“Every AI startup is effectively an extraction mechanism for OpenAI and Anthropic, reselling their tokens through their services at a massive discount because users are unwilling to pay the actual cost of using LLMs.”
Our reading, the sentence that was checked:
Every AI startup is effectively an extraction mechanism for OpenAI and Anthropic, reselling their tokens at a massive discount because users are unwilling to pay the actual cost of using LLMs.
Assessment: Unresolved Evidence as of
Where and when Ed Zitron wrote it
Where's Your Ed At, Annualized Enron Rate, paragraph 76.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment wyea-annualized-enron-rate-9c57e6-a1
We recorded 28 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 2026-10-13. This page does not claim the wording was captured then.
What the claim commits to
- Kind
- Causal claim
- 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
- every AI startup (scope, stated by the speaker)
- effectively (qualifier, stated by the speaker)
From the speaker
- Every AI startup
- an extraction mechanism for OpenAI and Anthropic
- reselling their tokens
- at a massive discount
- because users are unwilling to pay the actual cost
Our interpretation
- In context, AI startup means an AI application business using these providers' models, rather than every hardware, robotics, or model-training company.
- The plain reading is that each such startup materially subsidizes customers' token consumption because customers will not pay enough to cover the supplier cost.
- The comparison is aggregate realized customer payments against actual supplier charges after discounts, not the maximum usage permitted to one heavy user.
- The latest completed quarter before publication, July through September 2026, is used for current operating economics.
- The surrounding comparisons use reported margins and estimated token costs, so the financial basis is estimate.
Counts as supported if: Records covering the relevant startup population show material token resale subsidies at every firm, together with demand or pricing evidence establishing insufficient willingness to pay as the cause.
Counts as contradicted if: A relevant startup demonstrably covers its actual token supplier costs through customer payments without the claimed material subsidy, or reliable evidence disproves the asserted causal explanation.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Does the evidence support the cause?
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
Cost-recovery pricing and alternative-model development challenge the universal description. But the readable records do not establish a target-quarter counterexample on matched customer receipts and actual supplier charges. They also do not establish the claimed explanation about willingness to pay. The universal subsidy claim remains unproven; the available counterevidence is insufficient for a definitive contradiction under the contextual reading.
Limitations
- A literal reading covering unrelated AI businesses would be broader than the surrounding argument.
- Positive gross margin, token cost recovery, and overall profitability are different measures.
- The inaccessible margin report is excluded.
Evidence (2 items)
Context only: Cursor,
additional usage at cost.
Why it matters here: Cursor disclosed an arrangement allowing customers to pay for additional usage at cost, challenging a blanket account of unavoidable subsidies. Its date and limited coverage prevent treating it as proof of the startup's aggregate economics in the target quarter.
- Publisher
- Cursor
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- Measures
- Cursor, Price charged for additional model usage relative to stated cost Cost pass-through, company reported, 2025-07-04 to 2025-07-04
- Describes events of
- Public from
Context only: Harvey,
Kimi K3
Why it matters here: Harvey documents development using a model outside OpenAI and Anthropic. A research preview does not establish how much production usage changed or whether customer receipts covered token costs.
- Publisher
- Harvey
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- Describes events of
- Public from
What would settle it
Verified target-quarter billing and cost records for a relevant counterexample, or comprehensive evidence for the universal claim, plus evidence testing demand at cost-covering prices.
Where it was said (1)
Said once in the sources this register covers.
: Where's Your Ed At, Annualized Enron Rate
Every AI startup is effectively an extraction mechanism for OpenAI and Anthropic, reselling their tokens through their services at a massive discount because users are unwilling to pay the actual cost of using LLMs.
- Where
- paragraph 76
- Published
- 2026-10-13
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment wyea-annualized-enron-rate-9c57e6-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.