AI Is Too Expensive
We recorded 23 claims from this post. 23 can’t be checked. 6 more are held. Read the post at its source.
Claims from this post (23)
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
Observation
“If we assume that Microsoft’s other AI services grew 10% quarter-over-quarter, I estimate that Microsoft likely made around $17.9 billion in AI revenue in FY2025, or a little under a fifth of its capex.”
Our reading: Assuming Microsoft's other AI services grew 10% quarter-over-quarter, Zitron estimates Microsoft likely made around $17.9 billion in AI revenue in FY2025, a little under a fifth of its capex.
No horizon
Qualifier: assuming Microsoft's other AI services grew 10% quarter-over-quarter
Assessment: Can’t be checkedSaid
Observation
“I also cannot find an economic scenario where this pays itself off.”
Our reading: Zitron says he cannot find an economic scenario in which hyperscalers' AI capex pays itself off.
No horizon
Qualifier: I cannot find
Assessment: Can’t be checkedSaid
Causal claim
“If the margins were good, they’d be sharing the margins. If the revenues were good, they’d be sharing the revenues (and no, run rates aren’t revenues).”
Our reading: Hyperscalers' failure to disclose AI margins and revenues indicates they are not good.
No horizon
Qualifier: inferred from non-disclosure
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“And it all comes back to a very simple point: AI is too expensive.”
Our reading: AI is too expensive.
No horizon
Assessment: Can’t be checkedSaid
Forecast
“(Oracle is trying to compete and may die as a result)”
Our reading: Oracle may die as a result of trying to compete as an AI cloud provider.
No horizon
Qualifier: as a result of trying to compete in AI cloud
Assessment: Can’t be checkedSaid
Conditional forecast
“Sidenote: I have absolutely no idea what Meta does, and my chaos bet is that it starts renting out its compute to Anthropic or OpenAI when things get rough.”
Our reading: As a speculative bet, Meta may start renting its compute to Anthropic or OpenAI when things get rough.
No horizon
Only if: when things get rough
Assessment: Can’t be checkedSaid
Observation
“Anthropic has done a great deal of work to obfuscate how much it actually makes or spends, but I think it’s likely it burns even more than OpenAI, given the fact that it’s had to raise $75 billion in the last 6 months”
Our reading: Anthropic obfuscates its finances and likely burns even more money than OpenAI.
No horizon
Qualifier: inferred from $75B raised in 6 months
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“Numerous reports about how Anthropic will turn “cash flow positive” in either 2027 or 2028 are fantastical, illogical, entirely driven by ridiculous projections, and should never have been reported as anything other than an attempt by companies to mislead their investors.”
Our reading: Numerous reports that Anthropic will turn cash-flow positive in 2027 or 2028 are fantastical, illogical and driven entirely by ridiculous projections, and should never have been reported as anything other than an attempt by companies to mislead their investors.
No horizon
Assessment: Can’t be checkedSaid
Conditional forecast
“Because even if Anthropic were making $100 billion in annual revenue, it would probably be losing $300 billion or more to get there.”
Our reading: If Anthropic reached $100B annual revenue it would probably be losing $300B or more to get there.
No horizon
Only if: if Anthropic reached $100B annual revenue
Assessment: Can’t be checkedSaid
Observation
“The fact it had to raise $30 billion in February, $15 billion in April, and now $30 billion more in May all while allegedly pulling in more than $3 billion a month in revenue suggests that its COGS are fucking horrendous, and its growth is coming at a terrible financial cost.”
Our reading: Anthropic's raising $30 billion in February, $15 billion in April and $30 billion in May while allegedly bringing in more than $3 billion a month suggests its COGS are horrendous and its growth is coming at a terrible financial cost.
No horizon
Qualifier: allegedly
Assessment: Can’t be checkedSaid
Causal claim
“I think the answer is simple: their CFOs know that doing so would reveal their actual margins, which are hot dogshit with sprinkles on top.”
Our reading: Zitron thinks Anthropic's and OpenAI's CFOs know that going public would reveal their actual margins, which he calls extremely bad, and that this is why neither appears to be seriously considering an IPO.
No horizon
Qualifier: I think
Assessment: Can’t be checkedSaid
Forecast
“One important detail to keep in mind here is that as of a month or two ago, Anthropic moved all enterprise customers to token-based-billing, which will begin, I believe, a true stress-test of the true “value” of AI as costs skyrocket.”
Our reading: Zitron believes Anthropic's move of all enterprise customers to token-based billing will begin a true stress-test of AI's value as costs skyrocket.
