{"id":801687,"date":"2026-10-13T22:36:10","date_gmt":"2026-10-13T22:36:10","guid":{"rendered":"https:\/\/www.newsbeep.com\/uk\/801687\/"},"modified":"2026-10-13T22:36:10","modified_gmt":"2026-10-13T22:36:10","slug":"annualized-enron-rate","status":"publish","type":"post","link":"https:\/\/www.newsbeep.com\/uk\/801687\/","title":{"rendered":"Annualized Enron Rate"},"content":{"rendered":"

If you liked this piece, you should subscribe to my premium newsletter, and you can subscribe on the following links: $70 a year<\/a>, $18 a quarter<\/a>, or $7 a month<\/a>.<\/p>\n

In return, you get a weekly premium newsletter that dives deep into the finances of the AI bubble. Over the past few months, I\u2019ve published a two<\/a> part<\/a> guide to the world of AI debt, an analysis of the (dubious) impact of generative AI on the global economy<\/a>, deep analyses of the biggest players in the AI infrastructure space like NVIDIA (part one<\/a> and two<\/a>) and Oracle (part one<\/a> and two<\/a>), and the role private credit is playing in funding AI data centers<\/a>, and how private equity, reeling from years of successive failed software acquisitions, is doubling down on the AI bubble<\/a>.\u00a0<\/p>\n

On Friday, we\u2019ll take a look at the JOBS Act \u2014 or, the Obama-era legislation that created the regulatory environment that would ultimately prove essential for the AI bubble to inflate – and why it might be time to reform or entirely repeal it.<\/p>\n

If you want to get in touch \u2014 and especially if you have any juicy information about Anthropic, OpenAI, or any other companies in the AI bubble \u2014 hit me up on Signal at ezitron.76. I\u2019m also on IB on your Bloomberg Terminal.\u00a0<\/p>\n

Soundtrack: Cleopatrick – Hometown<\/a><\/p>\n

“I think the like, state of being an entrepreneur is like, how well do you make decisions in an information-poor environment, right?” – Mark Zuckerberg<\/a><\/p>\n

The entire AI bubble is inflated based on stupid, made up numbers that only exist to pump the bags of investors and make the AI startups look more successful than they are. It is a failure of journalism that they have been popularized, as they do not tell us very much about anything, other than that there\u2019s a new, bigger number to get excited about.<\/p>\n

Last week<\/a>, the markets dropped on the news that OpenAI\u2019s annualized recurring revenue (ARR) \u2014 which was reported as $70 billion<\/a> by Axios and confirmed by the Financial Times \u2014 was actually $50 billion<\/a>. The number in question was, according to the FT<\/a>, inflated by \u201cgrossing up\u201d the total to include sales through cloud partners, a discrepancy between how OpenAI and Anthropic make up a number that doesn\u2019t really tell anyone anything:<\/p>\n

Anthropic recognises gross revenue on sales through cloud partners. That is because Anthropic is the principal in the transaction, according to a person with knowledge of its accounting practices. The Claude maker\u2019s revenue run rate was calculated by annualising the prior 28 days of consumption-based revenue, plus recurring monthly subscription revenue, they added.<\/p>\n

OpenAI, meanwhile, counts only its own cut of proceeds from arrangements with partners such as Microsoft.<\/p>\n

The $70bn figure reported in September arose from attempts by investors to compare Anthropic and OpenAI performance on a like-for-like basis, according to a person close to OpenAI. However, OpenAI described the $70bn as a \u201cnet\u201d revenue metric when it was approached by the FT at that time.<\/p>\n

Back in August, Anthropic had told investors that it had hit $65 billion in annualized run rate<\/a>, but at no point defined what that meant, what period it referred to, what any of the underlying numbers are, or what \u201cgross\u201d versus \u201cnet\u201d bollocks it was playing. Per Irrational Analysis the calculus might have been even sillier than that<\/a>:<\/p>\n

Last month, there was a whole kerfuffel in AI\/semis\/finance circles on Anthropic July ARR. Two numbers were going around. I don\u2019t remember the numbers and frankly it does not matter. You will see.<\/p>\n

One ARR number was the traditional \u201ctrailing 28 days * 13\u201d number. Personally I hate this venture-capital clown metric but whatever a lot of people use this.<\/p>\n

The traditional ARR number was bad and implied deceleration in growth. So the people massively long Anthropic came up with a new ARR number that was July 31st * 365 days.<\/p>\n

This is the problem of saying what the revenues of a company might be based on numbers that are not its actual revenues. There is absolutely no value in using \u201cannualized run rates\u201d for anyone other than the company and its investors, and the fact that modern tech and business journalism continues to accept it for what are meant to be the largest and most-important companies in the world is a sign that something is very wrong with how we cover the tech industry.<\/p>\n

Sidenote: When one \u201cannualizes\u201d a period of revenue, the intention is usually to take a time period that\u2019s indicative of the consistent performance of the business and extrapolate from there what it might make on an (as you\u2019ve guessed) annual basis. This calculation is useful only when connected to specific periods and forms of revenue \u2014 recurring subscriptions and consistent contracts rather than, in the case of the AI bubble, consumption-based revenues that are not \u2018recurring\u2019 and thus mean very little when annualized.<\/p>\n

The other problem is that these run rates do not measure if revenues go up or down, because you\u2019re choosing the best period possible.<\/p>\n

