The AI Industry Is Lying To You
We recorded 31 claims from this post. 16 were checked against evidence. 15 can’t be checked. 25 more are held. Read the post at its source.
Claims from this post (31)
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 built on a vague sense of inevitability — that if everybody just believes hard enough that none of this can ever, ever go wrong that at some point all of the very obvious problems will just go away.”
Our reading: The AI bubble rests on a vague belief that it is inevitable and that its obvious problems will go away.
No horizon
Assessment: UnresolvedSaid
Observation
“Even if they were, there’s not enough power for them to turn on.”
Our reading: Even if the data center capacity described as under active development (about 61GW of IT load) were being built, there is not enough power for it to turn on.
No horizon Evidence as of
Limitation: Regional adequacy findings cannot automatically be applied to every US project.
Assessment: UnresolvedSaid
Observation
“Nevertheless, based on these data points, I’m comfortable estimating that North American data center absorption — as the IT load of data centers actually turned on and in operation — was at around 3GW for 2025, which would work out to about 3.9GW of total power.”
Our reading: About 3GW of IT load (about 3.9GW of power) of North American data center capacity was actually turned on in 2025.
No horizon Evidence as of
Limitation: Market coverage differs substantially between trackers.
Assessment: UnresolvedSaid
Observation
“NVIDIA’s overall data center segment revenue was $195.7 billion, which puts America’s data center purchases at around $135 billion, leaving around $44 billion of GPUs and associated technology uninstalled.”
Our reading: Around $44 billion of NVIDIA GPUs and associated gear sold to America in FY2026 remained uninstalled.
No horizon Evidence as of
Limitation: Revenue is not customer-held uninstalled inventory.
Assessment: UnresolvedSaid
Observation
“With the acceleration of NVIDIA’s GPU sales, it now takes about 6 months to install and operationalize a single quarter’s worth of sales.”
Our reading: It currently takes about six months to install and operationalize one quarter's worth of NVIDIA GPU sales.
No horizon Evidence as of
Limitation: Public financial and property-market reports are proxies for the physical quantities at issue.
Assessment: UnresolvedSaid
Forecast
“By the time they actually get all the Blackwells in Vera Rubin will be two years old!”
Our reading: By the time all purchased Blackwell GPUs are installed, Vera Rubin will be two years old.
No deadline stated Evidence as of
Limitation: Both the cohort and the starting point for Rubin's age need clearer definitions.
Assessment: UnresolvedSaid
Causal claim
“NVIDIA is selling GPUs years into the future, and its ability to grow, or even just maintain its current revenues, depends wholly on its ability to convince people that this is somehow rational.”
Our reading: NVIDIA is selling GPUs years into the future, and its ability to grow, or even just maintain its current revenues, depends wholly on its ability to convince people that this is somehow rational.
No horizon Evidence as of
Limitation: Customer announcements do not disclose all purchase terms or revenue recognition dates.
Assessment: UnresolvedSaid
Forecast
“Developer Mortensen claims the entire project will be completed by October 2026, which it obviously, blatantly won’t.”
Our reading: Stargate Abilene will not be fully completed by October 2026.
Not yet due Evidence as of
Limitation: Oracle's statement is available through a third-party transcript host.
Assessment: UnresolvedSaid
Conditional forecast
“At its current rate of construction, Stargate Abilene will be fully built sometime in late 2027.”
Our reading: At its current construction rate, Stargate Abilene will be fully built in late 2027.
Not yet due Evidence as of
Only if: at its current rate of construction
Assessment: Can’t be checkedSaid
Observation
“I imagine Microsoft has successfully brought online about 235MW of power for Fairwater.”
Our reading: Microsoft has probably brought only about 235MW of power online for its Fairwater Wisconsin data center.
No horizon
Assessment: UnresolvedSaid
Forecast
“All Nebius had to do to raise $3.75 billion in debt was to sign a deal with Meta for data center capacity that doesn’t exist and will likely take three to four years to build (it’s never happening).”
Our reading: The data center capacity Nebius contracted to Meta, which let it raise $3.75 billion in debt, does not exist, will likely take three to four years to build, and will never happen.
No deadline stated Evidence as of
Limitation: Company construction updates are self-reported.
Assessment: Can’t be checkedSaid
Observation
“The tech industry has taken advantage of an understandable lack of knowledge about construction or power timelines in the media to pump out endless stories about “data center capacity in progress” as a means of obfuscating an ever-growing scandal: that hundreds of billions of NVIDIA GPUs got sold to go in projects that may never be built.”
Our reading: Hundreds of billions of dollars of NVIDIA GPUs were sold for data center projects that may never be built, and the industry obscures this through media stories.
