“While OpenAI and Anthropic have raised nearly $300 billion in the last few years, their actual infrastructure costs — the GPUs and the data centers to power their services — were entirely funded by hyperscalers, likely costing another $250 billion in the process”
Our reading:
OpenAI's and Anthropic's infrastructure costs were entirely funded by hyperscalers, likely about $250 billion on top of nearly $300 billion they raised.
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 Abilene disclosures identify outside capital partners funding infrastructure used by OpenAI. They are material counterevidence to a literal claim that every funding dollar came from hyperscalers. But the same project involves Oracle, and the records do not settle whether Zitron means project capital or the economic commitments underwriting it. A categorical score would therefore depend on a financing interpretation he did not specify. His hedged $250 billion estimate is not treated as an exact realized-cost promise.
Caveats
- The project evidence does not measure the combined historical infrastructure costs of both labs.
- Outside financing does not by itself establish who ultimately bears the infrastructure cost.
- The beginning of the historical interval is unspecified.
Evidence (2 items)
Context only: Crusoe,
strategic capital partnerships
Why it matters: Crusoe identifies outside capital partnerships for the $15 billion Abilene joint venture. This challenges a literal all-capital-from-hyperscalers reading, but does not establish how funding obligations should be attributed after accounting for the tenant's commitments.
- Publisher
- Crusoe
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- Measures
- Crusoe, Blue Owl Capital and Primary Digital Infrastructure, Announced Abilene data-center joint-venture value, USD billions, company reported, ? to 2025-05-21
- About events on
- Published
Context only: OpenAI,
now up and running.
Why it matters: OpenAI identifies Abilene as an operating Stargate site and describes training and inference using Oracle-delivered equipment. This connects the project in e3 to OpenAI, while also showing why Oracle's role cannot be ignored when interpreting who funded it.
- Publisher
- OpenAI
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- About events on
- Published
Probably wrong?
One AI thinks this is probably wrong, and the other says it’s too early to tell. The evidence isn’t conclusive, so the verdict above stays unresolved. It counts only when you pick Speculative under “How sure?”.
- Anthropic’s AI: Crusoe says the $15B Abilene site, which serves OpenAI per OpenAI's own post, depended on capital partners Blue Owl and Primary Digital. That cuts against 'entirely funded by hyperscalers,' though the ultimate obligations are unclear.
How we judged it
- Kind of claim
- Observation
- Deadline
- None given. We don’t add one.
Where we had to interpret
- The firm assertion is that hyperscalers supplied all funding for the physical infrastructure serving the two labs.
- The plain reading distinguishes funding that infrastructure from raising equity for the labs. It does not automatically equate paying a cloud-service invoice with financing construction.
- The $250 billion figure is explicitly an estimate. Establishing that exact realized total is not a condition for supporting the categorical funding assertion.
- The nearly $300 billion raised and Microsoft's reported spending are premises, not amounts that may simply be added to produce an infrastructure total.
- The historical interval ends at publication. Its start remains unspecified because last few years does not establish a particular year.
- Oracle is not excluded merely because it is smaller than the three providers named later. CoreWeave's classification and the treatment of outside capital providers remain explicit ambiguities.
Right if: Records covering the relevant infrastructure establish that hyperscalers supplied its funding throughout the described period, with no material independent infrastructure funding. The hedged $250 billion estimate need not equal an audited total.
Wrong if: Records identify infrastructure actually serving either lab that was independently funded outside the hyperscaler group, with enough information to distinguish independent funding from a hyperscaler's borrowing, lease obligations or financing arrangements. A cloud invoice or an unidentified lender alone is insufficient.
We wrote this test after the outcome was known.
What would settle it
Project-level records identifying capital contributions, tenant guarantees and payment obligations for infrastructure serving both labs, with an explicit and consistently applied meaning of funded and hyperscaler.
We found 45 claims in this post. See them all.
Full record
Source and dates
Where's Your Ed At, The OpenAI Bubble, paragraph 32.
- 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
While OpenAI and Anthropic have raised nearly $300 billion in the last few years, their actual infrastructure costs — the GPUs and the data centers to power their services — were entirely funded by hyperscalers, likely costing another $250 billion in the process
- Where
- paragraph 32
- 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.