“Its models, again like OpenAI, are unprofitable, unsustainable and heavily-dependent on training data that's either running out or has already run out.”
Our reading:
Anthropic's models, like OpenAI's, are unprofitable, unsustainable and dependent on depleting training data.
Assessment: Unresolved
What happened
Was it true for the period described?
Outcome: Unresolved
We checked, and the evidence so far doesn’t settle it.
Judged on a plain reading. They named no measure, so we used the reading a typical informed reader would take. It is spelled out under “How we judged it”.
Evidence up to . Checked .
Why
Anthropic's filing substantiates a history of company losses. It does not establish the same metric as model profitability, and neither company losses nor data-scarcity research settles the full compound assertion. The evidence does not establish Anthropic's remaining training inputs or the OpenAI comparison. Calling the whole claim supported would substitute company finances for model economics and treat a general data concern as a company-specific finding.
Caveats
- The financial evidence measures consolidated GAAP losses, while the claim relies on estimated model economics.
- Unsustainable lacks a specified duration or financing threshold.
- Possible technical alternatives are not proof of adequate company-specific data supplies.
Evidence (2 items)
Context only: Anthropic, PBC / U.S. Securities and Exchange Commission,
We have a history of net losses
Why it matters: The filing includes audited 2024 consolidated losses, supporting the background concern about company finances. Consolidated GAAP losses do not establish the claimed estimated profitability of its models or inevitable financial failure.
- Publisher
- Anthropic, PBC / U.S. Securities and Exchange Commission
- Kind of source
- Regulatory filing, audited
- Measures
- Anthropic, Consolidated net loss, USD millions of dollars, gaap, 2024-01-01 to 2024-12-31
- About events on
- Published
Context only: arXiv,
synthetic data generation, transfer learning from data-rich domains, and data efficiency improvements might support further progress.
Why it matters: The research identifies possible alternatives to expanding human-written text datasets. This limits conclusions about unavoidable dependence on exhausted inputs, without proving that Anthropic had solved its data constraints.
- Publisher
- arXiv
- Kind of source
- Preprint, not peer reviewed
- About events on
- Published
How we judged it
- Kind of claim
- Observation
- Deadline
- None given. We don’t add one.
Where we had to interpret
- Its refers to Anthropic. The plain reading is that the model business does not cover its development and operating costs and depends on outside funding and scarce training inputs.
- Unsustainable is assessed as the current inability to fund the stated activity from its income, not an unstated prediction of inevitable bankruptcy.
- The surrounding discussion explicitly relies on reported 2024 financial results. Calendar 2024 is therefore the target period for the financial assessment.
- The financial basis is estimate because the context relies on press reporting. Company-wide cash burn and accounting losses are not interchangeable with model profitability.
- The ongoing data-supply clause concerns conditions at publication; the single target-period field cannot also represent that distinct February 2025 observation.
Right if: Comparable model-business revenue and fully allocated development and operating costs establish losses and dependence on external funding, while dated data-access evidence establishes material depletion of required training inputs. The OpenAI comparison also needs independent substantiation.
Wrong if: Comparable financial records establish that the model business covered the stated costs and financing needs, or data inventories establish ample required training inputs. A contradiction to one component must be identified separately.
We wrote this test after the outcome was known.
What would settle it
Comparable 2024 model-business revenues and allocated costs, financing requirements, and dated training-data inventories for Anthropic and OpenAI.
We found 52 claims in this post. See them all.
Full record
Source and dates
Where's Your Ed At, There Is No AI Revolution, paragraph 122.
- Said
- We read the source
- live page, full text, public
- Evidence up to
- Checked
We read this wording on 2026-10-28. The post may have been edited after 2025-02-24.
Where it was said (1)
Said once in the posts we read.
: Where's Your Ed At, There Is No AI Revolution
Its models, again like OpenAI, are unprofitable, unsustainable and heavily-dependent on training data that's either running out or has already run out.
- Where
- paragraph 122
- Published
- 2025-02-24
- We read it
- , live page
- Access
- full text, public
History
Versions
- Current: Unresolved. Evidence up to 2026-10-28.
What happened to this record
PublishedBy us
Published after both AI reviews and the final check passed.