I am not going to spend much time dwelling on gross versus net. It appears that, per Reuters<\/a>, 47% of Anthropic\u2019s 2025 revenues came from sales made by Google and Amazon, who take a 16% cut<\/a> as opposed to the roughly 20% that Microsoft takes from OpenAI. It appears that Anthropic for whatever reason counts every dollar of sales as part of its annualized run rate figure and OpenAI does not, for reasons that escape me, as I simply don\u2019t buy that it is \u201cmore ethical.\u201d\u00a0<\/p>\n

My evidence? The following quote from the FT, emphasis mine:<\/p>\n

The $70bn figure reported in September arose from attempts by investors to compare Anthropic and OpenAI performance on a like-for-like basis, according to a person close to OpenAI. However, OpenAI described the $70bn as a \u201cnet\u201d revenue metric when it was approached by the FT at that time.<\/p>\n

This is the same OpenAI that tried to claim that a $34 billion loss was actually an $8 billion loss when the FT asked<\/a> them about my reporting on its audited financials<\/a>.<\/p>\n

I must also say how utterly disappointed I was to see how Bloomberg credulously reported<\/a> that OpenAI \u201cexpected to hit $70 billion in annualized revenue by end of 2026,\u201d a transparent attempt to send the markets back up<\/a> with a story that tells the world less than nothing.\u00a0<\/p>\n

Wow, OpenAI, by the end of 2026, will increase a number that \u201crepresents a projection of [its] yearly sales based on a shorter period\u201d by $20 billion! What is the period in question? Who knows! How is that number derived? Who cares! Number go up! Put the story in the bag! Your job is not to cover the tech industry, it is to celebrate it! Stop thinking!<\/p>\n

Sidenote: another very obvious reason these companies use these numbers is SEO! If you try to Google OpenAI\u2019s projected revenues, you\u2019re guaranteed to get one of these stupid annualized figures. In reality, the company projects to make somewhere in the region of $30 billion in 2026<\/a>.\u00a0<\/p>\n

In other words, OpenAI\u2019s made up number was actually lower than Anthropic\u2019s made up number because of the way it made the number up, but by the end of the year, the made up number would be bigger than the other made up number. These made up numbers are reported by the largest and most well-respected news outlets of the world, who lend said made up numbers respectability despite the numbers in question being utter bullshit.<\/p>\n

I\u2019ll give you an example.<\/p>\n

For the first half of 2026, OpenAI made $12.4 billion in ($5.7 billion in Q1 and $6.7 billion in Q2) revenue per The Information<\/a> and the Wall Street Journal<\/a>, while Anthropic, per the WSJ<\/a> and the FT<\/a>, made $15.5 billion ($4.6 billion in Q1 and $10.9 billion in Q2). We do not have consistent coverage of their underlying costs, but in Q1 (per The Information), OpenAI spent $8.6 billion on training and $3.5 billion on \u201ccost of revenue,\u201d with no reporting on its sales and marketing spend where it allocates the cost of serving free users or discounts.\u00a0<\/p>\n

Anthropic said it had hit an annualized run rate (undefined period multiplied by undefined number) of $48 billion on May 28<\/a>, or roughly 430% more than its entire revenue in the second quarter. In fact, if you added up every dollar that both Anthropic and OpenAI made in the first two quarters, you\u2019d be $19.1 billion short of Anthropic\u2019s May \u201crun rate.\u201d<\/p>\n

We do not know how any of these numbers were actually conceived outside of they are somehow derived from a non-specific 28 day period. The information they convey is sparse, if it communicates anything at all.\u00a0<\/p>\n

Reporting \u201crun rates\u201d is a marketing exercise, and if you disagree, please answer me the following: what does a \u201c$65 billion annualized run rate\u201d tell you about Anthropic, exactly? It\u2019s MEANT to make you think it will make $65 billion a year in revenue, but that is not what it actually means. This number tells you nothing, and exists only to poison the well of information.\u00a0<\/p>\n

Without an actual period of time \u2014 and I mean an exact one, such as \u201cJuly 1 to July 31, 2026\u201d \u2014 and the exact way it was calculated, this number effectively means nothing, and reporting it without this information only serves OpenAI or Anthropic. You can say \u201cinvestors use it all the time\u201d all you want. Investors are fucking stupid for using it. If the argument is \u201cit shows how fast OpenAI is growing,\u201d why don\u2019t you find out the actual growth rate of the company?\u00a0<\/p>\n

The fact these numbers moved markets is a reflection of the definitive \u2014 and accepted \u2014 ignorance built into the entire AI bubble, where despite trillions of dollars on the line, we know very little about the actual underlying numbers, and those numbers we do know are inflated using techniques that would get you shot by an accountant.<\/p>\n

In fact, we don\u2019t know most of the numbers that would actually tell us whether any of this is a good idea.<\/p>\n

Reporting on the true revenues of the AI industry feels less like financial analysis and more like legal discovery, and it\u2019s fundamentally insane how little tangible, meaningful information that exists to feed the trillion-plus dollars in capital expenditures and endless full-court press in the media about how AI is the next big thing.\u00a0<\/p>\n

We do not know Google, Amazon or Oracle\u2019s AI revenues, and we have a single year \u2014 fiscal year 2026 \u2014 of Microsoft\u2019s AI revenues based on an estimate from Bloomberg<\/a> calculated from its disclosure of its OpenAI revenues in an SEC filing<\/a>. We do not know how much each individual service makes, how much their true revenue shares are with the AI labs or, as I discussed a few weeks ago, how much actual, operational AI capacity exists as a result of the trillion-plus dollars in capital expenditures<\/a>, though Bloomberg and I have both reported that Microsoft has around 2GW of it, even though it claimed to bring online a gigawatt in each of its last three fiscal quarters.\u00a0<\/p>\n

