger and faster.
Elon must be licking his chops right now, hoping that the "OpenAI problem" will just solve itself which bumps up X.ai as a competitor to Anthropic but under the guise and financial manipulation of all of SpaceX and it's subsidiaries to fool the public into thinking it is a long term player.
I'm really curious about something: how far will you go to support AI? Clearly they'll need to monetize things further, would you still use [whatever AI you are paying for] if the price was doubled? Tripled? Where would you stop and would you stop using AI altogether or would you look at competitors?
I’d easily pay multiple hundreds. Possibly a thousand a month.
If I were really forced to.
LLMs provide me about the same value as a car does.
I would probably still pay if the cost doubled, but I would also look at competitors, offline solutions, etc
We have benchmarks on our domain and it does there are models that are 2x to 10x cheaper for a small drop in percentage points in accuracy
Paying a thousand a month for a car is also very stupid.
Stretching the analogy, something that gets you from point A to point B for a fraction of the price without the same level of comfort is totally fine for me. For some of my tasks, that means using local models. For others it might mean a frontier-last-year kind of model. That's totally acceptable most of the time. For anything else I guess it's like renting a truck to move; just get the right vehicle as needed and pay the premium.
A $50k car used 1,000 miles per month probably costs close to a thousand per month, assuming 200k miles of life. I imagine this is not unusual in the US.
Is it smarter to totally cheap out and have an unreliable car that breaks down, stranding you, causing you to miss work though? If you're in a line of work that's customer facing, where having a beater of a car is going to hurt your job prospects? Without knowing the rest of the context, absolute statements based on absolute numbers is also dumb.
How many months is this car loan? What was the down payment? What's the interest rate? We haven't even asked what kind of car it is yet.
Cheap car =/= unreliable car. Expensive car =/= reliable car.
Agreed. For personal use it's already easily worth $100 a month (to me personally). More probably. For work, it's entirely based on its financial impact for a given role, and for some people/companies it will be worth the cost even at $X thousand per month per seat.
That's crazy. Can you provide some examples?
I’d pay thousands a month, if I had no cheaper choices, my productivity is now limited by the intelligence of AI, I’m basically a PM now.
What in the world are you working on?
I’m honestly just thinking about day to day utility in my personal life.
I've spent a grand total of $25 on AI ever, so apparently my answer is $25. But I'm not a big time software dev like the rest of you.
When I bought my last GPU, running AI models locally was a consideration though not the only one, and I have it set up but haven't used it much yet. I mostly use the free tiers of ChatGPT or Google to write the occasional script for me. I guess they're going to have to inject a truly unfathomable number of ads to get their money's worth.
I have a feeling my experience is closer to an average persons' than a dev, but it doesn't seem like they'll be able to monetize just from devs even if each one is spending thousands a month.
I'm not a coder but now work way faster than the coder I pay, stuff breaks but it's tenable and it's easier to get things to completion as the harnesses get better.
Don't give up just keep trying you can truly build personally life changing things. Don't look at it purely from a how do I sell this lense, just empower yourself with these tools while the getting is good
I have made life changing things with it, just not anything so life changing I'd consider paying more than $25. Stingy bastard, I am.
I hear you there.... too many years on my side paying over priced tools for the agencies, I'm well normalized on monthly subs and you can't compare these things to what we used to pay for. I've dropped 90% of the tools I used to sub to. But like others downstream have said no ones ever going to pay 500 to 1000 month open source models will just eat that margin up and flip the economics so ads it is lol.
For personal use not more than 30$/month.
For work, it depends, but if I have to spend more than a few hundreds bucks probably I'll start looking for alternatives (local models, Chinese providers, ecc)
PS: I'm in Italy, I guess in several parts of the world these figures are even smaller.
Max 60 bucks a month. More than that and I'd just move to local qwen 35b or some other cheaper model on openrouter.
Zero. It provides no value to me.
I will do nothing to “support” AI. Either it has utility or it doesn’t. I feel no loyalty or duty to help make it work if it doesn’t.
Anyway: Zero, as of right now.
I fully expect to be able to run useful LLMs on a machine I can justify buying for other reasons. I already can on the secondhand kit I own, and I don’t expect the cost-benefit analysis of local LLMs to ever really get worse.
If I ever need to pay for it, it will likely be to shift some of the capacity into the cloud for either business or pragmatic personal reasons (so I can just carry an iPad etc.)
I fully intend my expenditure to be negligible. Because once one realises that outspending others is impossible, only spending minimisation makes sense.
I foresee it potentially making sense for me to move some mature tools off a local LLM to openrouter, maybe. But probably to the same or similar models.
Never paid a cent, never will pay a cent. I have my principles.
It may put me at a disadvantage when it comes to quickly slop something together? But so far the free-to-use chat bots do as well for my needs.
I don’t and won’t support AI. For a while I paid 200€ a month and would have been happy to pay up to maybe 600€. However I don’t want to participate anymore in using such an anti-human technology and industry
I pay for good tools that I use.
I spend 30 - 60 bucks a year with Horizon Labs.
I spend 25 bucks a month on Cursor. Cursor replaced an OpenAI sub.
Both support hobby projects. If either cost increased I would spend some time testing local alternatives and probably drop them.
