Microsoft has lots of revenue streams tied to that capex outlay. Does OpenAI have similar revenue numbers to Microsoft?

OpenAI has a very healthy revenue stream in the form of other companies throwing money at them.

But to answer your question, no they aren’t even profitable by themselves.

> they aren’t even profitable

Depends on your definition of profitability, They are not recovering R&D and training costs, but they (and MS) are recouping inference costs from user subscription and API revenue with a healthy operating margin.

Today they will not survive if they stop investing in R&D, but they do have to slow down at some point. It looks like they and other big players are betting on a moat they hope to build with the $100B DCs and ASICs that open weight models or others cannot compete with.

This will be either because training will be too expensive (few entities have the budget for $10B+ on training and no need to monetize it) and even those kind of models where available may be impossible to run inference with off the shelf GPUs, i.e. these models can only run on ASICS, which only large players will have access to[1].

In this scenario corporations will have to pay them the money for the best models, when that happens OpenAI can slow down R&D and become profitable with capex considered.

[1] This is natural progression in a compute bottle-necked sector, we saw a similar evolution from CPU to ASICS and GPU in the crypto few years ago. It is slightly distorted comparison due to the switch from PoW to PoS and intentional design for GPU for some coins, even then you needed DC scale operations in a cheap power location to be profitable.

They will have an endless wave of commoditization chasing behind them. NVIDIA will continue to market chips to anyone who will buy... Well anyone who is allowed to buy, considering the recent export restrictions. On that note, if OpenAI is in bed with the US government with this to some degree, I would expect tariffs, expert restrictions, and all of that to continue to conveniently align with their business objectives.

If the frontier models generate huge revenue from big government and intelligence and corporate contracts, then I can see a dynamo kicking off with the business model. The missing link is probably that there need to be continual breakthroughs that massively increase the power of AI rather than it tapering off with diminishing returns for bigger training/inference capital outlay. Obviously, openAI is leveraging against that view as well.

Maybe the most important part is that all of these huge names are involved in the project to some degree. Well, they're all cross-linked in the entire AI enterprise, really, like OpenAI Microsoft, so once all the players give preference to each other, it sort of creates a moat in and of itself, unless foreign sovereign wealth funds start spinning up massive stargate initiatives as well.

We'll see. Europe has been behind the ball in tech developments like this historically, and China, although this might be a bit of a stretch to claim, does seem to be held back by their need for control and censorship when it comes to what these models can do. They want them to be focused tools that help society, but the American companies want much more, and they want power in their own hands and power in their user's hands. So much like the first round where American big tech took over the world, maybe it's prime to happen again as the AI industry continues to scale.

Why would China censoring Tiananmen Square/whatever out of their LLMs be anymore harmful to the training process when the US controlled LLMs also censor certain topics, eg "how do I make meth?" or "how do I make a nuclear bomb?".

Because China censors very common words and phrases such as "harmonized", "shameless", "lifelong", "river crabbed", "me too". This is because Chinese citizens uses puns and common phrases initially to get around censors.

OpenAI models refuse to translate subtitles because they contain violence, sex, or racism.

That’s just a different flavour of enforced right-think.

They are absolutely different flavors. OpenAI is not being told by the government to censor violence, sex or racism - they're being told that by their executives.

News flash: household-name businesses aren't going to repeat slurs if the media will use it to defame them. Nevermind the fact that people will (rightfully) hold you legally accountable and demand your testimony when ChatGPT starts offering unsupervised chemistry lessons - the threat of bad PR is all that is required to censor their models.

There's no agenda removing porn from ChatGPT any more than there's an agenda removing porn from the App Store or YouTube. It's about shrewd identity politics, not prudish shadow government conspiracies against you seeing sex and being bigoted.

Sigh. No. Censorship is censorship is censorship. That is true even if you happen to like and can generate a plausible defense of US version that happens to be business friendly ( as opposed to China's ruling party friendly ).

> Censorship is censorship is censorship

"if your company doesn't present hardcore fisting pornography to five year olds you're a tyrant" is a heck of a take, even for hacker news.

It is not a take. It is simple position of 'just because you call something as involuntary semen injection does not make it any less of a rape'. I like things that are clear and well defined. And so I repeat:

Censorship is censorship is censorship.

