Pundit ClaimsAI claims by Gary Marcus and Ed Zitron, with the evidence

“The capacity never existed, this deal will never happen, and you all know it and are either hopelessly ignorant or craven beyond words.”

The OpenAI Bubble, paragraph 135.

Our reading:

The OpenAI-driven Oracle compute deal behind Oracle's RPO will never happen; the capacity never existed.

Assessment: Unresolved

What happened

Did the forecast come true?

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

On July 22, 2025, OpenAI reported operating capacity and early workloads at Abilene, while describing additional Oracle capacity separately [e1]. On September 10, 2026, Oracle reported delivering 131,000 GPUs at Abilene during fiscal Q1, June 1–August 31, 2026 [e2]. The Abilene operations challenge a broad denial that any Oracle–OpenAI capacity existed [e1]. Neither disclosure connects both substantive delivery and payment to the disputed agreement or documents its definitive termination without performance [e1] [e2]. Neither component is settled.

Caveats

  • The July 2025 disclosure distinguishes Abilene from the additional Oracle capacity and does not establish that Abilene performed the disputed agreement.
  • Oracle's delivery report does not verify payment under that agreement.
  • The June–August 2026 delivery total does not identify how much was delivered before the July 15, 2026 post.
  • The evidence does not establish whether the historical capacity assertion concerns Abilene or the additional promised buildout.

Evidence (2 items)

  • Context only: OpenAI,

    up and running.

    Why it matters: This challenges a broad reading that no Oracle–OpenAI capacity existed. It does not establish that Abilene constituted performance of the particular additional-capacity agreement.

    Publisher
    OpenAI
    Kind of source
    Company disclosure: the company describing itself, not an independent finding
    About events on
    Published
  • Context only: Oracle, transcript hosted by MarketBeat,

    We delivered 131,000 GPUs

    Why it matters: Management reports substantial Abilene delivery during fiscal Q1. Delivery at this site does not by itself establish the contract attribution and payment required by the resolution rule.

    Publisher
    Oracle, transcript hosted by MarketBeat
    Kind of source
    Company disclosure: the company describing itself, not an independent finding
    Measures
    Oracle, GPUs delivered at Abilene GPUs, company reported, 2026-06-01 to 2026-08-31
    About events on
    Published

How we judged it

Kind of claim
Forecast
Deadline
None given. We don’t add one.

Where we had to interpret

  • The main forecast concerns performance of the Oracle–OpenAI compute agreement identified in the context, not merely signing an agreement.
  • Actual delivery and paid use attributable to that agreement would refute the ordinary meaning of 'will never happen'; completion of every promised facility is not required.
  • The historical capacity assertion concerns capacity associated with that agreement before the post. Capacity under a separate or earlier agreement cannot automatically settle it.
  • The surrounding text supplies a $300 billion contract figure but does not identify its accounting or reporting basis.

Right if: Definitive termination or extinguishment of the identified agreement without substantive delivery or paid use, together with records establishing that its promised capacity had not existed before the post.

Wrong if: Documented delivery and payment for compute under the identified agreement, or direct evidence that the capacity denied by the historical assertion already existed.

We wrote this test after the outcome was known.

What would settle it

Identify the disputed agreement. Verified substantive delivery and payment under it would contradict the forecast under the early-resolution rule. Records tying operational capacity before the post to that agreement would contradict the historical assertion. Supporting both components requires definitive extinguishment without performance and records showing that the promised capacity had not existed before the post.

We found 41 claims in this post. See them all.

Full record

Source and dates

Where's Your Ed At, The OpenAI Bubble, paragraph 135.

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

    The capacity never existed, this deal will never happen, and you all know it and are either hopelessly ignorant or craven beyond words.

    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

  1. PublishedBy us

    Published after both AI reviews and the final check passed.

  2. Held backBy us

    The final check failed: a cited source didn’t say what the write-up said.