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

“But it’ll have to at some point unless they intend to write the GPUs off.”

Where're All The AI Chips?, paragraph 120.

Our reading:

Hyperscaler depreciation will have to spike at some point unless they intend to write the GPUs off.

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 29, 2026, Meta reported that higher depreciation contributed to infrastructure cost growth in Q2 2026 [e1]. This predates the September 22 forecast and does not establish a subsequent sharp increase attributable to the relevant GPUs [e1]. It also does not establish retention for use or write-off intentions across Amazon, Alphabet, Microsoft, and Meta [e1]. The evidence therefore leaves the conditional forecast unresolved [e1].

Caveats

  • The forecast has no deadline or numerical definition of a depreciation spike.
  • The relevant GPU assets are not precisely identified.
  • The evidence does not establish whether the write-off exception applies.
  • The packet contains no verified post-publication disclosure meeting the early-resolution rule.

Evidence (1 item)

  • Context only: Meta Platforms,

    higher depreciation

    Why it matters: Meta reported depreciation-related cost pressure before the post. This establishes neither a subsequent group-wide spike nor intentions concerning the GPUs in question.

    Publisher
    Meta Platforms
    Kind of source
    Company disclosure: the company describing itself, not an independent finding
    About events on
    Published

How we judged it

Kind of claim
Conditional forecast
Deadline
“at some point”

Conditions

  • unless they intend to write the GPUs off (their words)

Where we had to interpret

  • Hyperscalers means Amazon, Alphabet, Microsoft and Meta, the four companies discussed in the surrounding passage.
  • The ordinary reading is a sharp increase in reported depreciation as the discussed GPU investments enter service. No numerical threshold or deadline is supplied.
  • The exception concerns an intention to write off the relevant GPUs. Silence about write-offs does not establish the absence of that intention.
  • The financial measure is company-reported depreciation expense, in the companies' reporting currency of US dollars.

Right if: Dated disclosures establish that the relevant GPUs are being retained for use and that their entry into service produces a sharp increase in depreciation across the named hyperscalers.

Wrong if: The relevant GPU investments complete their use or disposition without a sharp depreciation increase and without the write-off exception applying. A period without a spike is insufficient.

We wrote this test after the outcome was known.

What would settle it

Post-publication disclosures establishing that the relevant GPUs are retained for use and that entering service causes a sharp depreciation increase across the named companies. Contradiction would require completed use or disposition of those assets without that increase and without the write-off exception applying.

We found 53 claims in this post. See them all.

Full record

Source and dates

Where's Your Ed At, Where're All The AI Chips?, paragraph 120.

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 2026-09-22.

Where it was said (1)

Said once in the posts we read.

  • : Where's Your Ed At, Where're All The AI Chips?

    But it’ll have to at some point unless they intend to write the GPUs off.

    Published
    2026-09-22
    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

    Both reviewers disagreed with the draft’s verdict.