Ok, I guess I'm #TeamProCensorship, then. So is almost everyone.

I am not sure if it will surprise you, but your affiliation or the size of your 'team' is largely irrelevant from my perspective. That said, I am mildly surprised you were able to accept the new self-image as willing censor though. Most people struggle with that ( edit: hence the 'this is not censorship' facade ).

They're accepting your definition of censorship to highlight how fucking stupid it is. Is Hacker News a censorship haven because I flagged the "How to have Sex with Cars" post uploaded yesterday? Am I a tyrant for trying to oppress that poor user's voice? No. I'm upholding the guidelines of a privately owned and moderated community.

"Censorship is censorship is censorship" is the sort of defense you'd rely on if you were caught selling guns and kiddie porn on the internet. It's not the sort of defense OpenAI needs to use though, because they have a semblance of self-preservation instinct and would rather not let ChatGPT say something capable of pissing off the IMF or ADL. Call that "censorship" all you want - it's like canvassing for your right to yell 'fire!' in a movie theater.

<< IMF or ADL.

Friend, neither of those is a body that can say constitution in US is null and void. Nor to they get to pick and choose which speech is kosher. It is not up to those orgs to decide.

<< They're accepting your definition of censorship to highlight how fucking stupid it is.

They are accepting it, because there is no way it cannot not be accepted. Now.. just because there is some cognitive dissonance over what should logically follow is a separate issue entirely.

Best I can do is spread some seeds..

Yes, that's true. It's very rare for people to be able to value actual free speech. Most people think they do until they hear something they don't like

However private individuals or companies deciding not to offer certain products is an expression of free speech.

i.e. denying someone who is running an online platform/community or training an LLM model or whatever the right to remove or not provide specific content is a clearly limiting their right to freedom of expression.

Usually a sign of great discussion when someone responds with "sigh" to a reasonably presented argument.

I don't know why people care if they're being censored by government officials or private billionaires. What difference does it make at the end of the day? why is one worse than the other?

Because you aren't being "censored" by billionaires at all. They have made the business decision to reduce the usefulness of their AI to prevent their liability from being legally, or even socially, held accountable.

Again, consider my example about YouTube - it's not illegal for Google to put pornography on YouTube. They still moderate it out though, not because they want to "censor" their users but because amateur porn is a liability nightmare to moderate. Similarly, I don't think ChatGPT's limitations qualify as censorship.

Okay, i mean you can say censorship isn't censorship if you want? This is my point, why are you treating limits placed on your expression/sharing/information differently based on what type of person is doing it?

Because fundamentally it's the same type of censorship as someone deciding to not sell porn magazines, videos or the the Anarchist Cookbook in their newsstand/bookstore/etc. back in the day. They judged (probably quite rightly) that it's not good for business.

Of course the market being extremely concentrated and effectively an oligopoly even in the best case does shine a somewhat different light on it. Until/unless open models catch up both quality and accessibility wise.

Is "Pooh" also censored?

See https://en.wikipedia.org/wiki/Censorship_of_Winnie-the-Pooh_...

Don't forget "Winnie the Pooh"!

They want their LLMs explicitly approved to align with the values of the regime. Not necessarily a bad thing, or at least that avenue wasn't my point. It does get in the way of going fast and breaking things though, and on the other side there is an outright accelerationist pseudo-cult.

Ignoring the moral dimension for a second, I do wonder if it is harder to implement a rather cohesive, but far-reaching censorship in the chinese style, or the more outrage-driven type of "censorship" required of American companies. In the West we have the left pre-occupied with -isms and -phobias, and the right with blasphemy and perceived attacks on their politics.

With the hard shift to the right and Trump coming into office, especially the last bit will be interesting. There is a pretty substantial tension between factual reporting and not offending right-wing ideology: Should a model consider "both sides" about topics with with clear and broad scientific consensus if it might offend Trumpists? (Two examples that come to mind was the recent "The Nazis were actually left wing" and "There are only two genders".)