No horizon
Assessment: Can’t be checkedSaid
Conditional forecast
“Furthermore, if the supply of AI compute is a bottleneck, this suggests that when (or if) that bottleneck is ever cleared, there will suddenly be a massive supply glut, lowering the overall value of the data centers in progress…”
Our reading: If AI compute supply is a bottleneck, this suggests that when (or if) the bottleneck is ever cleared there will suddenly be a massive supply glut that lowers the value of data centers in progress, which are all filled with Blackwell GPUs that will be two or three years old by the time those data centers are finally turned on.
No horizon
Only if: if the supply of AI compute is a bottleneck
Assessment: Can’t be checkedSaid
Rhetoric
“The entire AI bubble is a god damn mirage. Every single “multi-gigawatt” data center you hear about is a pipedream,”
Our reading: The AI bubble is a mirage and every announced multi-gigawatt data center is a pipedream funded by private credit.
No horizon
Assessment: Can’t be checkedSaid
Observation
“2025’s $178.5 billion in US-based data center deals doesn’t appear to be resulting in any immediate (or even future) benefit to anybody involved.”
Our reading: The $178.5 billion of US data center deals in 2025 does not appear to benefit anyone involved.
No horizon
Assessment: UnresolvedSaid
Forecast
“At some point reality will set in and spending on NVIDIA GPUs will have to decline.”
Our reading: Spending on NVIDIA GPUs will eventually have to decline.
No deadline stated Evidence as of
Limitation: Supplier revenue is not identical to customer cash spending.
Assessment: UnresolvedSaid
Conditional forecast
“Simple questions like “where are the GPUs going?” and “how many actual GPUs have been installed?” are left unanswered as article after article gets written about massive, multi-billion dollar compute deals for data centers that won’t be built before, at this rate, 2030.”
Our reading: At the current rate, data centers in the announced multi-billion dollar compute deals will not be built before 2030.
Not yet due Evidence as of
Only if: at the current rate
Assessment: UnresolvedSaid
Causal claim
“And I’d argue it’s convenient to blame this solely on power issues, when the reality is clearly based on construction timelines that never made any sense to begin with. If it was just a power issue, more data centers would be near or at the finish line, waiting for power to be turned on.”
Our reading: Data center delays come from unrealistic construction timelines, not only power, because otherwise more would be finished and waiting for power.
No horizon Evidence as of
Limitation: Aggregate delay statistics do not establish project-level causation.
Assessment: UnresolvedSaid
Observation
“Then there’s the very, very obvious scandal that NVIDIA, the largest company on the stock market, is making hundreds of billions of dollars of revenue on chips that aren’t being installed.”
Our reading: NVIDIA is earning hundreds of billions of dollars of revenue on chips that are not being installed.
No horizon Evidence as of
Limitation: The annual revenue figure covers a different period from the contract's October 27, 2025–January 25, 2026 reference quarter [e1].
Assessment: Can’t be checkedSaid
Conditional forecast
“Assuming that Vera Rubin actually ships in 2026, it’s reasonable to believe that people will be installing these things well into 2028, if not further, and that’s assuming everything doesn’t collapse by then.”
Our reading: If Vera Rubin ships in 2026, it will probably still be getting installed well into 2028 or later, unless things collapse first.
No horizon
Only if: Vera Rubin ships in 2026
Assessment: Can’t be checkedSaid
Observation
“While the indictment isn’t specific as to the breakdown, it confirms that some Blackwell GPUs made it to China, and I’d wager quite a few.”
Our reading: Likely quite a few NVIDIA Blackwell GPUs were smuggled to China.
No horizon
Assessment: Can’t be checkedSaid
Forecast
“The arrest of Liaw has, it seems, ruffled some feathers in Washington, and I would not be shocked to see Huang sat before a congressional inquiry at some point.”
Our reading: Jensen Huang may at some point be called before a congressional inquiry about GPU diversion to China.
No horizon
Assessment: UnresolvedSaid
Observation
“So, I want to explain something very clearly for you, because it’s important you understand how fucked up shit has become: hyperscalers are forcing everybody in their companies to use AI tools as much as possible, tying compensation and performance use to token burn, and actively encouraging non-technical people to vibe-code features that actually reach production.”
Our reading: Hyperscalers are forcing everybody in their companies to use AI tools as much as possible, tying compensation and performance to token burn, and actively encouraging non-technical people to vibe-code features that actually reach production.
No horizon Evidence as of
Limitation: Several sources summarize internal records that were not publicly available.
Assessment: Can’t be checkedSaid
Forecast
“Despite the furious (and exhausting) marketing campaign around “the power of AI code,” I believe that these events are just the beginning of the true consequences of AI coding tools: the slow destruction of the tech industry’s software stack.”
Our reading: The Meta and Amazon incidents are only the start of AI coding tools slowly destroying the tech industry's software stack.