We do not know the revenues or much about the economics of [include funding totals) Perplexity, Harvey, Instinct, Reflection, Cognition, Replit, Lovable, and Factory, all of whom have raised somewhere in the region of $15.8 billion in funding, and have a combined valuation of $215.8 billion. We do not and will never know the true revenues or costs of Cursor\u2019s operations as a result of its acquisition by SpaceX<\/a>, and as a reminder, annualized revenues are not actual indicators of the economic health \u2014 the revenues, costs or cash flows \u2014 of any of these companies.\u00a0<\/p>\n

Perhaps this was permissible in the past because so much less capital was being annihilated in the name of startups, but things have escalated so much further than any other era, including the Dot Com bubble.<\/p>\n

Look, the entire AI bubble has been inflated based on the implicit belief that there will be demand \u2014 from AI startups and elsewhere \u2014 for AI compute, to the point that Goldman Sachs says that hyperscalers need AI startups to make $425 billion in annual revenues just for the hyperscalers to break even<\/a>\u2026and we don\u2019t even know, outside of Anthropic and OpenAI, how much any AI startup actually makes.<\/p>\n

\"\"<\/p>\n

What we do have is \u201cannualized run rates\u201d pegged to non-specific time periods and non-specific numbers that tell us, for example, that Cognition hit $1 billion<\/a> in late September, a number that means next to nothing. Even when The Information<\/a> was able to report that Cognition would burn $800 million in 2026, it resorted to saying that it was \u201cgenerating around $900 million in annualized revenue, or $75 million a month\u201d as of August 27 2026 \u2014 adding that executives had \u201cprojected it would hit $1.5 billion in annualized revenue by end of year,\u201d meaning that two journalists were able to successfully find out its actual expected burn, that it burned $200 million in Q2 2026\u2026but not the actual revenues that the company generated on a quarterly basis.<\/p>\n

Let\u2019s be clear about something: the reason we\u2019re not getting these numbers is that these companies are not actually making these amounts on an annual basis. If Cognition was actually going to hit anything close to $1 billion \u2014 and let\u2019s be honest, it\u2019s likely less than $500 million! \u2014 it\u2019d leak it versus using these weird monthly snapshots.<\/p>\n

It is astonishing that we, four years into the AI bubble, do not know the revenues of more than a handful of AI companies outside of OpenAI and Anthropic. All we have is a mishmash of annualized run rates, and at this point I have to wonder if it\u2019s a failure of journalism, successful opsec, even more successful marketing, or that these companies don\u2019t have great accounting.\u00a0<\/p>\n

But really, go and look up basically every AI company. Perplexity? Nothing. Harvey? Nothing. Cursor? Years of reporting \u2014 endless thousands of words spent \u2014 and nary one report I can find, outside of these fucking annualized run rates.\u00a0<\/p>\n

Part of the reason, I imagine, is this chart from Ramp<\/a> that I\u2019ve brought up before:<\/p>\n

\"\"<\/p>\n

From what I hear, the vast majority of those 1% of customers are AI startups that power their products using OpenAI and Anthropic\u2019s models, predominantly using monthly subscriptions that allow users to burn way more in tokens than their subscriptions actually make.<\/p>\n

For example, SemiAnalysis did a deep study<\/a> into the amount of tokens you could burn on the various different AI subscriptions out there, finding that you could burn anywhere from $83 to $118 a month for $20 and from $1,264 to $1,331 a month for $200 on Cognition, dropping to $46 on a $20-a-month plan and $507 for $200-a-month on Cursor when using Anthropic\u2019s models.<\/p>\n

\"\"<\/p>\n

While OpenAI\u2019s models are being dropped from Cursor, SemiAnalysis found you could burn $154 a month using OpenAI\u2019s models on Cognition\u2019s $20-a-month plan, and $1,736 a month on its $200-a-month plan.<\/p>\n

The various boosters reading this and making noises like the Skesis from Dark Crystal<\/a> will say \u201cbut these companies get discounts!\u201d as if that makes up for the fact that customers are able to burn two to six times their actual subscription every month. They\u2019ll also argue that \u201cthis is the gym model, and most customers don\u2019t spend that much,\u201d a statement that they have no proof for other than that it makes them feel better.\u00a0<\/p>\n

The problem \u2014 and the reason I\u2019d argue that the actual revenues of these companies are so shrouded in secrecy \u2014 is that the underlying economics do not match up when you consider that effectively every AI startup has to resell tokens at a discount. It doesn\u2019t really matter how much revenue you make if you\u2019re sending twice or thrice or five times that amount out the door to Anthropic and OpenAI, but saying a larger \u201cannualized run rate\u201d makes the amounts you\u2019re raising far less worrying than what I imagine are particularly-shrimpy sales.<\/p>\n

AI Startups Are An Extraction Machine That Turns Venture Capital Dollars Into Anthropic and OpenAI\u2019s Revenues<\/p>\n

Let\u2019s take Harvey \u2014 which raised $550 million at a $15.5 billion valuation in September<\/a> \u2014 as an example. As of early August, Harvey had hit $350 million in annualized revenue<\/a>, up from $190 million in January. If we assume it uses the same \u201ctrailing 28 days\u201d model as Anthropic, this means that it made $26.9 million in a 28-day period ending sometime in August, up from $14.6 million in a period that ended some time in January 2026<\/a>.<\/p>\n