Because when a small group of elites with permament term and no elections decides what is allowed and what isn't... and has full control of silencing what's not allowed and any meta discussion about the silencing itself... is different from when an elected government decides it, and then anyone is free to raise a stink on whatever is their version of twitter today without worrying about being disappeared tomorrow

It's not an elected government if you're talking about the US. These policies are also all decided by "elites with permanent term and no elections" you realize right?

> It's not an elected government if you're talking about the US

If you don't believe US has elections then straighten up your tinfoil hat:)

Maybe you'll say next the earth is flat, if you think people have nothing better to do but to find ways to lie to you.

I don't feel like this was a good faith interpretation of my comment. What i'm saying is that in the US and China, censorship is decided by unelected officials. In one case it's CPC in another case it's corporate executives

That makes even less sense as a comparison. Sure Instagram censored anti-Trump posts for a day but in case you didn't notice you are free to discuss that without fearing suppression or jail.

Come on man why are you doing this? I didn't say censorship was the same in america or china you're just trying to find something to disagree with

> Why would China censoring Tiananmen Square/whatever out of their LLMs be anymore harmful to the training process when the US controlled LLMs also censor certain topics, eg "how do I make meth?" or "how do I make a nuclear bomb?".

I was explaining why it is more harmful and thought you were arguing it is not harmful?

I was just making a very simple narrow claim: Censorship in the west and china are both done by unelected people. Note that i didn't say china was good, censorship was equivalent or anything else you're trying to argue. my literal only point was:

Censorship in the west and china are both done by unelected people

OK you are technically correct. We can talk how "appointed by elected people" makes a difference...

Because falsifying history seems worse than restricting meth production, at least to me.

Though I see no reason whatsoever why LLM should be blocked from answering "how do I make a nuclear bomb?" query.

Thanks for the detailed breakdown. This is an important nuance to my short reply.

Didn’t it just come out they are losing money on the pro subscriptions?

    > they (and MS) are recouping inference costs from user subscription and API revenue with a healthy operating margin.

I tried to Google for more information. I tried this search: <<is openai inference profitable?>>

I didn't find any reliable sources about OpenAI. All sources that I could find state this is not true -- inference costs are far higher than subscription fees.

I hate to ask this on HN... but, can you provide a source? Or tell us how do you know?

I don't have any qualified source and this metric would be likely be quite confidential even internally.

It is just an educated guess factoring costs of running similar/comparable models to 4o or 4o-mini per token, and how azure commitments work with OpenAI models[2], also knowing that Plus subscriptions are probably more profitable[1] than API calls.

It would be hard for even OpenAI to know with any certainty because they are not paying for Azure credits like a normal company. The costs are deeply intertwined with Azure and would be hard to split given the nature of the MS relationship[3]

----

[1] This is from experience of running LibreChat using 4o versus ChatGPT Plus for ~200 users, subscriptions should quite profitable than raw API by a order of 3 to 4x, of course different types of users and adoption levels will be there my sample while not small is not likely representative of their typical user base.

[2] MS has less incentive to subsidize than say OpenAI themselves

[3] Azure is quite profitable in the aggregate, while possibly subsidizing OpenAI APIs, any such subsidy has not shown up meaningfully in Microsoft financial reports.

Sam Altman posted to Twitter:

    > Insane thing: we are currently losing money on OpenAI pro subscriptions! people use it much more than we expected.

Ref: https://techstartups.com/2025/01/06/openai-is-losing-money-o...

Next he will tell us how open OpenAI is

Are they spending $10B/year on training?

It was my impression that OpenAI was struggling to make money on their $200 pro subscription, because they've underestimate how much people would use it (https://www.theregister.com/2025/01/06/altman_gpt_profits/).

So I do question if OpenAI is even able to make a profit, even if you remove training and R&D. The $20 plan may be more profitable, but now it will need to cover the R&D and training, plus whatever they lose on Pro.

I am paying for o1 Pro but since Deepseek R1 came out I stopped using it. So there goes $200/mo of their revenue ;)

> but they (and MS) are recouping inference costs from user subscription and API revenue with a healthy operating margin.

As far as I am aware the only information from within OpenAI one way or another is from their financial documents circulated to investors:

> The fund-raising material also signaled that OpenAI would need to continue raising money over the next year because its expenses grew in tandem with the number of people using its products.

