Annualized Enron Rate
We recorded 28 claims from this post. 16 were checked against evidence. 12 can’t be checked. 14 more are held. Read the post at its source.
Claims from this post (28)
In the order they appear in the post. Each quotation is the speaker’s words in this post. A label such as “Supported” or “Can’t be checked” is the claim’s current assessment; its page has the evidence and the reasons.
Assessment: Can’t be checkedSaid
Rhetoric
“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.”
Our reading: The entire AI bubble is inflated based on made-up numbers that exist only to pump investors' holdings and make AI startups look more successful than they are.
No horizon
Scope: the entire AI bubble
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“There is absolutely no value in using “annualized run rates” 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.”
Our reading: Annualized run rates have absolutely no value for anyone other than the company and its investors, and tech and business journalism's continued acceptance of them for the largest companies is a sign that something is very wrong with tech coverage.
No horizon
Scope: for anyone other than the company and its investors
Assessment: Can’t be checkedSaid
Observation
“I must also say how utterly disappointed I was to see how Bloomberg credulously reported that OpenAI “expected to hit $70 billion in annualized revenue by end of 2026,” a transparent attempt to send the markets back up with a story that tells the world less than nothing.”
Our reading: OpenAI's projection of $70 billion annualized revenue by end of 2026, credulously reported by Bloomberg, was a transparent attempt to lift markets and conveys no information.
No horizon
Assessment: UnresolvedSaid
Observation
“In reality, the company projects to make somewhere in the region of $30 billion in 2026.”
Our reading: OpenAI projects actual 2026 revenue of roughly $30 billion.
No horizon Evidence as of
Limitation: The decisive reporting describes private company materials through another news organization.
Assessment: UnresolvedSaid
Observation
“In fact, if you added up every dollar that both Anthropic and OpenAI made in the first two quarters, you’d be $19.1 billion short of Anthropic’s May “run rate.””
Our reading: Combined H1 2026 revenue of Anthropic and OpenAI was $19.1 billion short of Anthropic's May 2026 $48 billion run rate.
No horizon Evidence as of
Limitation: The packet does not independently establish OpenAI's $12.4B H1 revenue figure [e1].
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“Reporting “run rates” is a marketing exercise, and if you disagree, please answer me the following: what does a “$65 billion annualized run rate” tell you about Anthropic, exactly?”
Our reading: Reporting run rates, such as Anthropic's '$65 billion annualized run rate', is a marketing exercise.
No horizon
Assessment: Can’t be checkedSaid
Rhetoric
“Investors are fucking stupid for using it.”
Our reading: Investors are stupid for using annualized run rates.
No horizon
Assessment: SupportedSaid
Observation
“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.”
Our reading: Microsoft has only around 2GW of operational AI capacity despite claiming to bring a gigawatt online in each of its last three fiscal quarters.
No horizon Evidence as of
Limitation: Support concerns the reported AI-specific estimate, not an audited inventory of every source of Microsoft's AI compute.
Assessment: Can’t be checkedSaid
Causal claim
“Let’s be clear about something: the reason we’re not getting these numbers is that these companies are not actually making these amounts on an annual basis.”
Our reading: AI startups withhold real revenue figures because they do not actually earn their claimed annualized amounts.
No horizon
Assessment: Can’t be checkedSaid
Causal claim
“The problem — and the reason I’d argue that the actual revenues of these companies are so shrouded in secrecy — is that the underlying economics do not match up when you consider that effectively every AI startup has to resell tokens at a discount.”
Our reading: AI startups' revenues are kept secret because effectively every one must resell tokens at a discount, so the economics do not work.
No horizon
Assessment: UnresolvedSaid
Causal claim
“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.”
Our reading: In AI, more demand produces larger losses, and the most engaged customers are the least profitable.
No horizon Evidence as of
Limitation: The claim is assessed as a general pattern rather than a literal universal law.
Assessment: UnresolvedSaid
Conditional forecast
“And if they can’t do that, large chunks of OpenAI and Anthropic’s revenues disappear.”
Our reading: If AI startups can no longer keep raising money, large chunks of OpenAI's and Anthropic's revenues will disappear.
No deadline stated Evidence as of
Only if: if AI startups can no longer raise unlimited money
Assessment: UnresolvedSaid
Observation
“The Revenues80% of Anthropic and OpenAI’s 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.”
Our reading: 80% of Anthropic's and OpenAI's revenues come from 1% of their customers, the vast majority of which are AI startups that spend 300% to 1000% of their revenue on AI tokens.
No horizon Evidence as of
Limitation: A sample limitation is not proof that the full-population claim is false.
Assessment: UnresolvedSaid
Causal claim
“As a result, every AI startup runs at a loss. This necessitates raising massive amounts of capital every few months.”
Our reading: Because AI startups have very small revenues even with discounted tokens, every AI startup runs at a loss, which forces them to raise massive amounts of capital every few months.
No horizon Evidence as of
Limitation: No comparable target-quarter accounts establish losses across the defined startup population.
Assessment: UnresolvedSaid
Causal claim
“This means that Anthropic and OpenAI get a massive bump of new revenues — funded by AI startups that are allowing users to burn tokens without restraint — with every release.”
Our reading: With every model release, Anthropic and OpenAI get a massive bump of new revenues funded by AI startups that let users burn tokens without restraint.
No horizon Evidence as of
Limitation: The companies, metrics, and coverage of the available records differ.
Assessment: UnresolvedSaid
Causal claim
“It doesn’t really matter what big businesses are or are not using Claude or ChatGPT — the underlying revenues of these companies are held up by venture capital by proxy.”