Horizon Labs especially, I know that they have been matched by open models and are mostly a convenience at this point.
I make an important distinction between cloud services and local AI. My lifetime spending on cloud AI is probably less than $500, and I don't intend to spend any more. But I've already dropped $2.5k on new hardware for local inference, and could easily see myself spending more in the future. In fact, I'm regularly browsing for deals. I would also be open to paying for local models, if there was a way to make that compatible with fully open models.
AI is so important, I want to have it under my control. Even if I have to pay a penalty in terms of capabilities.
It depends on how much time it saves me and how much I make per hour generally, right?
If AI allows me to cut my time to do something in half on average or allows me to do 2x more it would be worth it to pay up to what my monthly income was before assuming my income scaled with my output.
> assuming my income scaled with my output
The problem is that hasn’t been the case for 50 years
I suppose it could be if you did freelance or your own business.
Well probably your income will decrease slightly as your field gets automated. So you'll output twice the much for the same effort and make a bit less money.
I don’t think there’s a single person out there that will ‘support’ AI
Maybe it’s just your phrasing but people will only pay for what works, no one is loony enough to support a trillion dollar industry out of the kindness of their heart or spirit of innovation
Businesses know exactly how much they make or save thanks to AI. Take your hourly wage and count how many hours you save, and you know what it's worth to you. People who use AI for real world tasks would probably mostly accept double or even triple.
Beginning to see why he needed seven trillion dollars.
A few things to note - the financial literacy here is... sometimes lacking? 1. Revenue GROWTH is 3.5x; Expense GROWTH -> Slightly less than 3x. There's a path to profitability 2. However, the COSTS probably assume a 5 (or longer) year depreciation on GPUs. If that assumption dies, the whole thing goes down.
If R&D costs don't go up - where does the moat come from? Cheaper players catch up with 'good enough' and will erode their revenue. Most of human tasks just don't require that much intelligence.
They're racing toward 'superintelligence' that recursively self-improves.
No indication we're anywhere close to reaching it.
Going to be an interesting year to say the least.
I think something people are missing in the headlines. The actual losses were 60b with 17b removed from the bottom line figure. To quote a reddit post "removing $17.87 billion in costs via that “net loss attributable to noncontrolling members capital”"
Ha, not a problem.
Look, for coding and a lot of other things, AI is awesome.
But the here's the killer. I have a dinky 16gb VRAM card, and that's kind of the sweet spot for the level of AI I actually want. I don't want it doing too much, I'd rather create slowly than have it one shot something that I have to then pore over later.
Feels like a company investing kazillions in, i don't know, air-conditioning or building wi-fi. Yes, it's going to be around, and also no one's gonna need THAT MUCH.
I'm a simple guy and I don't understand the "sales and marketing" cost.
I don't like these products. I have several negative opinions on them. To the extent they work and there is a customer base what marketing could you /possibly/ be engaged in? Doesn't the product sort of market itself? Or another way is this a product that you can market to expand your MAUs?
It's so polarizing I can't imagine how that $5.7B is being spent.
It costs money to get influencers to set up kool-aid stands on their platforms.
I've seen physical billboards in the Portland, OR area for OpenAI, so I guess that accounts for at least part of it. Not really sure what kind of return they're getting on those but apparently they can just do whatever they want, even if they're losing money.
I didn't look at the financials but the subscription product is heavily discounted relative to the API pricing and that difference could well be booked as a marketing expense. They also have a string of grant and similar initiatives (like $50M each) that could be marketing. There's a lot of stuff they could assign at least partially to marketing, and it sounds like they spend money pretty freely.
I've seen lots of ads saying I should use chatgpt to plan a workout or give me recipes. Thats apparently the killer app for 95% of the population at this point.
Don’t forget changing the background of a picture. This alone can triple the GDP.
Practically printing money!!!
That aligns pretty well with a past job. Those two areas were very popular user interests. Third one was cosmetics like skincare routine.
The irony being that the majority of the American population doesn’t work out or cook
They have a large and rapidly growing enterprise sales organization. If you want to sell to enterprises you need account executives, solutions engineers, forward deployed engineers, etc.
I cannot consume any content anywhere without being slapped in the face with an unending stream of OpenAI ads and paid plugs. I'd guess most of that money is going directly to Google and Facebook.
They need marketing because they have competition that essentially offers an identical product. Why should a consumer choose openai over anthropic or whatever else there is? The answer is not obvious.
OpenAI will make fully autonomous killing machines while Anthropic wont.
Land mines are my favorite fully autonomous killing machine. They've also been around for a while.
The worst ones looked like brightly coloured children's toys.
There's perhaps a metaphor or two lurking about bait and switch tactics.
They are paying influencers to pretend they use LLMs, and discredit Chinese models: https://www.wired.com/story/super-pac-backed-by-openai-and-p...
Half of the comments on this site at any given moment are from bots or shills shilling OpenAI and Anthropic. Now include Reddit, Twitter and everywhere else with a tech audience, paying for all that "organic" marketing doesn't come cheap.
Wait, shit, really? How do I collect the money? Tweet at @Sama to collect, or...?
>It's so polarizing I can't imagine how that $5.7B is being spent.