Since December 2025, Harvey has raised a total of $910 million across<\/a> three<\/a> rounds<\/a>, or roughly 33.8 times the estimated underlying revenues of its most-recent run rate, which increased by 45.7% in that period.<\/p>\n

Bloomberg claimed in late September<\/a> that Harvey\u2019s Gross Margins had gone from 50% in January 2026 to negative 50% in June 2026 to \u201cpositive again\u201d (I imagine in September) due to using open source models, but something about those numbers is deeply weird to me, mostly because I\u2019ve never, ever heard of a company having had that happen.\u00a0<\/p>\n

And none of that explains why a company that makes tens of millions of dollars in revenue needs to raise hundreds of millions of dollars every few months, unless, of course, most of those dollars are flowing immediately to Anthropic and OpenAI, which is, it turns out, exactly what happened:<\/p>\n

The $15.6 billion legal startup Harvey built its business around training AI models like OpenAI\u2019s GPT-4 to do specialized work for lawyers. Recently, however, soaring artificial intelligence costs have nudged it to rethink its dependence on the AI giants.<\/p>\n

After a March update to its AI agents, Harvey\u2019s customer usage spiked, but its gross margins dropped sharply \u2014 from about 50% at the beginning of the year to -50% by June, according to a person familiar with the matter.<\/p>\n

This is the antithesis of any good business. In basically any industry, the more demand you have, the more money you make, but in AI, the more demand you get, the more money you lose, and your most engaged customers are also your least-profitable.\u00a0<\/p>\n

However you might rationalize the underlying economics of a company like Harvey, the numbers we do know are extremely worrying \u2014 $900 million raised since December 2025, and, in my most-optimistic calculation, around $175 million (though likely less than $150 million) in revenue through the end of August, for a company that raised another $600 million across<\/a> two<\/a> other rounds in 2025. That\u2019s $1.66 billion in about a year and a half to make an unprofitable company with less than $30 million in monthly revenue that is somehow worth $15.5 billion.<\/p>\n

Harvey has categorically failed to find product-market fit, unless we qualify that as \u201cliterally finding a customer.\u201d It\u2019s very clear that its underlying product cannot support itself, and that the company is entirely-dependent on venture capital, raising at least $160 million every three to six months since February 2025. It is abundantly clear that Harvey\u2019s customers \u2014 like every AI startup \u2014 do not actually want to pay what its services cost, which means it does not have product-market fit, nor what I\u2019d call a \u201cbusiness.\u201d<\/p>\n

Then there\u2019s Cognition, which has raised a combined $3 billion in 2026 alone (in May<\/a> and September<\/a>), with its run rate going from $492 million to $1 billion in that period. If we assume the same calculation as Anthropic uses, that means revenues were $37.8 million and $76.9 million respectively, and that its last two rounds of funding are roughly 39 times the size of its 28-day trailing revenues. It is unclear what has substantively changed about the company, other than that the $400 million it raised in September 2025<\/a> was insufficient to last.<\/p>\n

I know there are going to be venture capitalists who argue that \u201cthese companies are just raising opportunistically,\u201d but that doesn\u2019t explain why every AI startup raises so much money in so little time, and how one reconciles that distance with any reality outside of \u201cwe can raise unlimited money forever.\u201d\u00a0<\/p>\n

And if they can\u2019t do that, large chunks of OpenAI and Anthropic\u2019s revenues disappear.\u00a0<\/p>\n

Let me summarize all of this in simple terms, as it\u2019s all terribly important.<\/p>\n

The Revenues80% of Anthropic and OpenAI\u2019s revenues come from 1% of their customers, the vast majority of which are AI startups that spend anywhere from 300% to 1000% of their revenue on AI tokens.AI Startups Run Services at a Loss, Allowing Users to Burn Thousands of Dollars of Tokens for Hundreds of Dollars a Month, Requiring Constant Venture FundingCustomers of AI startups are repulsed by the actual cost of tokens, meaning that for every dollar of subscription revenue, AI startups must allow customers to burn anywhere from 2 times to 10 times their revenue in tokens. For example, Cognition\u2019s $200-a-month users can spend upwards of $1200 a month in OpenAI or Anthropic tokens.\u00a0Even with these \u201cdiscounted\u201d tokens, AI startups have very small revenues.As a result, every AI startup runs at a loss. This necessitates raising massive amounts of capital every few months.Every New Model Release Artificially Boosts Revenues Through AI StartupsWhen a new model releases, every AI startup must immediately integrate it, because that\u2019s what their customers expect, and because the customers in question are curious, they\u2019re likely to use the new models more on release.This means that Anthropic and OpenAI get a massive bump of new revenues \u2014 funded by AI startups that are allowing users to burn tokens without restraint \u2014 with every release.All Of This Is Only Made Possible By Perpetual Venture Funding, And Once It Expires, It Eats Anthropic and OpenAI\u2019s Revenues AliveAll it takes for one of these companies to stop being able to spend tens or hundreds of millions of dollars a month on OpenAI and Anthropic\u2019s services to meaningfully reduce their revenue.It doesn\u2019t really matter what big businesses are or are not using Claude or ChatGPT \u2014 the underlying revenues of these companies are held up by venture capital by proxy.<\/p>\n

Every AI startup is effectively an extraction mechanism for OpenAI and Anthropic, reselling their tokens through their services at a massive discount because users are unwilling to pay the actual cost of using LLMs.\u00a0<\/p>\n

Using annualized run rates allows everyone to feel like this isn\u2019t a problem, because it\u2019s far easier to rationalize that Cognition has raised $3 billion this year and is worth $48 billion if it has $1 billion in annualized run rate rather than $80 million a month in revenue.\u00a0<\/p>\n