Our reading: Anthropic's and OpenAI's revenues are held up by venture capital by proxy via AI startups.
No horizon Evidence as of
Limitation: The sample does not identify how much of either provider's total revenue came from AI startups [e1].
Assessment: UnresolvedSaid
Causal claim
“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.”
Our reading: Every AI startup is effectively an extraction mechanism for OpenAI and Anthropic, reselling their tokens at a massive discount because users are unwilling to pay the actual cost of using LLMs.
No horizon Evidence as of
Limitation: A literal reading covering unrelated AI businesses would be broader than the surrounding argument.
Assessment: Can’t be checkedSaid
Causal claim
“The obfuscation of the actual underlying financial of these companies is intentional, because it makes it harder to do the kinds of comparisons I’ve done today,”
Our reading: The obscuring of AI startups' actual underlying financials is intentional, because it makes the kinds of comparisons he made in the post harder to do.
No horizon
Assessment: UnresolvedSaid
Observation
“Again and again, I’ve found companies paying way, way less than the supposed going rate of compute.”
Our reading: Zitron has repeatedly found companies paying far less than the supposed going rate of compute.
No horizon Evidence as of
Limitation: The survey supplies contractual rates, but no verified comparison readings from the identified public benchmark [e1].
Assessment: ContradictedSaid
Observation
“Your depreciation costs here work out to about $1.54-per-GPU-per-hour, meaning that you’re losing nearly half of your revenue for the first six years…of a five year long contract.”
Our reading: In the Canopy Wave B300 contract, depreciation of about $1.54 per GPU-hour means the provider loses nearly half its revenue over the first six years of what is only a five-year contract.
No horizon Evidence as of
Limitation: The assessment is conditional arithmetic under the stated cost assumption, not a finding about actual company accounts.
Assessment: UnresolvedSaid
Observation
“As I said a few weeks ago, 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 about how much they’re bringing online.”
Our reading: There is no firm understanding of how much AI compute exists, and Microsoft is actively misleading the public about capacity coming online.
No horizon Evidence as of
Limitation: The independent estimates cover selected developers and distinguish compute use from ownership; they do not provide a complete operational census [e2].
Assessment: UnresolvedSaid
Observation
“As a result, basically any AI compute capacity estimates are guesses.”
Our reading: Essentially all AI compute capacity estimates are guesses.
No horizon Evidence as of
Limitation: The research is not a representative audit of the estimation industry.
Assessment: Can’t be checkedSaid
Observation
“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 “how does that work?” or “how did you get that number?””
Our reading: The vast majority of words written and billions invested in the AI bubble rest on stacked assumptions, several of which fall apart when you ask simple questions such as how something works or how a number was derived.
No horizon
Scope: several of which
Assessment: UnresolvedSaid
Observation
“their immediate response is to say that 2026 will be fine because of quarters with discounted compute costs and adjusted operating profits that don’t actually resolve the underlying problem that these companies have $1.3 trillion in compute commitments, with at least $413 billion of them being uncancellable.”
Our reading: Discounted compute and adjusted profits do not solve the problem that Anthropic and OpenAI have $1.3 trillion in compute commitments, at least $413 billion uncancellable.
No horizon Evidence as of
Limitation: The reported nominal amounts are not reconciled to payments already made, overlap or the precise remaining uncancellable portion [e1, e2].
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“This is extremely bad and extremely important to resolve.”
Our reading: The lack of knowledge about AI companies' financial condition is very bad and should be resolved.
No horizon
Assessment: UnresolvedSaid
Causal claim
“Anthropic and OpenAI’s 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.”
Our reading: Anthropic's and OpenAI's growth depends on AI startups being able to spend increasing amounts on tokens every month, without fail, for what appears to be the rest of time.
No horizon Evidence as of
Limitation: Revenue growth is an interpretation of the speaker's unspecified growth measure.
Assessment: Can’t be checkedSaid
Rhetoric
“Most coverage of AI lives in its own bubble where nothing bad has happened or could happen, and that every promise will come true.”
Our reading: Most AI coverage lives in its own bubble in which nothing bad has happened or could happen and every promise will come true.
No horizon
Scope: most coverage of AI
Assessment: Can’t be checkedSaid
Rhetoric
“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’re talking to is critical to not losing your shirt.”
Our reading: The AI bubble is inflated by bogus numbers pushed by the ignorant, the craven or the easily led, and being able to tell which of these you are talking to is critical to avoiding heavy losses.
No horizon
Held from this post (14)
These records did not pass every check, so nothing about them is published. Held is not a finding.
- The two reviews read the claim differently
- The two reviews read the claim differently
- The criteria are ambiguous. Our reading, not assessed: Zitron can find no compelling evidence that there was even a gigawatt of compute available for Anthropic to buy, despite New York Times-reported investor statements that it would have 5GW by the end of 2026.
- The two reviews reached different assessments. Our reading, not assessed: Google's, Amazon's and Oracle's AI revenues are not publicly known, and only a single year of Microsoft's AI revenues, fiscal year 2026, is available, as a Bloomberg estimate calculated from Microsoft's SEC-filing disclosure of its OpenAI revenues.
- The two reviews read the claim differently
- The two reviews read the claim differently
- The two reviews read the claim differently
- The two reviews read the claim differently
- The two reviews read the claim differently
- The two reviews read the claim differently
- The two reviews read the claim differently
- The two reviews reached different assessments. Our reading, not assessed: It appears that Anthropic counts every dollar of sales in its annualized run rate figure while OpenAI does not, and Zitron does not buy OpenAI's explanation that its approach is 'more ethical.'
- The two reviews read the claim differently
- The two reviews read the claim differently