“Otherwise, there’s very little evidence — and boy, have I looked — that there’s more than a few billion in demand for AI compute, and that’s being generous.”
Our reading:
Outside the largest compute consumers (Microsoft, Google, Amazon, CoreWeave, Meta and Oracle), there is very little evidence of more than a few billion dollars in demand for AI compute, and that is being generous.
Assessment: Unresolved
What happened
Was it true for the period described?
Outcome: Unresolved
We checked, and the evidence so far doesn’t settle it.
Judged on a plain reading. They named no measure, so we used the reading a typical informed reader would take. It is spelled out under “How we judged it”.
Evidence up to . Checked .
Why
The search found substantial evidence of xAI compute investment outside the named group. That weighs against the broad suggestion that such demand has little evidence. But hardware value, equipment financing and compute-service spending are different measures, and Zitron supplies no period or accounting basis for his dollar comparison. These records do not justify either a quarterly ceiling or a definitive aggregate contradiction.
Caveats
- The unknown accounting basis prevents a supported or contradicted monetary score.
- The historical examples are context, not outcome evidence for the publication-date snapshot.
Evidence (2 items)
Context only: arXiv; Konstantin F. Pilz and coauthors,
estimated hardware cost
Why it matters: This is material evidence of substantial compute investment outside the six named consumers. It measures an earlier hardware stock, so it cannot establish a July 2026 annual or quarterly demand total.
- Publisher
- arXiv; Konstantin F. Pilz and coauthors
- Kind of source
- Preprint, not peer reviewed
- Measures
- xAI, Estimated hardware acquisition cost of Colossus, USD billions, estimate, ? to 2025-03
- About events on
- Published
Context only: Apollo Global Management,
NVIDIA invested in VCI
Why it matters: The announcement describes a $5.4 billion acquisition and lease of compute infrastructure to an xAI subsidiary, supported by a $3.5 billion capital solution. This is relevant outside demand, but the transaction value is neither a disclosed annual rental bill nor an additional amount to add mechanically to the Colossus estimate.
- Publisher
- Apollo Global Management
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- Measures
- Valor Compute Infrastructure; xAI subsidiary as lessee, Announced acquisition-and-lease transaction value, USD billions, company reported, ? to 2026-01-07
- About events on
- Published
Probably wrong?
Both AIs think this is probably wrong. The evidence isn’t conclusive, so the verdict above stays unresolved. It counts only when you pick Reasonable or Speculative under “How sure?”.
- Anthropic’s AI: xAI alone sits outside the named consumers, with a Colossus cluster estimated at $7B in hardware (Pilz et al.) and a $5.4B xAI GPU acquisition-and-lease deal (Apollo, Jan 2026). Both clearly exceed a few billion, despite basis differences.
- OpenAI’s AI: Apollo disclosed a $5.4 billion xAI compute transaction, while independent researchers estimated Colossus hardware at $7 billion. Without adding potentially overlapping figures, these show substantial demand outside his list.
How we judged it
- Kind of claim
- Observation
- Deadline
- None given. We don’t add one.
Conditions
- outside Microsoft, Google, Amazon, CoreWeave, Meta and Oracle (their words)
- and that's being generous (their words)
Where we had to interpret
- The plain reading concerns publicly evidenced demand outside the six named consumers, as it stood when the post was published.
- The target date is a snapshot of that public evidence. It does not restrict demand to a quarter or convert the claim into an annual spending ceiling.
- The source supplies neither a spending interval nor a distinction between hardware purchases, compute-service spending and outstanding contracts. These measures must remain separate.
- The surrounding passage attributes substantial demand to OpenAI and Anthropic through the named providers. Counting that same demand again as outside demand would not fairly test the passage.
- A few billion is an approximate scale, not a specified numerical cutoff. No precise accounting basis is supplied.
Right if: A sufficiently comprehensive public accounting of demand outside the named consumers shows only a few billion dollars on a clearly identified, comparable measurement basis. Failure to find contracts alone is insufficient.
Wrong if: Public records available by the post show demand outside the named consumers substantially exceeding a few billion dollars on a comparable basis, without double-counting provider and customer transactions. A result that depends on substituting cumulative hardware value for annual service spending is insufficient.
We wrote this test after the outcome was known.
What would settle it
A defined transaction basis and spending interval, followed by a non-overlapping inventory of qualifying customer demand publicly evidenced by July 15, 2026.
We found 45 claims in this post. See them all.
Full record
Source and dates
Where's Your Ed At, The OpenAI Bubble, paragraph 28.
- Said
- We read the source
- live page, full text, public
- Evidence up to
- Checked
We read this wording on 2026-10-28. The post may have been edited after 2026-07-15.
Where it was said (1)
Said once in the posts we read.
: Where's Your Ed At, The OpenAI Bubble
Otherwise, there’s very little evidence — and boy, have I looked — that there’s more than a few billion in demand for AI compute, and that’s being generous.
- Where
- paragraph 28
- Published
- 2026-07-15
- We read it
- , live page
- Access
- full text, public
History
Versions
- Current: Unresolved. Evidence up to 2026-10-28.
What happened to this record
PublishedBy us
Published after both AI reviews and the final check passed.
Held backBy us
A reviewer said our test wasn’t fair to what was actually said.