The obfuscation of the actual underlying financial of these companies is intentional, because it makes it harder to do the kinds of comparisons I\u2019ve done today,\u00a0<\/p>\n

If you disagree, start giving me actual revenue numbers for these companies instead of your magic elf number you made up or got told. If you\u2019re an investor in these companies and have not been told the actual most recent quarterly revenue, there\u2019s a reason \u2014 the real one sucks!\u00a0<\/p>\n

If you know it, leak it if it\u2019s so impressive. You know where to find me. Only one of us is being weird, and it\u2019s the cloister of you who condone using bad information because you know the real number \u2014 the actual quarterly or annual revenue \u2014 is probably going to make people panic.\u00a0<\/p>\n

These companies are also, at least in theory, the ones that pay for AI compute, which means that we also have a very, very poor understanding of the actual customers behind those hundreds of billions of NVIDIA GPUs.<\/p>\n

In fact, that\u2019s a great place to go next.<\/p>\n

People are going to get mad at me for this because they are cowards, but I posit that we have little understanding of the true price that people are paying for AI GPUs, or about the underlying economics of AI data centers writ large.<\/p>\n

Yes, yes, I know, you\u2019re going to tell me that SiliconData has an index<\/a>, and that that index shows \u201cGPU rental prices going up,\u201d but these prices overwhelmingly favor spot pricing \u2014 as in what it costs to rent a GPU right now \u2014 versus what companies are paying on a long term basis.<\/p>\n

It\u2019s kind of like valuing a shipping company based on skyrocketing spot rates during a capacity shortage rather than the year (or years)-long rates that they regularly charge. You know, kind of like annualizing the data to make things look better.<\/p>\n

SiliconData claims to create a \u201csingle standardized reading,\u201d but does so using a mishmash of non-specific data that \u201cstandardizes for rental term length, cluster scale and interconnect\u201d in a way that only ever seems to benefit the providers.\u00a0<\/p>\n

The reason that SiliconData\u2019s index has become so popular is that we otherwise lack many good measurements, which isn\u2019t to say that its measurements are good, just that they\u2019re the best we\u2019ve got at the moment. While I think the company is doing its best with what it\u2019s got, measuring GPU rates in general like this only seeks to confuse the market through abrupt shifts upwards and downwards based on what I imagine are larger contracts ending or beginning \u2014 not that we know, because SiliconData does not actually explain what\u2019s changed or why it might have changed.<\/p>\n

This is a huge issue, because NVIDIA has sold $535 billion in GPUs since 2022, and we still don\u2019t know on any reliable basis how much their customers can rent them for. We don\u2019t know with much certainty how much it costs to run one, or have a firm grasp on what an hourly comparison would be \u2014 as in how many dollars an hour it cost to run one \u2014 outside of analyst estimates (and my own model<\/a>) that all seem to vary wildly and always end up at a place that says \u201cwhat a great investment!\u201d<\/p>\n

There\u2019s also some compelling evidence that the actual long-term rental rates for these things might not be as high as we think.<\/p>\n

I realize some people might read what I\u2019m saying and immediately want to pelt me with rocks, but I\u2019m just telling you what I\u2019ve found, and if I\u2019m honest, I think I\u2019m doing a better job than the vast majority of people covering this shit, in that I\u2019m trying to actually show what people are paying versus aggregating various publicly-available sources of information to simmer out info-slop.<\/p>\n

Again and again, I\u2019ve found companies paying way, way less than the supposed going rate of compute.<\/p>\n

Krea \u2014 B200s at $2.80-per-hour-per-GPU, vs. SiliconData rate of $5.94-per-hour<\/p>\n

In late April, The Information ran a story<\/a> about startups renting AI GPUs, specifically focusing on NVIDIA\u2019s B200s, which SiliconData claims rent at $5.94 an hour, emphasis mine:<\/p>\n

Six months ago, Krea signed a six-month contract to rent several hundred Nvidia Blackwell chips for $2.80 an hour per chip after several cloud providers competed aggressively for the startup\u2019s business. But in the last month, when the startup began looking for more AI servers to train a model from scratch, some sales representatives at cloud providers wouldn\u2019t pick up the phone, said co-founder and CEO Victor Perez.<\/p>\n

[a few months later]\u2026In the end, the founders reached an agreement to pay $3.70 an hour for another several hundred Blackwell chips in a one-year commitment\u201432% higher than their last deal. And even that price seemed cheap, given the chip prices he had seen from other providers.<\/p>\n

To be clear, the B200 GPU, as part of an 8-GPU pod, works out to about $62,500 a GPU. At this rate, 32% of the first six years of its useful life would be absorbed by depreciation, assuming that it was sold in its entirety for every single hour of every single day of those years.\u00a0<\/p>\n

OpenAI \u2014 Stargate Abilene \u2014 GB200s at $2.54-per-hour-per-GPU, vs. SiliconData rate of $5.94-per-GPU-per-hour (for B200)<\/p>\n

Then there\u2019s Oracle\u2019s under-construction \u201cStargate Abilene\u201d data center that landowner Lancium believes will make around $10 billion in annual revenue<\/a> (see slide 5). At $10 billion a year and 450,000 GB200 GPUs, which works out to \u2014 if we assume a take-or-pay rate \u2014 to about $2.54-per-hour-per-GPU. If we assume that each of these is inside a GB200 NVL72 at about $3.5 million a piece, that\u2019s about $1.61 per-GPU-per-hour of revenue for the first six years before you get to the COGS.<\/p>\n