Subscriptions are the lions share of their revenue (73%). It's possible they are making money on the average Plus or Enterprise subscription but given the above claim they definitely aren't making enough to cover the cost of inference for free users.

https://www.nytimes.com/2024/09/27/technology/openai-chatgpt...

Given the release of the new DeepSeek R1 model [0], OpenAI’s future revenue stream is probably more at risk than it was a week ago.

[0] - https://arstechnica.com/ai/2025/01/china-is-catching-up-with...

OpenAI will not exist in 5 years, I'm calling it now. First movers to market dont always win, and they will surely lose.

Google was first mover.

In what way? They weren't the first search engine, or advertising on the web?

In terms of ai and OpenAI leapfrogged them

Yahoo, AOL, Alta Vista (others too) all were search engines on the web before Google's Sept 1998 existence.

Lycos, Metacrawler, Dogpile. The list goes on

Sure, but we are talking ai and the fact that google was first in this space.

The first in what? Not in search nor Generative AI.

Why would you think search. Google wasn't first for search. They were first for page rank

Google researchers invented the transformer

Who if not Google was the first in generative ai? They invented transformers and diffusion, the cornerstones of text and image generati, respectively.

They weren't the first to meaningfully commercialise either, though. That remains with OpenAI for both (GPT-3/ChatGPT and DALL-E 2).

The question is what's going to be OpenAI's Adwords.

if your birth year starts with 2, I can see why you might think that

Not necessarily. DeepSeek will probably only threaten the API usage of OpenAI, which could also be banned in the US if it's too sucessful. API usage is not a main revenue for OpenAI (it is for Anthropic last time I checked). The main competitor for R1 is o1, which isn't gnerally available yet.

DeepSeek is an open source model. You can download it and run it locally on your laptop already.

So any OpenAI user ( or competitor even) could take it and run a hosted model. You can even tweak the weights if you wanted to.

Why pay for OpenAI access when you can just run your own and save the money?

The one your laptop can run does not rival what OpenAI offers for money. Still, the issue is not whether third party can run it, it's just the OpenAI seems not putting API as their main product.

LM Studio version is here: https://lmstudio.ai/model/deepseek-r1-llama-8b

That's like saying I have a healthy revenue stream from my credit card.

Not quite. In 2 years their revenue has ~20x from 200M ARR to 3.7B ARR. The inference costs I believe pay for themselves (in fact are quite profitable). So what they're putting on their investor's credit cards are the costs of employees & model training. Given it's projected to be a multi-trillion dollar industry and they're seen as a market leader, investors are more than happy to throw in interest free cash flow now in exchange for variable future interest in the form of stocks.

That's not quite the same thing at all as your credit card's revenue stream as you have a ~18%+ monthly interest rate on that revenue stream. If you recall AMZN (& all startups really) have this mode early in their business where they're over-spending on R&D to grow more quickly than their free cash flow otherwise allows to stay ahead of competition and dominate the market. Indeed if investors agree and your business is actually strong, this is a strong play because you're leveraging some future value into today's growth.

All well and good, but how well will it work if the pattern continues that the best open models are less than a year behind what OpenAI is doing?

How long can they maintain their position at the top without the insane cashflow?

One system will be god like and then it doesn't matter

These types of responses always strike me as dogmatic.

Reminds me of the crypto craze where people were claiming that Bitcoin was going to replace all world currencies.

Only the usa Is running a Manhattan project. Nothing to see here really. Go back to bed

Have they built their own ASICs for inference like Google and Microsoft have? Or are they using NVIDIA chips exclusively for inference as well?

The rumors I've heard are that they have a hardware team targeting a 2026 release, but no productions ASICs at the moment.

Platform economics "works" in theory only upto a point. Its super inefficient if you zoom out and look not at system level but ecosystem level. It hasn't lasted long enough to hit failure cases. Just wait a few years.

As to openai, given deepseek and the fact lot of use cases dont even need real time inference its not obvious this story will end well.

I also can't see it ending well for OpenAI. This seems like it's going to be a commodity market with a race to the bottom on pricing. I read that NVIDIA has a roughly 1000% (10x) profit margin on H100's, which means that someone like Google making their own TPUs has a massive cost advantage.

