“Anthropic and OpenAI make up the vast majority of all AI revenues and compute capacity. I estimate 70% of all revenues and capacity demand, if not higher.”
Our reading, the sentence that was checked:
Anthropic and OpenAI make up the vast majority of all AI revenues and compute capacity, which he estimates at 70% of all revenues and capacity demand, if not higher.
Assessment: Unresolved Evidence as of
Where and when Ed Zitron wrote it
Where's Your Ed At, Am I Meant To Be Impressed?, paragraph 98.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment wyea-am-i-meant-to-be-impressed-3a7559-a1
We recorded 45 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 2026-05-06. 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)
Conditions
- author's estimate (qualifier, stated by the speaker)
From the speaker
- Anthropic and OpenAI
- vast majority of all AI revenues and compute capacity
- I estimate 70% of all revenues and capacity demand, if not higher
Our interpretation
- Assess a combined share of approximately 70% or more separately for revenue and compute; the two shares cannot be averaged.
- The ordinary reading concerns the AI industry broadly. A narrower frontier-model, enterprise-API or selected-cloud market cannot silently replace all AI.
- Use January–March 2026, the latest completed reporting quarter before the post, for the ongoing observation.
- Revenue means realized revenue on a consistent market perimeter. Compute means operational capacity attributable to the labs, with contracted future demand reported separately.
- The author's estimate is a confidence qualifier, not a conditional antecedent.
- The contextual heading's 75% GPU figure is not substituted for the 70% threshold in the selected claim.
Counts as supported if: Comparable industry-wide estimates place both labs' combined realized revenue share and operational compute share at approximately 70% or higher during the target period.
Counts as contradicted if: A defensible, comparable industry denominator places either combined share materially below 70%, allowing for estimation uncertainty.
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
Independent compute estimates materially challenge the claimed concentration: the retrieved December 2025 estimates put the two labs well below 70%. They cannot settle a January–March 2026 observation without evidence for the intervening change. The search also found a potentially relevant revenue study, but read_page did not retrieve its substantive data. Neither component has a complete, matching-period evidentiary record.
Limitations
- The compute counterevidence is explicitly retained as context because its measurement period differs.
- The compute estimates have substantial uncertainty.
- A narrower LLM market would not automatically measure all AI revenue.
- The metric_scope covers revenue; compute is a separate nonfinancial measurement.
Evidence (2 items)
Context only: Epoch AI, Gradient Updates; Josh You,
not using
Why it matters here: The analysis estimates that even five leading labs together used less than half of global operational AI compute at the end of 2025. This materially challenges the broad concentration premise, but does not measure January–March 2026.
- Publisher
- Epoch AI, Gradient Updates; Josh You
- Kind of source
- Independent measurement
- Describes events of
- Public from
Context only: Epoch AI,
in 2025
Why it matters here: The table's median estimates put the two labs at 47.9% of compute among five leading labs alone. This is material earlier-period counterevidence, with substantial estimation uncertainty and no direct first-quarter measurement.
- Publisher
- Epoch AI
- Kind of source
- Independent measurement
- Arithmetic
- 100 × (1,743,000 + 1,190,000) / (1,743,000 + 1,190,000 + 1,583,000 + 996,000 + 615,000) (inputs: OpenAI: 1,743,000 H100 equivalents; Anthropic: 1,190,000 H100 equivalents; Google DeepMind: 1,583,000 H100 equivalents; Meta Superintelligence Labs: 996,000 H100 equivalents; SpaceXAI: 615,000 H100 equivalents; 47.9%, rounded to one decimal place; ratio of reported medians, not a confidence interval.)
- Describes events of
- Public from
What would settle it
Accessible first-quarter industry revenue data with consistent accounting and a matching compute allocation dataset, including the rest of the AI market and explicit uncertainty bounds.
Where it was said (1)
Said once in the sources this register covers.
: Where's Your Ed At, Am I Meant To Be Impressed?
Anthropic and OpenAI make up the vast majority of all AI revenues and compute capacity. I estimate 70% of all revenues and capacity demand, if not higher.
- Where
- paragraph 98
- Published
- 2026-05-06
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment wyea-am-i-meant-to-be-impressed-3a7559-a1: Current. Outcome: Unresolved. Evidence as of 2026-10-28. Public date not recorded.
Recorded events
Assessment publishedBy this site
First assessment published after both reviews and the final-text check passed.