“On the back end, AI startups are annihilating cash, with up until recently Anthropic allowing you to burn upwards of $8 in compute for every dollar of your subscription. OpenAI allows you to do the same, though it’s hard to gauge by how much.”
Our reading, the sentence that was checked:
AI startups are annihilating cash: until recently Anthropic let subscribers burn upwards of $8 in compute for every dollar of their subscription, and OpenAI lets them do the same, though it is hard to gauge by how much.
Assessment: Unresolved Evidence as of
Where and when Ed Zitron wrote it
Where's Your Ed At, AI's Economics Don't Make Sense, paragraph 21.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment wyea-ais-economics-dont-make-sense-cc2bd2-a1
We recorded 31 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 2026-04-28. 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
- up until recently (Anthropic) (qualifier, stated by the speaker)
- hard to gauge by how much (OpenAI) (qualifier, stated by the speaker)
From the speaker
- up until recently
- Anthropic allowing you to burn upwards of $8 in compute for every dollar of your subscription
- OpenAI allows you to do the same
- hard to gauge by how much
Our interpretation
- The ordinary reading is permitted heavy-user consumption, not average consumption across subscribers.
- Anthropic's threshold is at least approximately eight dollars of actual compute expense per subscription dollar.
- For OpenAI, 'the same' means subsidized usage exceeding subscription fees; the following qualification does not commit OpenAI to the same eight-to-one ratio.
- The surrounding explanation identifies compute as what the provider pays to serve the model. API retail-equivalent value is a different measure.
- Use the recent pre-publication period, January 1 through April 28, 2026, as an explicit interpretation of 'up until recently'.
- The linked independent analysis makes the Anthropic figure an estimate, not a company-reported accounting figure.
Counts as supported if: Dated usage and actual-cost measurements show that Anthropic permitted approximately eight or more dollars of compute expense per subscription dollar, and OpenAI permitted compute expense above subscription fees, during the relevant period.
Counts as contradicted if: Complete permitted-usage and actual-cost records rule out the Anthropic threshold or show that OpenAI's permitted usage could not exceed subscription fees. An API-price mismatch alone does not prove the actual-cost assertion false.
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 linked analysis substantiates a narrower API-price comparison: Claude allowances were valued at roughly eight or more times subscription fees. It does not establish that Anthropic incurred that much compute expense. Earlier reporting supports OpenAI Pro losses, but not the specified 2026 state. The metric mismatch leaves the actual-cost claim unresolved rather than disproved.
Limitations
- The independent allowance analysis is not a provider-cost audit.
- API retail prices cannot be substituted for internal compute expenses.
- The OpenAI evidence predates the target period.
Evidence (2 items)
Context only: she-llac.com,
Compared to API pricing
Why it matters here: The table estimates $163 of API-equivalent usage for a $20 Pro subscription, labeled 8.1 times, and 13.5 times for Max plans. This explains the cited figure but does not measure Anthropic's actual compute expense.
- Publisher
- she-llac.com
- Kind of source
- Independent measurement
- Measures
- Anthropic, API retail-equivalent value of estimated subscription allowances divided by subscription price, USD USD of API-equivalent value per USD of subscription fees, estimate, 2026-01-01 to 2026-01-31
- Describes events of
- Public from
Context only: Fortune,
losing money
Why it matters here: This is an executive acknowledgment of losses associated with Pro usage in early 2025. It is outside the contracted period and supplies neither a compute-cost ratio nor evidence of the 2026 state.
- Where
- Article opening
- Publisher
- Fortune
- Kind of source
- News report
- Measures
- OpenAI — ChatGPT Pro, Subscription-plan losses, definition unspecified, USD, company reported, ? to 2025-01
- Describes events of
- Public from
What would settle it
Historical allowance measurements paired with actual or defensibly estimated serving costs for Anthropic and OpenAI in the relevant pre-publication period.
Where it was said (1)
Said once in the sources this register covers.
: Where's Your Ed At, AI's Economics Don't Make Sense
On the back end, AI startups are annihilating cash, with up until recently Anthropic allowing you to burn upwards of $8 in compute for every dollar of your subscription. OpenAI allows you to do the same, though it’s hard to gauge by how much.
- Where
- paragraph 21
- Published
- 2026-04-28
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment wyea-ais-economics-dont-make-sense-cc2bd2-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.
About assessment wyea-ais-economics-dont-make-sense-cc2bd2-a1.