No horizon
Assessment: Can’t be checkedSaid
Observation
“The problem is that while LLMs can write “all” code, that doesn’t mean the code is good, or that somebody can read the code and understand its intention (as these models do not think), or that having a lot of code is a good thing both in the present and in the future of any company built using generative code.”
Our reading: LLM-written code is not necessarily good or understandable, and LLMs do not think.
No horizon
Assessment: Can’t be checkedSaid
Forecast
“But what I do know is that big tech is setting itself up for crisis after crisis, especially when Anthropic and OpenAI stop subsidizing their models to the tune of allowing people to spend $2500 or more on a $200-a-month subscription.”
Our reading: Big tech is setting itself up for crisis after crisis, especially when Anthropic and OpenAI stop subsidizing their models to the extent of letting people spend $2,500 or more on a $200-a-month subscription.
No horizon
Qualifier: especially when Anthropic and OpenAI stop subsidizing their models
Assessment: Can’t be checkedSaid
Forecast
“Generative code is a digital ecological disaster, one that will take years to repair thanks to company remits to write as much code as fast as possible.”
Our reading: The damage from generative code is a disaster that will take years to repair.
No horizon
Assessment: Can’t be checkedSaid
Recommendation or value judgment
“Every single person responsible must be held accountable, especially for the calamities to come as lazily-managed software companies see the consequences of building their software on sand.”
Our reading: Everyone responsible for pushing generative code must be held accountable for the calamities to come.
No horizon
Assessment: Can’t be checkedSaid
Rhetoric
“In the end, everything about AI is built on lies.”
Our reading: Everything about AI is built on lies.
No horizon
Assessment: UnresolvedSaid
Observation
“Anthropic’s constant flow of “annualized” revenues ended up equating to literally $5 billion in revenue in four years, on $25 billion or more in salaries and compute.”
Our reading: Over four years Anthropic made about $5 billion in actual revenue against $25 billion or more in salaries and compute.
No horizon Evidence as of
Limitation: The speaker does not identify the four years; the contract provisionally uses 2022–2025.
Assessment: UnresolvedSaid
Causal claim
“Software engineers aren’t being replaced — they’re being laid off because the software that’s meant to replace them is too expensive, while in practice not replacing anybody at all.”
Our reading: AI is not replacing software engineers; they are being laid off because AI tooling is so expensive.
No horizon Evidence as of
Limitation: Corporate explanations may omit other motives.
Held from this post (25)
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 two reviews read the claim differently
- The two reviews reached different assessments. Our reading, not assessed: AI is expensive, unproductive and harms the learning and effectiveness of its users, which poisons the tech industry's future.
- The two reviews reached different assessments. Our reading, not assessed: The concept of a hundred-megawatt data center is barely a few years old, Zitron cannot find a built, in-service gigawatt data center of any kind, and OpenAI cannot afford to pay its bills.
- The criteria are ambiguous. Our reading, not assessed: The supposedly planned OpenAI extension to Stargate Abilene never existed and was never going to happen.
- The two reviews reached different assessments. Our reading, not assessed: The heavy, mandated AI use at hyperscalers should not be read as a sign that AI has taken a leap in efficacy; its actual outcomes are a labyrinth of half-functional internal dashboards that measure random user data or convert files, spending hours to save minutes at some theoretical point.
- The record does not meet the current rules. Our reading, not assessed: Zitron does not think many of the projects behind the roughly 61GW of IT load listed as under active development in the US are actually being built right now, citing only $178.5 billion in data center debt deals last year against more than a trillion dollars of GPUs that load would require.
- The two reviews read the claim differently
- The two reviews read the claim differently
- The two reviews read the claim differently
- The criteria are ambiguous. Our reading, not assessed: Most of the hundreds of billions of dollars of GPUs sold are sitting uninstalled.
- The two reviews read the claim differently
- The criteria are ambiguous. Our reading, not assessed: If roughly 3GW of IT load came online in America, that should in theory mean tens of billions of dollars of revenue, yet nobody appears to be showing massive amounts of revenue from these data centers.
- Attribution is uncertain
- The two reviews read the claim differently
- The two reviews read the claim differently
- The two reviews read the claim differently
- The criteria are ambiguous. Our reading, not assessed: AI makes very little revenue, lacks tangible productivity benefits, and seems, at scale, to actively harm the productivity and efficacy of the workers being forced to use it.
- The record does not meet the current rules. Our reading, not assessed: OpenAI's ~$8.67B Azure spend and Anthropic's ~$2.66B AWS spend through Sept 2025 suggest real AI demand is weak and concentrated in a few companies.
- The two reviews read the claim differently
- The criteria are ambiguous. Our reading, not assessed: No one appears to profit from renting out AI compute.
- The two reviews read the claim differently
- The two reviews read the claim differently
- The two reviews read the claim differently