Moore's law seems to be against them too... hardware getting more powerful, small models getting more powerful... Not at all obvious that companies will need to rely on cloud models vs running locally (licencing models from whoever wants that market). Also, a lot of corporate use probably isn't that time critical, and can afford to run slower and cheaper.

Of course the US government could choose to wreck free-market economics by mandating powerful models to be run in "secure" cloud environments, but unless other countries did same that might put US at competitive price disadvantage.

They do get a lot of customers buying their stuff, but on top of that, a company with unique IP and mindshare can get investors to open their wallet easily enough; I keep thinking of AMD that was not or barely profitable for like 15 years in a row.

Serious question - why Texas???

Texas is a world leader in renewable energy. Easy permitting, lots of space, lots of existing grid infrastructure from the o&g industry.

Why do you think datacenters have actually been built in Oregon?

What's the relevance of this article to the comment?

The comment is talking like Texas is a bastion of energy, when they've made the news multiple years in a row about how entire cities collapse over what most states would consider trivial amounts of snow. It is hard to be a world leader in energy and to also have a wiki on your recurring energy crisis issues.

Any downsides?

Texas.

Texas is a world leader in racism, misogyny, homophobia, transphobia, flagrant disrespect for women's autonomy over their own bodies, religious fanaticism, terrible education system, frequent school shootings, armed road ragers, anti-environmental coal rollers, mentally unstable angry gun freaks, presidential assassins who can actually aim well, sunshine towns, white supremacists, violent highly armed far right militias, Confederate Lost Cause rebels, slavery apologists, violent insurrectionists, pardoned violent criminals, mouth breathing MAGA cult members, critical thinking opponents, decrepit electrical power infrastructure, pathological traffic jams, lack of public transportation and bicycle infrastructure, corrupt government, voter suppression, totally unjustified arrogance, enormous explosions from unregulated industries, etc.

https://en.wikipedia.org/wiki/West_Fertilizer_Company_explos...

https://www.texastribune.org/2025/01/21/texans-insurrection-...

What about hurricanes? Extreme weather events which spew a lot of water, wind and debris around might potentially do a lot of damage to a data center.

My kneejerk response was to point to the incoming administration, but the fact Stargate has been in the works for more than a year now says to me it's because of tax credits.

It's where the energy is for this project.

This is unfortunately paywalled but a good writeup on how the datacenter came to be: https://www.theinformation.com/articles/why-openai-and-oracl...

I'm not a subscriber so I can't read it, which startup are they referring to in the headline?

They're referring to Crusoe (crusoe.ai)

A company that will surely still exist in 4 years time.

Lots of back door deals. Just expect more government things put in TX just like the Army built that place in Austin, when we have plenty of dead bases that could be reused

:/

Probably for the same reason that Silcon Valley has been moving there slowly and quietly for a while now.

Because rich people inevitably don't like taxes? And maybe forest fires?

because california is not a very permissive landscape and believes there should be a community-involved process with veto for essentially every large private action

This was a very long winded way to spell money

i’m curious, do you feel that you are contributing to the discussion?

Do you feel your prior comment was sincere and fully captured the reasons for which some firms and individuals have transitioned from California to Texas?

Natural gas to power the turbines while the nuclear plant are built, I guess. Also is Texas more open to large-scale development than elsewhere?

Any downsides?

Existing underinvestment in infrastructure and its maintenance, extreme weather, water resource limitations, some human rights issues.

.

I'm not sure that's how capitalism works.

Who is "we"?

This isn't your money

It is not. But this kind of money does have impact for society in any field. So, this a proper concern.

Meanwhile, Azure has failed to keep up with the last 2-3 generations of both Intel and AMD server processors. They’re available only in “preview” or in a very limited number of regions.

I wonder if this is a sign of the global economic downturn pausing cloud migrations or AI sucking the oxygen out of the room.

global economic downturn? what?

it’s absolutely the second one, this is a commonality across many orgs i’ve talked to who cannot get their CPU request met bc of GPU spend

Isn't it more likely a reaction to xAI now having the most training compute?

How is compute only 50% of their expenses?

I'd guess salaries and such for all the devs and researchers make up a significant portion of the other half

That seems almost impossible.