In every way imaginable and then more, looks like beyond the imagination :)
>I don't like these products. I have several negative opinions on them.
You're not alone, and the crowd seems to be building at the same time enthusiasts are proliferating too.
So much widespread negativity I would guess that's about what it's expected to cost to fully overcome resistance and objections. Which must be bigger than we think, they sure have more information than us.
I wonder how effective the marketing is (not much it seems).
I was watching a World Cup match last week and one of the TV ads during half time was something to the tune of ChatGPT being used by kids to improve their street soccer skills. This was Brazilian TV. Anyone even remotely familiar with Brazil would find this ad deeply, thoroughly out of touch. I can't think of a worse chatbot pitch than that.
Suspicious lack of pro-AI comments here
You mean the lack of pro-Anthropic/OpenAI comments, who are gambling tokens at their casinos and won't admit that they are very expensive.
This is because people here are quietly realizing that they fell for the "token-maxxing" marketing drive which was complete BS for you to gamble more money on tokens as the big AI labs gave heavily subsidized token prices they cannot afford.
Jevon's paradox does not exist at those companies, but it certainly exists at the Chinese AI Labs at Deepseek, Alibaba, z.AI and Xiaomi.
>This is because people here are quietly realizing that they fell for the "token-maxxing" marketing drive which was complete BS for you to gamble more money on tokens as the big AI labs gave heavily subsidized token prices they cannot afford.
Good callout. All these "trends" in AI were definitely from the AI companies themselves in order to push the sales of more tokens. What's after agent orchestration? Whatever it is, it will involve a big spend.
their PR department is probably still trying to figure out what narrative the bots should follow for this one.
Apparently the narrative the bots are following is that "this proves that inference is profitable!"
It's the thing to do in HN comments. Downvote anything AI related and armchair diagnosing AI coding as psychosis. :/
Luckily, we've seen this before. Doom and gloom when smartphones came out. And then the same again when mobile development was preferred and there was an outcry from the web dev crowd and constant downvoting of phone apps.
The last fad was crypto and web3. Once AI hit the scene, it's pretty much dead now.
Good analysis. But who cares? It takes a long time for companies to figure out how to become profitable. And I honestly believe that OpenAI/Anthropic etc. have done humanity a huge favor. The money they're burning is not yours or mine. They're institutional investor money. So, again, who cares?
It will become profitable. Local models and local on-laptop inference will get good enough. This argument has been made for decades. It's not like everyone is walking around hosting email and photos on their personal machines. Sometimes it takes a large investment to make servers and clouds for this stuff possible.
We need to get away from this idea that in order for one thing to succeed, the other must fail. We also need to stop thinking in binary and accept that all these things (profitability, local models, powerful laptops, etc.) can all happily coexist.
Buyers of consumer storage, RAM, and GPUs care. People affected by the data center buildouts care. Workers losing jobs due to underpriced tokens care. People on the receiving end of AI slop care.
People aren't affected by PC component prices. The MacBook Neo was introduced and selling extraordinarily well for $500 during this component price crisis. 99% of the population isn't building their own PCs or smartphones.
People who don’t want Macs or who want more powerful hardware are.
They haven't done humanity a favor at all. The innovation that these LLMs have produced has been small. A few fun math theorems where the answer was gleaned from a pattern in the training data. Great,.but it doesn't change the world one bit.
That latest drug for pancreatic cancer? Yeah, all human. After the trillions already spent, AI hasn't come up with any new medications, no new inventions to save lives... Nothing
We're only a few years into it, and yet all generations of folks are using it for all kinds of things and getting joy out of it. That's positive impact. Even folks outside of tech are having fun with it. That is a positive change for humanity. Similar to Radio, TV, smartphones, internet, microwaves and PCs.
It's already being used in the medical field in many different ways, and I believe it will be able to fold new proteins to help make new drugs. It's coming.
They already fold proteins with simulations.
Just because people are using it doesn't mean it's a net good. Lots of people use social media and that's just rotting brains and making people far more polarized than they ever were before.
It's use in medicine hasn't resulted in anything meaningful. Nobody's medical bill has gotten cheaper and nobody has lived longer or healthier because of anything AI did.
https://www.the-scientist.com/chatgpt-and-alphafold-help-des...
https://arxiv.org/pdf/2503.15204
> They already fold proteins with simulations.
I like how you're trying to argue with the Nobel committee here.
Did you read it? It just said AI assisted with the process. So did the Internet. AI didn't come up with the solution or even the idea. It simply helped prune the search space, which is valuable, but not worth the trillions of dollars invested and all the added CO2.
> The money they're burning is not yours or mine. They're institutional investor money. So, again, who cares?
This is not happening in a vacuum. A lot of index funds and retirement accounts have bought into AI and AI adjacent companies, many with stakes in OpenAI. If OpenAI keels over, even when private, it will affect a lot of americans. If they IPO, it's even worse.
Index funds are based on a variety of tech stock. This whole "if they keel over" has been beaten to death ever since Tesla is surpassed Ford in market cap. And then Twitter was bought. No market crash. There will be some market corrections, but nothing be alarmed about.
