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

“Hyperscalers are getting increasingly worse at turning their capex into operational capacity.”

Where're All The AI Chips?, paragraph 116. The full statement is at the source.

Our reading, the sentence that was checked:

Hyperscalers are getting increasingly worse at turning capex into operational capacity.

Assessment: Unresolved Evidence as of

Where and when Ed Zitron wrote it

Where's Your Ed At, Where're All The AI Chips?, paragraph 116.

Said
Source observed
live page, full text, public
Evidence as of
the cutoff for this assessment
Assessment recorded
assessment wyea-wherere-all-the-ai-chips-53487e-a1

We recorded 35 claims from this post. See every claim from this post.

The wording above is the version this site observed on 2026-10-28. The source may have been edited after 2026-09-22. This page does not claim the wording was captured then.

What the claim commits to

Kind
Observation
Deadline, as the speaker wrote it
None stated. This site does not add one.
Deadline state when assessed
No horizon
Deadline state on 2026-10-29
No horizon (the release’s reference date)

From the speaker

  • Hyperscalers
  • getting increasingly worse
  • turning their capex into operational capacity

Our interpretation

  • Use Microsoft, Alphabet, Amazon and Meta from the surrounding discussion.
  • The ordinary reading is a deteriorating conversion process: comparable spending takes longer to become operational, or less becomes operational at a comparable age.
  • The endpoint is the quarter ended June 30, 2026, the latest completed reporting quarter publicly reportable before the post.
  • The historical baseline is unspecified and remains open rather than being selected to fit the evidence.
  • A falling D&A-to-capex ratio is a proposed proxy, not the outcome itself.

Counts as supported if: Comparable historical spending cohorts show a broad deterioration in commissioning time or the share operational at a fixed age, after accounting for asset mix and cost changes.

Counts as contradicted if: Comparable records show stable or improving conversion across the group. An improvement at one company is material counterevidence but does not settle a group-level trend.

Retrospective: these criteria were written after the outcome was public.

Current assessment

Was it true for the period described?

Outcome: Unresolved

The record was checked, and the evidence available by the cutoff does not settle it.

Scored on a plain reading. The speaker named no measure for this claim, so it was judged on the reading an informed reader would take, stated below under our interpretation. Both reviews had to find that reading fair.

Evidence as of
Deadline state when assessed
No horizon
Assessment recorded
Policy version
2026-10-28.7

Rationale

The broad deterioration claim is not established by the records reviewed. Microsoft reports nearly halving GPU dock-to-live times in its largest regions, which points in the opposite direction for one deployment stage. That does not resolve the whole industry or the time from initial spending to operation. Accounting ratios and asset balances do not supply the missing comparable commissioning history.

Limitations

  • The direct operational evidence covers one company and one deployment stage.
  • Different readings of conversion efficiency could produce different results.

Evidence (1 item)

  • Context only: Microsoft,

    Over the last fiscal year, we have reduced dock-to-live times for new GPUs in our largest regions by nearly 50%.

    Why it matters here: This is material counterevidence to a general deterioration claim. It covers Microsoft GPU installation in its largest regions, not the full spending-to-service cycle or all hyperscalers.

    Publisher
    Microsoft
    Kind of source
    Company disclosure: the company describing itself, not an independent finding
    Describes events of
    Public from

What would settle it

Historical commissioning records for comparable spending cohorts across the four companies, with a stated baseline and consistent treatment of asset costs and deployment stages.

Where it was said (1)

Said once in the sources this register covers.

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

    Hyperscalers are getting increasingly worse at turning their capex into operational capacity.

    Published
    2026-09-22
    Source observed
    , live page
    Access
    full text, public

History

Assessment versions

Recorded events

  1. Assessment publishedBy this site

    First assessment published after both reviews and the final-text check passed.

    About assessment wyea-wherere-all-the-ai-chips-53487e-a1.