No horizon
Qualifier: I believe
Assessment: Can’t be checkedSaid
Causal claim
“Executives have, across the board, given their engineers free reign to burn as many tokens as they’d like, and while I severely doubt that Anthropic actually hit $50 billion in annualized revenue outside of not-quite-fraudulent non-GAAP measurements, I believe its revenue growth has come from an artificial boost from a tech industry searching for a reason to pay somebody money.”
Our reading: Executives across the board have given engineers free rein to burn tokens, Zitron severely doubts Anthropic actually hit $50 billion in annualized revenue outside of not-quite-fraudulent non-GAAP measurements, and he believes its revenue growth has come from an artificial boost from a tech industry searching for a reason to pay somebody money.
No horizon
Scope: outside of not-quite-fraudulent non-GAAP measurements
Assessment: Can’t be checkedSaid
Observation
“To be very clear about what I mean, I think there is currently an AI token binge across both Anthropic and OpenAI.”
Our reading: Zitron thinks there is currently an AI token binge across both Anthropic and OpenAI.
No horizon
Qualifier: I think
Assessment: Can’t be checkedSaid
Causal claim
“This might be because Anthropic obfuscates the data that might tell customers the real costs.”
Our reading: Executives pushing AI use without knowing what it costs might be because Anthropic obfuscates the data that might tell customers the real costs.
No horizon
Qualifier: might be because
Assessment: Can’t be checkedSaid
Observation
“While I’m not accusing Anthropic of anything untoward, massive, multi-million dollar contracts that involve individuals burning thousands or tens of thousands of dollars’ worth of tokens with no service level agreement, transparency or true granularity into the burn is a perfect setup for a company — not saying it’s Anthropic! — to do something dastardly with those numbers.”
Our reading: Multi-million-dollar token contracts with no service level agreement, transparency or true granularity into usage are a perfect setup for a company to do something dastardly with the numbers, though Zitron explicitly says he is not accusing Anthropic.
No horizon
Scope: no service level agreement, transparency or true granularity
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“Unless, of course, you can’t actually measure how many tokens a particular task can take with much accuracy, in which case every single AI token budget is bullshit.”
Our reading: If you can't measure with much accuracy how many tokens a particular task takes, then every AI token budget is bullshit.
No horizon
Only if: if you can't measure how many tokens a task takes with much accuracy
Assessment: Can’t be checkedSaid
Rhetoric
“Before AI, this wasn’t as much of a problem, in the sense that inefficiencies and wasted hours weren’t directly connected to a chatbot that is specifically designed to burn money.”
Our reading: AI chatbots are designed to burn money, tying workplace inefficiency directly to a growing cost.
No horizon
Assessment: Can’t be checkedSaid
Causal claim
“Organizations aren’t burning millions or hundreds of millions of dollars a year on AI because it’s good, they’re doing it because they are run by people who do not know what the fuck they’re doing.”
Our reading: Organizations spend heavily on AI because of incompetent leadership, not because AI is good.
No horizon
Assessment: Can’t be checkedSaid
Rhetoric
“Anthropic is a company purpose-built to con people into giving it by money by making people feel smart.”
Our reading: Anthropic is a company purpose-built to con people into giving it money by making them feel smart.
No horizon
Assessment: Can’t be checkedSaid
Rhetoric
“Nobody wants this, nobody wanted it since the beginning, it was forced upon everyone, and to pretend otherwise is laughable and offensive.”
Our reading: There is no genuine demand for generative AI; it was forced on everyone.
No horizon
Assessment: Can’t be checkedSaid
Rhetoric
“Generative AI was purpose-built to grift an economy run by executives and managers who don’t actually do any work.”
Our reading: Generative AI was built to grift executives and managers who don't do real work.
No horizon
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“We should make everybody who obsessed over theoretical ideas about what AI can or will do ashamed for their intellectual deceit or constant ignorance.”
Our reading: AI boosters who obsessed over theoretical AI capabilities should be shamed for their deceit or ignorance.
No horizon
Held from this post (6)
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 reached different assessments. Our reading, not assessed: Generative AI is good at imitating work and will claim it can do tasks it cannot.
- The two reviews read the claim differently
- The two reviews reached different assessments. Our reading, not assessed: Anthropic cannot afford to make around $100 billion in annualized revenue, because there is no mechanism for it to bring its costs down and more compute does not make OpenAI's or Anthropic's services cheaper to offer.
- The two reviews read the claim differently
- The two reviews read the claim differently