I want to see the person who thought they were losing only hundreds of millions
This headline is not what I would read from this. The numbers are more favorable than the general tone of rumors, and point towards the expected shape of a fast-growing R&D heavy business.
Sam didn't lie, they are in fact a non profit.
You are an LLM dataset's worst nightmare. XD
I do not think that means what you think it means
I'm just here for Ed's victory lap.
Of course it does. They all do. Anybody who though otherwise wasn't paying attention.
This title is not how I'd actually interpret the results.
Glad to see more sane takes in the comments. All these articles on their current financials are missing the point.
OpenAI is doing pretty well.
Capital expenditure is required to deliver on 1) better models 2) better infra and 3) better products. Insane CapEx is required to do all the above + compete with Google, Meta, Microsoft, Apple, Anthropic, etc. etc. etc. who are all trying to do the same. These financials are sane, considering the scenario.
What is the right way to deal with Ed Zitron articles because he’s historically extremely inaccurate and makes wild claims.
People ignore all his horrendous takes from last year and still eat this years “analyses” like it’s Gods words.
He has been predicting the doom for years and years now and it is strange to see HN still putting credence here.
This is what he said around a week back
“ One of my sources has come forward and brought me a story that will possibly burst the AI bubble. The reason they brought this to me is that I’ve shown — and will continue to show — that I actually give a shit about this industry and the people in it.
If you’re wondering what the story is, know that it’s the information I’ve wanted for years, delivered as I have always wanted it, and I will treat it with the reverence it deserves. Imagine what the worst possible thing for me to get would be and you’re probably close.
I expect it to be out in the next two weeks, and you’ll know exactly when it runs. There’ll be a podcast and a newsletter, and very likely follow-on coverage elsewhere.
I can guarantee you it’ll be worth it, and you’ll be stunned by what I report.”
This is qanon tier stuff. He’s been pulling this shtick for a while and people still haven’t caught on.
Yeah if this is his “information he wanted for years” it’s pretty abysmal in terms of crashing the “ai bubble”.
Yeah he has zero credentials and authority and an agenda to push. Not to mention most of his articles are financially and technically illiterate and full of mistakes and inaccuracies.
No idea why his shit keeps getting submitted.
I think there's some fundamental thing in his writing that speaks to people -- they want AI to fail and they want a prophet to give them reasons to think so.
It's simpler than that, some people just like sardonic writing. I don't know if I believe Ed any more than some AI cheerleader. But his writing is proper relaxing compared to hype rants that I wouldn't blame someone for suspecting to be coke-fueled.
Genuine question, do you have examples of inaccuracies and mistakes? If you ignore his caustic tone and predictions what I’ve seen reported by Ed Zitron has been accurate.
I concur. His tone greatly undermines the value of the facts he reports. Sometimes / oftentimes his analysis is off the mark, but I have not found him to be reporting falsehoods or inaccuracies.
It’s very very easy to find inaccuracies. In fact it’s hard to find any prediction he got right.
Here’s a compilation of things he got wrong. It’s not small things btw https://news.ycombinator.com/item?id=48447549
I don't think reasoning and agentic is an improvement over what came before. It's like going back to batch processing once you've had access to an interactive terminal.
> I don't think reasoning and agentic is an improvement over what came before.
Respectfully, it shows that you haven't been using agentic models or reasoning models. I would advice you to go and use them and make an opinion afterwards. If you have come to this conclusion after extensively using these models then I don't know what to say. I guess you are the audience for Ed Zitron.
So the Ed Zitron audience like the REPL and the hypesters like the compiler?
Here’s a small list https://news.ycombinator.com/item?id=48447549
That's really not convincing. I personally don't care much about predictions, nobody is an oracle who can predict the future. If someone looks at SpaceX, talk about the financial engineering that went in the stock, then predict incorrectly that the price will go down, the meaningful information isn't the prediction, it's the analysis. The price might never go down due to how insane and detached from reality the stock market is when it comes to Elon Musk.
The facts I've seen in his reports seem to reflect reality as far as I can tell, he is correct that software companies switched from very low Capex to be extremely Capex heavy. And that announced datacenter aren't getting built. And that AI labs do not have a business model. And we've been since a few years in a financial bubble. And companies shifting to full agentic didn't take pricing into account until the switch from subscriptions to API pricing. And that nobody can say how much the use of agents cost beforehand (because both output tokens and the amount of tokens required for a given task cannot be predicted in advance). Etc.
> I personally don't care much about predictions
I don't know what to say other than I now know the audience Ed Zitron writes for.
Are you aware he's not an investor or stock analyst? His short term predictions of the market and the industry are irrelevant and doesn't invalidate the general thesis or his reported data. Have you seen mistakes and inaccuracies that aren't related to predictions?
As a side note I do personally have a thing for caustic writing, even if I wouldn't agree with his analysis I would still be happy reading some of his articles. Reminds me of blog writers from 2010
Hang on; You said financially and technically illiterate. You didn't just say the predictions were inaccurate. GP is asking for evidence of illiteracy.
https://news.ycombinator.com/item?id=48553079
Ed Zitron is always prematurely correct, which is the worst kind of inaccurate (apparently).