Sharon AI\/ESDS Cloud \u2014 B300s at $2.80-per-hour-per-GPU, vs. SiliconData rate of $6.65-per-GPU-per-hour<\/p>\n

Then there\u2019s this contract between Australian neocloud Sharon AI and ESDS Cloud FZ-LLC<\/a>, a Dubai-based subsidiary of Indian enterprise cloud company ESDS<\/a>, renting out 8208 NVIDIA B300 GPUs at $3.30 a month over 60 months (five years). If we assume a cost of $540,000 per eight-GPU B300 cluster, that leaves us with about $1.54 in depreciation over six years, leaving us with about $1.869-per-hour-per-GPU of actual revenue for the first six years, in a contract that is five years long.<\/p>\n

Canopy Wave \u2014 B300s at $4.3-per-hour-per-GPU for three years, $3.50-per-hour-per-GPU for two years (total five) vs. SiliconData rate of $6.65-per-GPU-per-hour (Average = $3.98-per-GPU-per-hour)<\/p>\n

Here\u2019s another. A contract I found<\/a> between neocloud Kidz AI (which rents capacity from a company called Limestone<\/a>, or, more-specifically, its subsidiary Catalyst Compute) and AI inference company Canopy Wave priced 256 B300 GPUs at $4.30 an hour (per GPU) for the first three years and $3.50 an hour for the last two of the five-year-long contract.<\/p>\n

Your depreciation costs here work out to about $1.54-per-GPU-per-hour, meaning that you\u2019re losing nearly half of your revenue for the first six years\u2026of a five year long contract.<\/p>\n

Sure hope there aren\u2019t new kinds of GPUs \u2014 or an abundance of older GPUs \u2014 after these things end!<\/p>\n

QumulusAI \u2014 B300s at $3.70-per-GPU-per-hour for three years vs. SiliconData rate of $6.65-per-GPU-per-hour\u00a0<\/p>\n

QumulusAI is a former cryptocurrency mining company that converted into renting GPUs. In this contract<\/a>, Qumulus is renting out 128 B300 nodes \u2014 so around 1,024 B300 GPUs \u2014 at $3.70-per-GPU-per-hour, sitting at about $1.284-per-GPU-per-hour of depreciation.<\/p>\n

Again, sure hope there aren\u2019t a bunch of data centers under construction that have the same chips as these!<\/p>\n

WhiteFiber for BaseTen \u2014 B300s at $4.516-per-GPU-per-hour for three years vs. SiliconData rate of $6.65-per-GPU-per-hour<\/p>\n

Buried deep in the earnings report of AI infrastructure company WhiteFiber<\/a> sits a three-year-long contract with buzzy AI inference firm Baseten, which raised $1.5 billion in June<\/a>:<\/p>\n

In August 2026, we entered into a service order with a new customer, BaseTen Labs, Inc., to provide services utilizing a total of 1,392 B300 GPUs. The service order has an initial term of 36 months beginning on the services commencement date, after which it automatically renews for successive one-month periods unless terminated by either party. The service order represents an aggregate revenue opportunity of approximately $165.2 million. The deployment and revenue generation is scheduled to begin in November 2026.<\/p>\n

This all works out to about $1.284-per-GPU-per-hour of depreciation on $4.51-per-GPU-per-hour, which is still more than two dollars less than what SiliconData quotes.<\/p>\n

As mentioned, BaseTen just raised a ton of money, so it\u2019s willing to pay through the nose, and even then it\u2019s paying so much less than spot.<\/p>\n

There\u2019s also another contract with a customer in Iceland that works out to about $3.46-per-GPU-per-hour:<\/p>\n

In August 2026, we entered into a new five-year service order with an existing customer to provide services utilizing a total of 576 Nvidia B300s GPUs in Iceland. The service order represents an aggregate revenue opportunity of approximately $87.5 million over its initial term, with additional potential upside through revenue sharing. Service under this agreement is targeted to commence in December 2026.We Don\u2019t Even Know How Many AI Data Centers Exist<\/p>\n

As I said a few weeks ago<\/a>, we do not really have any firm understanding of how much AI compute actually exists, to the point that the largest companies in the world (read: Microsoft) are actively misleading the public<\/a> about how much they\u2019re bringing online.<\/p>\n

We also have no clear understanding of how much data center capacity is actually under construction, because nobody actually wants to tell us, and reports like this from Sightline Climate<\/a> claiming 190GW \u201cin planning\u201d lump in every imaginable kind of data center with ones specializing in AI compute.<\/p>\n

As a result, basically any AI compute capacity estimates are guesses. I\u2019ve seen somebody say there\u2019s \u201c50GW of capacity,\u201d I\u2019ve got guys in my mentions saying that actually Bloomberg is wrong in saying that Microsoft has 2GW of capacity<\/a> and the real number is 5GW based on nebulous \u201cconversations\u201d that only ever end up as a positive for the AI industry.\u00a0<\/p>\n

In a vacuum of good information, people simply start making things up, which is why The Information published one of the more insane things I\u2019ve read in a while<\/a>, emphasis mine:<\/p>\n

In a time of server scarcity, speed matters. The Louisiana campus plans to consume 330 MW of power to produce 245 MW in usable compute capacity. If that capacity can get online even one month faster than its current schedule calls for, Anthropic could generate an extra $1.3 billion or more in revenue, based on investors\u2019 and analysts\u2019 estimates. That assumes it uses all the capacity to serve up answers rather than develop new models.<\/p>\n

