“AI does not have a “profit lever,” because the raw costs of providing access to AI models are so high (and they’re only increasing) that the basic economics of how the tech industry sells software don’t make sense.”
Our reading, the sentence that was checked:
AI businesses have no profit lever because model access costs are high and rising.
Assessment: Contradicted Evidence as of
Where and when Ed Zitron wrote it
Where's Your Ed At, The Case Against Generative AI, paragraph 207.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment wyea-the-case-against-generative-ai-2a33b7-a1
We recorded 46 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 2025-09-29. 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)
From the speaker
- AI does not have a 'profit lever'
- raw costs of providing access to AI models are so high
- they're only increasing
- basic economics of how the tech industry sells software don't make sense
Our interpretation
- A 'profit lever' is the Silicon Valley pattern he describes: grow fast and burn money, then turn profitable as revenue outgrows costs.
- Plain reading: AI model and software businesses cannot reach operating profit as they scale, because serving costs stay too high relative to revenue.
- A lever can only be tested by whether it later works. So the test window runs from the post date to the evaluation date.
- The universal claim ('AI does not have') is contradicted if a leading AI model business reaches operating profit through scale, with revenue outgrowing costs.
- Absolute compute spending rising does not by itself support the claim, if costs fall as a share of revenue.
Counts as supported if: Leading AI businesses stay at operating losses as they scale, and serving costs stay at or above revenue, or keep rising as a share of it.
Counts as contradicted if: A leading AI model provider reaches GAAP operating profit because revenue growth outpaced cost growth, with compute costs falling as a share of revenue.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Does the evidence support the cause?
Outcome: Contradicted
The dated evidence goes against what the claim committed to, under the criteria frozen before grading.
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
Zitron argued that AI has no 'profit lever' because serving costs are too high and only rising. Anthropic's S-1 shows the lever working. Its operating loss narrowed each quarter after Q3 2025, and it earned $1.0 billion of GAAP operating income in Q2 2026 'as revenue growth outpaced the growth in total operating expenses.' Compute fell from about 201% of revenue in H1 2025 to 55% in H1 2026. Absolute compute spending did rise, so that narrow point holds. But the conclusion that AI economics cannot turn profitable is contradicted for a leading model developer. Net losses continue, mostly from non-cash charges.
Limitations
- The result turns on reading 'profit lever' as reaching operating profit through scale.
- The evidence covers one company. Software resellers like Replit and Cursor may differ.
- The operating profit covers one quarter, and Anthropic says results will fluctuate.
Evidence (3 items)
Goes against the claim: Anthropic, PBC (SEC Form S-1),
we generated operating income of $1.0 billion for the three months ended June 30, 2026, as revenue growth outpaced the growth in total operating expenses.
Why it matters here: The second-largest model developer reached GAAP operating income because revenue outgrew costs. This is the 'profit lever' the claim says AI lacks.
- Publisher
- Anthropic, PBC (SEC Form S-1)
- Kind of source
- Regulatory filing, unaudited
- Measures
- Anthropic, operating income, USD billions, gaap, 2026-04-01 to 2026-06-30
- Describes events of
- Public from
Goes against the claim: Anthropic, PBC (SEC Form S-1),
The result of our revenue growth, compute efficiency, and operating efficiency has been a significant improvement in Adjusted Operating Income, which was $1.6 billion for the six months ended June 30, 2026
Why it matters here: Anthropic credits revenue growth and compute efficiency for moving from large losses to near break-even GAAP operations and positive adjusted operating income. Its percentage-of-revenue table shows compute falling from 201% of revenue in H1 2025 to 55% in H1 2026.
- Publisher
- Anthropic, PBC (SEC Form S-1)
- Kind of source
- Regulatory filing, unaudited
- Measures
- Anthropic, operating loss and adjusted operating income, USD millions, gaap, 2026-01-01 to 2026-06-30
- Describes events of
- Public from
Context only: Anthropic, PBC (SEC Form S-1),
we may not achieve or sustain profitability in the future.
Why it matters here: Net losses continue, but they are driven mainly by non-cash fair-value charges. Anthropic itself warns it may not sustain profitability. This limits how far the operating-profit evidence reaches.
- Publisher
- Anthropic, PBC (SEC Form S-1)
- Kind of source
- Regulatory filing
- Describes events of
- Public from
Where it was said (1)
Said once in the sources this register covers.
: Where's Your Ed At, The Case Against Generative AI
AI does not have a “profit lever,” because the raw costs of providing access to AI models are so high (and they’re only increasing) that the basic economics of how the tech industry sells software don’t make sense.
- Where
- paragraph 207
- Published
- 2025-09-29
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment wyea-the-case-against-generative-ai-2a33b7-a1: Current. Outcome: Contradicted. 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-the-case-against-generative-ai-2a33b7-a1.