I don't think people ignore anything. Every single Ed Zitron post on HN has dozens of top-level comment exactly like yours, "No no no don't listen to Ed, he's a hack and AI is great".
“I had a guaranteed military sale with ED 209, renovation program, spare parts for twenty-five years… Who cares if it worked or not?!?”
It's possible that I'm just not up to date with current news, but I'm having trouble connecting this quote to the article. Or really even understanding the quote at all. Can you elaborate?
The commenter above seems to be describing late stage capitalism, where businesses exist mainly to milk investors, as told by bad boy tech executive Dick Jones in the 1980's action movie RoboCop.
dystopian robocop reference
Remember when Nvidia gave us HBM for the 1080 ti and then took it away because it was "too expensive for consumer products"? I remember.
I feel like the 1080 ti is like a prophet of the current crisis, these companies are buying $10k paperweights per user to MAYBE... LUCKILY... charge what... $200 a year? and that is for every 1/100 users.
this same 10k hardware will be outdated in a couple of years...
It just doesn't make financial sense, if you couldn't sell standalone GPUs that people PAID for with HBM in them, what makes you think that you can sell a POSSIBLE subscription utilizing a $10k+ GPU?
This is the most obvious bubble of all time.
The fact that people here are looking at these numbers and saying "this is fine" is absolutely bonkers.
Basically, it's a company that's not sustainable for two separate reasons. The first one is that they have an extremely high overhead. SG&A of 55% is really bad. The seconds reason is that their R&D costs are truly astronomical. They could probably cut those costs to some extent, but they're not going to cut them to nothing. They're already losing ground to Anthropic even with this much R&D.
To put it differently, even if OpenAI cut its R&D and inference costs by half, they would still be leaking money like a sieve.
> SG&A
= SG&A stands for Selling, General, and Administrative expenses
These companies are clearly calling things that are R&D that aren't R&D.
If you're building a model that lasts a few months before it's no longer the most current one, and maybe a year before it's completely unusable by anybody, then that should just be COGS.
Doing that, however, would betray the real problem with this business model.
Calling it capex with an appropriate depreciation schedule is more appropriate.
They are also likely overestimating the useful lifespan of the hardware. They keep extending the number of years on the GPUs to make the accounting look better.
When are these GPUs going to be available on the second hand market?
Presumably when the power consumption costs more than the cost of replacement.
It’s not so much that these GPUs stop working after 3 years, but that newer GPUs can handle more requests with less power for the same purchase price. So the useful value of the GPU degrades until eventually it’s cheaper to replace than to keep running.
Standard depreciation is 3-5 years.
If the supply side constraints remains the same, I doubt they'll be releasing their GPUs as they could be considered strategic assets. Their current moat is largely hardware right.
In few years open weight models may be good enough for anything but advanced usecases. With right hardware, competition may grab the lower end of market using open models. There's also potential loss of interesting training data from real conversations.
I see more downsides than upsides.
This is the venture model now though. Spend until profitable. Uber did it. It seems OpenAI could do it as well given we seem to be in a 2 horse race for foundation models and having capital to get better pushes them further ahead.
Gemini is number 3 in this race
Before Uber did it, Amazon had been doing it for almost two decades. It's nothing new. There is a difference between 1 billion and 20 billion in losses, though. Amazon in, I forget, 2014? Ran a profitable quarter with I think $1 in profits, simply to prove they were in control of their finances, and "we can stop any time we want". Sam gets a lot of shade, but he's been around the YC block once or twice, I suspect whatever risk they're taking on is at least somewhat measured.
Amazon structured their entire operation to look like this but as you indicated, could have switched to a porfit-making, dividen-paying company more than a decade ago, that just wasn't their strategy. The same can not be said for OpenAI. Even if they slashed their R&D, their marketing and sales costs are extremely high for a tech company. On paper they look more like a utility and those are not worth double-digit multiples; they compete with t-bills and GICs
Looking at the fact that third parties are making a profit offering XYZ third party open models on OpenRouter, it stands to reason that OpenAI could turn off their R&D, marketing, hype jedi, legal departments and just sell GPT9.999 and turn a profit.
Again like in the Amazon analogy, I don't think they're done growing, and unfortunately, I think they've positioned themselves (perhaps intentionally) as too big to fail, and need to continue growth at all costs.
I'm glad I'm not OAI's CFO sounds like a stressful job trying to justify/account for whatever Sam says to the board, or whatever the board demands. Sam hasn't said hardly anything since about February so I'm guessing the CFO simply bends to the will of the board these days. But that's speculation.
it stands to reason that OpenAI could turn off their R&D, marketing, hype jedi, legal departments and just sell GPT9.999 and turn a profit.
That rests on 2 assumptions:
1) That inference on OpenAI's frontier models is actually cost competitive with open models. Their high SG&A suggests otherwise.
2) That slashing R&D won't lead to a marketshare collapse when everyone (remaining) moves to Anthropic to get on their frontier models. All evidence suggests otherwise again, with Anthropic already exerting enormous competitive pressure on OpenAI's marketshare.
I think OpenAI is in a terribly tenuous position: they're getting squeezed from Anthropic (on the high end) and open models on the low end. A lot of companies in a lot of industries suffered this fate. Getting stuck in the middle is not a good thing!