I asked Okoye for an explanation, and he gave me one in the comments of the article:<\/p>\n

Facility rent: Hut 8\u2019s lease for this 245 MW site is $7.0B over 15 years. That is about $39M a month.Compute: IREN\u2019s 5-year, $9.7B contract with Microsoft covers about 200 MW of GB300 GPUs. That is about $9.7M perMW-year, or about $200M a month at 245 MW. It\u2019s a GPU contract, so treat it as a benchmark for this TPU site, not its actual price.Lab revenue: the $1.3B is the midpoint of the $50M to $75M per MW-year estimates Ann cites [referring to a comment from the author Ann Davis Vaughan].[Ed note: Those numbers that she cites come from another article<\/a> where Ric Boada, the CEO of ten-month-old neocloud Volta, saying:]He figures that the frontier AI labs are generating about $75 million in revenue per megawatt of use a year for inference output compared to an annualized cost to build and provide the infrastructure of $10 million to $15 million per megawatt a year.\u00a0Each layer is paid out of the one above it. Even at the low end, about $1.0B a month, lab revenue is roughly 5x the compute bill. That gap is why labs push datacenter developers so hard on timelines.<\/p>\n

All of this hinges on some very wishy-washy numbers about the per-megawatt revenue of AI labs, which in turn hinge on wishy-washy numbers about how much compute labs have. For example, the New York Times reported that Anthropic would have 5GW of capacity by the end of 2026, up from 1.5GW at the end of 2025<\/a>, according to statements made to its investors, but I can find no compelling evidence that there was even a gigawatt of compute for it to buy.\u00a0<\/p>\n

If we assume these statements are accurate, then:<\/p>\n

Anthropic, which made $4.6 billion in revenue against (theoretically) 1.5GW of capacity, made about $3 million a megawatt in 2025.In 2026, assuming it makes $35 billion on 5GW of capacity, it would make $7 million a megawatt.<\/p>\n

The peanut gallery will scream and oink at me for not working out how much is inference versus training, to which I say nobody fucking knows how much of its capacity is actually diverted to inference and I don\u2019t even think they know how much capacity Anthropic has writ large!<\/p>\n

We are basing the vast majority of words written and billions invested in the AI bubble on assumptions built on assumptions built on assumptions, several of which fall apart when you ask simple questions like \u201chow does that work?\u201d or \u201chow did you get that number?\u201d\u00a0<\/p>\n

This would all be significantly less-worrying if it wasn\u2019t for the fact that everything is built on the idea that AI is going to be the panacea for the entire tech industry!\u00a0<\/p>\n

Put another way, NVIDIA plans to sell $670 billion or more GPUs next year<\/a> on top of the more than $500 billion it\u2019s already sold and $800 billion has been invested in AI startups mostly by people with little or no idea about the underlying economics of the data centers or their customers!<\/p>\n

A few months ago, a dear friend of mine told me that \u201cnobody knows what they\u2019re investing in,\u201d and it\u2019s taken me a while to fully understand what he meant.<\/p>\n

The entire world believes we\u2019re in an AI-powered economic boom of data center capex and AI startup growth based on flimsy information sources that vary from fanciful to utterly hysterical, constantly guessing about how much capacity OpenAI or Anthropic or any of their hyperscale sugar daddies actually have operational, or how much a theoretical data center could charge under the best of circumstances to a customer that may or may not actually exist.<\/p>\n

We are four fucking years into the AI bubble, have little or no understanding of the actual economics of AI, and that which we do have is extremely grim. Yet when you try and tell an AI booster (or many members of the media!) that Anthropic lost $8 billion<\/a> and OpenAI lost $20.97 billion<\/a> in 2025, their immediate response is to say that 2026 will be fine because of quarters with discounted compute costs<\/a> and adjusted operating profits<\/a> that don\u2019t actually resolve the underlying problem that these companies have $1.3 trillion in compute commitments, with at least $413 billion of them being uncancellable<\/a>.<\/p>\n

The hysteria has grown so significant that Gavin Baker of investment firm Atreides (invested in both Anthropic and SpaceX) has simply chosen to suggest that \u201cnobody is taking them seriously<\/a>,\u201d only to disappear into the ether when asked<\/a> about which of the contracts he believes are \u201cnot taken seriously\u201d and \u201cby whom,\u201d probably because he\u2019s talking out of his ass.\u00a0<\/p>\n

This argument is one you make when you are flat out of ideas, and simply hope that people will leave you alone if you sound authoritative enough, and is known in debate circles as the \u201cnuh uh\u201d technique. The reason Baker is deploying it is that he, like every AI booster, knows that while you can fool the press and investors with made up numbers, there\u2019s very little that \u201cannualized run rates\u201d or \u201cAnthropic ending the year with $100 billion in revenue<\/a>\u201d can do when Broadcom CEO Hock Tan, who recently likened AI investors selling their investments on the news of OpenAI’s adjusted financials to degenerate gamblers<\/a>, asks Dario Amodei to pay some of the $161.2 billion in non-cancellable contracts he signed.<\/p>\n

We live in a climate of intense hysteria and profound confusion, made worse by a media ecosystem that currently appears more interested in sustaining narratives than investigating their underlying truths. The vast majority of hyperscaler AI revenues come from Anthropic and OpenAI\u2019s compute spend<\/a> that is dependent on near-perpetual funding, much like the vast majority of Anthropic and OpenAI\u2019s revenues come from AI startups that depend on near-perpetual funding to continue their spend.\u00a0<\/p>\n

And to make matters worse, we quite literally have no idea about the underlying financial condition of any other AI companies outside of the main AI labs, and no, annualized run rates don\u2019t mean jack shit.\u00a0<\/p>\n

This is extremely bad and extremely important to resolve. Anthropic and OpenAI\u2019s growth depends on the AI startup ecosystem being able to spend increasing amounts on tokens every single month, without fail, for what appears to be the rest of time. Said startups\u2019 ability to keep doing so depends on venture capital being willing to sling them hundreds of millions of dollars a year each, with each investment increasing the paper value of the company but adding little or no actual qualitative improvement other than their ability to keep playing at the Sam Altman Casino.<\/p>\n

I genuinely don\u2019t know why most of the tech and business media is ignoring this problem. Is it that it\u2019s too difficult to comprehend? Is it that you, as a writer, truly think that all of this will work out? Is it that you think that Anthropic and OpenAI will just \u201cnegotiate out\u201d of their contracts? If that\u2019s the case, what happens to all the data center capacity? Is it that there\u2019s advertiser or editorial pressure to be \u201cpositive on AI\u201d? I truly want to know!\u00a0<\/p>\n

Anyway, I\u2019ve realized now that \u201cnobody knows what they\u2019re investing in\u201d applies to basically everything. Most people writing about AI believe it\u2019ll all work out because nothing bad has happened or they have a single statistic that affirms their worldview. Most people are still saying \u201cit\u2019s like Amazon Web Services!\u201d or \u201cit\u2019s like Uber!\u201d or \u201cit\u2019ll be like the Dot Com Bubble,\u201d even though those are not correct or sensible or rational takes that actually mean much other than \u201cmoney burn good actually,\u201d and that thin level of rationality goes up to the highest rungs of finance and journalism to a degree that turns my fucking stomach.<\/p>\n

Most coverage of AI lives in its own bubble where nothing bad has happened or could happen, and that every promise will come true. Every deranged, multi-faceted \u201cmodel\u201d that shows OpenAI or Anthropic will print money in the future is an attempt to get away from very, very simple mathematical problems.\u00a0<\/p>\n

And, fundamentally, I\u2019m just not sure anyone gives that much of a shit about being correct outside of sounding as close to the consensus as possible, which is why it\u2019s been so profoundly strange hearing people say things recently that I\u2019ve been saying for years.<\/p>\n

You can shove your fingers or whatever other objects you\u2019ve got handy in your ears and say \u201cla la la\u201d all you want, but Jefferies\u2019 Chris Wood<\/a> just said that \u201cthe most likely long-term outcome of the AI boom will be massive capital destruction in the US<\/a>,\u201d and Barclays just said the following<\/a> (aside: that lab revenue number is completely wrong, but whatever):<\/p>\n

\u2026let’s face facts; the combined disclosed lab revenues of circa $150bn and total Al revenues running at around $250bn is not going to cut it when the ecosystem needs well in excess of $1trn to run even the maintenance capex of this up front spend and I’ve yet been presented with a credible framework as to where this money is going to come from other than pie-in-the-sky TAM numbers whimsically inserted in prospectuses.<\/p>\n

I want to force boosters to read this like the guy at the end of A Clockwork Orange. I want them to answer me, in detail, why this is actually fine.\u00a0<\/p>\n

I already know the answer: they\u2019d produce some sort of asinine number that shows that growth is parabolic, that there\u2019s insatiable demand for AI compute, that Anthropic and OpenAI are the next Google and Facebook combined, that there\u2019s not enough compute to go around, or another anecdote that is literally anything other than a meaningful response.\u00a0<\/p>\n

There is little difference between how the various run rates, projections, \u201cmodels\u201d and \u201cframeworks\u201d used to justify future AI investment are used and the underlying tools of a religious con, except the only community you get here is made up of zyned-up twenty-somethings and the elder millennials desperate for their approval.<\/p>\n

The AI bubble is inflated by bullshit numbers peddled by the ignorant, craven or easily-led. Your ability to distinguish between which of them you\u2019re talking to is critical to not losing your shirt.<\/p>\n

If you liked this piece, you should subscribe to my premium newsletter. It\u2019s $70 a year<\/a>, $18 a quarter<\/a>, or $7 a month<\/a>, and in return you get a weekly newsletter that\u2019s usually anywhere from 10,000 to 18,000 words and provides vast, detailed analyses of the biggest events and companies in the AI bubble.<\/p>\n

If you want to get in touch \u2014 and especially if you have any juicy information about Anthropic, OpenAI, or any other companies in the AI bubble \u2014 hit me up on Signal at ezitron.76. I\u2019m also on IB on The Terminal.<\/p>\n","protected":false},"excerpt":{"rendered":"If you liked this piece, you should subscribe to my premium newsletter, and you can subscribe on the…\n","protected":false},"author":2,"featured_media":801688,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[20],"tags":[554,733,4308,86,56,54,55],"class_list":["post-801687","post","type-post","status-publish","format-standard","has-post-thumbnail","category-artificial-intelligence","tag-ai","tag-artificial-intelligence","tag-artificialintelligence","tag-technology","tag-uk","tag-united-kingdom","tag-unitedkingdom"],"_links":{"self":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/posts\/801687","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/comments?post=801687"}],"version-history":[{"count":0,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/posts\/801687\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/media\/801688"}],"wp:attachment":[{"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/media?parent=801687"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/categories?post=801687"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.newsbeep.com\/uk\/wp-json\/wp\/v2\/tags?post=801687"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}