It wouldn't surprise me if they have an unadvertised model router. Like, it's extremely clear you're chatting with a lobotomized model if you use voice mode in their app. Wild speculation here but I'm reasonably confident that as a $20/mo user I am not getting the same level of max thinking model that my enterprise account gets for the same question, despite both being labeled as the same model. Nowhere in their literature does it say $20/mo users get the exact same model/thinking effort, either
I’m not sure Sam is actually well-regarded around the YC block? Didn’t he lie about being the chairman of the board of YC? That’s what it says on his Wikipedia page.
Let’s not forget the whole fiasco where he was already almost fired from OpenAI.
His business track record is basically one failed location-based social media company.
I think it’s starting to become clear that OpenAI is going to be the first casualty of the AI race, and I think these undisciplined operations are a big part of the reason.
A major tell is how Apple Intelligence is seemingly steering away from OpenAI and is embracing Google instead.
Anthropic has the most useful B2B tooling and found their product niche, and they have the model leads in that niche.
xAI gets financially shielded by being a part of a gigantic financial instrument and the Elon Musk reality distortion field. Cursor has a similar product market fit as Anthropic and gets to consolidate with xAI.
Google and Microsoft get to use AI within their highly profitable ecosystems.
Apple gets to mostly sit out but act as one of the biggest toll booths for everyone else.
You've seen Sam Altman's interviews, yet you still think him a competent man? I think he's rather the embodiment of the death of meritocracy as an idea.
Uber did not need this much money to cross the finish line, and faced less competition.
More importanly they have moat
This. The fact that no one seems to understand that Anth and OAI don’t have a moat is beyond me.
Or they do. If their moat is so weak, where's Grok and Gemini in this race?
Gemini is fairly well used in business, anyone using the Google office suite uses Gemini.
Grok is only used when you want to sexually harass people on Twitter.
They are working on the moat: corruption.
Uber’s situation was different, though. The reason Uber were bleeding money is because they purposefully made all their rides cheap to undercut the taxi businesses. People used Uber because it was cheaper than renting a taxi.
Now you can’t really find taxis anywhere, even at airports it’s a lot more difficult than it used to be.
Once the taxi business was disrupted enough, Uber’s pricing skyrocketed and customers had basically no other options for competition on pricing.
OpenAI basically created a new market. There is no AI chatbot incumbent to disrupt and swallow.
People use AI because it is cheaper than paying humans to think. Soon you won’t really be able to find human thinkers.
Some humans will need to interpret the thinking and apply it somewhere and take some responsibility for those decisions. If you think AI can do all that end to end it’s a different question but we’re nowhere near that right now.
Definitely, I’m not saying that AI can entirely replace humans. But AI is definitely replacing parts of many jobs. If AI companies raise their rates to be profitable, and it turns out that paying for profitable AI is not worth it vs paying for humans, that might be a sticky situation.
There will always be a competitor that can undercut the inference market. There is no "moat" given that you can self host decently capable LLM agents like Qwen3.6 on not super expensive hardware, like an AMD R9700, and still get competitive speeds to most cloud interfaces.
If you can self host it that easily, any Joe can scale it out much like shared web hosting, and shared web hosting or even dedicated rented boxes has always been cheaper than the big cloud providers.
I don't think OpenAI or Anthropic can reasonable compete in the long term if they can't achieve "AGI", and they won't, no matter what shareholders desire.
The hot-take that humans are going to stop thinking because of AI is the only convincing evidence of the idea i’ve come across so far.
Actually the point is total cost wise outside of subsidy it is not cheaper than humans. the bigger problem is as the parent said open AI created a market. It is selling a commodity service with investor funds. There is no moat. your second sentence soon you won't be able to find human thinkers is on its face absurd, assuming the human race continues. Thinking is the human ecological niche.
Thinking doesn’t disappear because of OpenAI’s shitty LLM’s.
It certainly disappear for their customers however.
You still can get taxis, at least in Australia. And they hound you at the front of the airport.
They just consistently cost more and have worse service even after uber increased prices.
Japan too. never thought I'd see it here but a taxi driver took the long way after a work drinking party. I guess he thought we were too drunk to notice. Well my boss sure did and lost his mind at the guy.
Likely the continued existence of taxis are keeping Uber's prices in check in the Australian market.
Uber will be running an optimisation model and be charging the maximum market can sustain, with additional goals such as eliminating competition and not being shut down by regulators.
My family and I have gone back to using car services for rides to the airport b/c "Uber XL" seems to include a WIDE variety of vehicles in terms of size and cleanliness.
A car service is about the same cost, the car looks brand new and clean and the driver is helpful.
What’s the price multiplier?
Uber/Lyft takeover had little to do with price (though, yes, they were cheaper) and everything to do with reliability and overall quality of service. Even though ride sharing industry lost money in subsidy arms race and side bets it was fundamentally sound in major metros since early on (similar to how Amazon was fundamentally sound from early on, despite not recognizing profit for a long time). Popular "analyses" kept equating Uber/Lyft with firms losing money on every sale with no path to fix it but the demand was always there as riders had already left taxis and transit on reliability and convenience grounds.
But there are competitors. The race is to corner the market.
They aim to undercut labor.
For now, businesses are getting addicted to cheap tokens. As the screws get turned, business will debate whether they should spend budget on humans or tokens. What's further devastating is that humans are also becoming addicted to cheap tokens. Much human output is nowadays a token slopfest. People are becoming dumber too. So the real business question will be spending budget on token monkeys or tokens.
> They aim to undercut labor.
Which doesn't work the same way at all. With taxis, making them unprofitable leads to a long-lasting lack of taxis. When lots of jobs are lost, it actually becomes easier to hire someone with the right experience.
Supply-wise yes.
But when lots of jobs are lost, consumer spending is lost, and it becomes harder to sustain a business (whether B2C or B2B) and afford to hire someone...
It depends on how long you can keep those people un- or underemployed. I think engineers are rapidly bleeding experience even while being employed if all they do is prompting.
It might work very much the same. Discourage a cohort of CS grads into following another career path. Give businesses enough time to fully commit to “agentic workflows” such that they don’t have the expertise for in-house engineering anymore. Completely spaghettify every code base such that only AI would be willing and able to implement new features in it. Let customers lower their expectations of quality to meet what AI can product. By the time they crank up the token price, it may be hard or impossible for businesses just to switch back to human engineers.
If you knew about how much man power it takes to maintain, evaluate and improve agentic workflows, I don’t think you would write such a thing. In this context, AI is a jobs program for permanent employment.
Uber's situation is exactly the same. OpenAI is offering inference for a bunch of industries at prices that make it more competitive than hiring humans to do the same work.
If the break-even price to actually provide the service wasn't actually economic compared to humans, would there be nearly as much of a market? That's the real question. OpenAI is basically betting that they can live long enough that AI systems get built around them, which creates enough of a lock-in that they still have customers when prices increase by a lot.
I think you underestimate the price by a few orders of magnitude where it makes sense to pay a model instead of a human. If someone earning 200,000 a year gets replaced by paying 500 a day to Anthropic or OpenAI their employer comes out ahead.
There's likely always going to be value in limiting the number of $200k+ SWEs you have to pay. But that's not the interesting case.
What about the $10k/year offshored employees that are getting replaced by AI call centers? If that were the break even, then once you close down the whole building and develop the systems to not need them, then how much would inference costs have to go up before all that gets unwound and handed back to humans? It's more than you think - there's real margin there.
The uber situation was even more insidious than that. It wasn't like college students were calling cabs to go to bars in 2013. Uber created a market. It was essentially a mind virus. Gee now I can go to this place all for $7. Chum the water, establish the new pattern of living that people won't ever back away from, then twist the knife and raise prices knowing they won't revert back to whatever Old Way now long forgotten or not even engaged with by the upcoming generation.
Many such cases.
> It wasn't like college students were calling cabs to go to bars in 2013.
well, in 2009, we did.
The AI chatbot incumbent is the swath of human professionals, no?
The problem here is that open weight models are already good enough for a majority of process automation and intelligence tasks and that is where a good chunk of efficiency corporate dollars are. So there's an ever shrinking slice of inference that will hit the frontier models and inference is where the insane margins are. Now to be fair, Claude co-work and Claude code/Codex do seem magical today and these potentially will continue to be high margin/leverage plays. Frontier models are also likely to push towards decision making - so we'll have to see how it shakes out but the bottom end is already commoditized and it is getting bigger and bigger.
Yeah insane that people think it'll be okay in the long run but wondering how much different the financial status of other such company would be? Not much I guess.
I can't imagine there's a large variation in costs per token for inference and training costs between companies, and since they're all basically doing exactly the same thing, and competing on price... yeah.
I'd agree for OpenAI and Anthropic, but Google is a different story, because they're renting TPUs to both OpenAI and Anthropic, and presumably making reasonable margins on them given how supply constrained the entire industry is.
That's a good point!
They're not really losing ground to Anthropic. 5.5 was a bit better than 4.8. Fable was good, and was a jump over 4.8, but only incremental over 5.5.
Anthropic is also likely losing money, right?
I never get throttled by Codex at $20/mo however Claude throttles faster at same rate. I like Claude’s output in terms of code however
In other words, Codex likely has a lower demand:capacity ratio compared to Claude. Which can mean either OpenAI did a better job at building out capacity, or the demand for Claude outstripped Anthropic's projections. Or both.
Or Codex models are more efficient that Claude. Plus the two things you mentioned.
Or Claude is better at getting people to move to more expensive membership tiers. From reading here, it seems like Claude still has a lot of users. If Claude has lower limits for their $20 plan, it stands to reason that people are paying for more expensive plans to get similar levels of usage. This assumes they aren’t reducing demand through the throttling, which is a big assumption.
I’d love to know what Anthropic’s comparable numbers look like.
huh, I felt like (gpt5.5 < opus4.8 << fable) in terms of code quality, and (g ~= o < f) in terms of pure pass rate; which one did you mean? curious about your typical workflow/tasks
I want to like gpt5.5 but it's like an evil genie: bugs are fixed, features are implemented, you are now a proud owner of a 2kloc file with a single function that makes you wish you had keybindings for horizontal scrolling
GPT 5.5 regularly got things correct where opus 4.8 got things wrong. Just today, it was launching an executable and putting the environments after the launch. It should have been
Env1=val env2=val ./executable ....
But instead it did
./executable
Then set the envvars. Chatgpt 5.5 made no such mistake.
According to Wash they are operationally profitable
https://www.wsj.com/tech/ai/mind-blowing-growth-is-about-to-...
Some question it
https://www.wheresyoured.at/anthropics-profitability-swindle...
>The projections, which were reviewed by The Wall Street Journal
No, they're just projections of profitability
You're over inflating the S which is expected to increase as now they are "going to market" G&A is within expectations.
Revenue is still growing faster than costs and gross margins have continued to improve.
The real question is when they can start spending less on R&D and still compete.
> The real question is when they can start spending less on R&D and still compete.
As a company making SOTA models? Never.
Not here to entirely disagree its bonkers, but Tesla lost about 6b until 2022 when it got profitable and has since returned a healthy multiple of its prior losses as profits.
Tesla didn't have any real competition until recently from chinese side. 2nd OpenAI is a software company unlike Tesla
Amazon and Uber are other examples of businesses that looked like total basket cases for years. I remember reading, and at the time being persuaded by, articles arguing that Uber was doomed because there are no real economies of scale in livery services, and so the minute they began to achieve a dominant position and hike prices, countless competitors would easily swoop in and undercut them. Didn't turn out that way.
Profits are irreducibly profitable, regardless.
Tesla's losses were always a small fraction of annual revenue. OpenAI loses multiples of annual revenue.
> They're already losing ground to Anthropic even with this much R&D.
Do we know how bad/good misAnthropic is doing financially?
Petitioning the corrupt head of state to force Anthropic out of business seems to be part of the business model.
We are watching an experiment: how high the Tower of Babel can stand of it's built with AI slop.
"According to the narrative story in Genesis 11, the city received the name "Babel" from the Hebrew verb bālal,[e] meaning to jumble or to confuse, after Yahweh distorted the common language of humankind.[11] According to Encyclopædia Britannica, this reflects word play due to the Hebrew terms for Babylon and "to confuse" having similar pronunciation.[7]" (Wikipedia)
Leaked: OpenAI is a rapidly scaling startup, has economics similar to other startups
If anything this is MORE evidence that the infinite money printer will be coming online any second now! Yep aaaaany second now... OH THERE IT- awww one of you guys wasn't praying hard enough.
I'm not surprised
AI is a huge bubble, just as dot-com was a huge bubble (I remember when people spent huge amounts of money to have key .com domains; as much as people paid for linux.com back then, the domain essentially went nowhere), and just as buying houses for too much money with loaned money was a huge bubble.
Just as with the two previous bubble, we’re seeing companies hemorrhaging huge amounts of money, and when the dust settles the market is going to crash big time like it did with the two previous bubbles.
Unlike previous bubbles, this bubble isn’t giving people high paying jobs until everything crashes (programmers with the dot-com bubble; construction people during the real estate bubble), but it very annoyingly is making memory and SSD storage cost far too much causing computers to cost about 150% as the cost two years ago before the AI bubble was in full force, forcing Apple to make a “MacBook Neo” model with the absolute minimum of ram and SSD storage space.
Like the dot-com bubble, we will have very few winners left (with dot-com, the big winners were Amazon and Google) but unlike the previous bubble, it’s incredible how political this particular bubble is (i.e. the controversy around Grok).
what a surprise! who would have thought, right?
Pardon my French, but yeah, no shit?
AI companies are black holes for money the way delivery companies are (or were, considering the money people are willing to pay these days).
Most of them will disappear alongside the money people have bet on them.
Anyone remember how immensely incorrect most of HN commenters were on Uber's eventual profitability? For years we heard endless admonishment of Uber being an unsound business model. They made $10b in profit last year, $150b company at 18 P/E ratio. I would take the average HN opinion of business profitability with a grain of salt.
Who needed leaks to know that?
Everyone's financial literacy seems to evaporate when discussing AI companies. They assume that companies need to be profitable or they're a bubble waiting to burst.
The whole point of the company is that they are investing a huge amount of money upfront in order to make models that are better and better, and thus have a higher productivity multiplier.
They are very profitable on inference, they just know that the race to AGI requires a huge amount of investment, compute, getting the best researchers, etc.
I think that the issue most people have is that the degree to which they would need to be profitable in order to pay back their debt is not realistic. It is unlikely that they would be able to get that large a portion of US GDP and if they did then there will likely be riots in the streets.
WHy would there be riots in the streets?
People are gonna lose so much money on their upcoming IPO lol
it will be so satisfying to see them crash and burn
Ed Zitron has proven trump wrong so many times it's going to be hilarious how right it will come out on this
They know it is a scam, but it doesn’t matter as it is now too late.
That ship has sailed long ago into the IPO sunset.
That’s absurd. Why couldn’t it still fail, especially when their last raise was at 20x revenue or more? These numbers are horrendous.
It can fail, but the cost will be pushed on small retail investors, pension funds, index funds etc. The investors and managers that made it fail and waste money will be rewarded and will remain rich. It will be the "socialize losses" situation.