Anthropic says in a revised filing that it may soon reach its own threshold for AI-driven accelerated progress

What happened

Anthropic, the company behind the Claude AI models, filed a revised S-1A on Monday, November 2, 2026. TechCrunch reported that much attention focused on the updated financials, which show rising revenue alongside rising expenses. TechCrunch also highlighted new language in the filing's Risk Factors section about AI systems that help build more capable AI.

According to TechCrunch, the filing appears to warn that the company is close to creating a model capable of recursive self-improvement, often shortened to RSI. The company also said that none of its current models have crossed that line.

What the filing says

AI is already part of Anthropic's research

The filing says Anthropic uses AI in its own research and development. It cautions that this practice could pose a "considerable safety risk," especially if it sharply speeds up progress. The company said this risk falls under its Responsible Scaling Policy, which lays out how it plans to monitor and reduce risks as it approaches recursive self-improvement.

Past jumps that went unnoticed

TechCrunch reported that Anthropic added a statement that its growing use of AI could advance model capabilities faster than the company can build and test matching safety, alignment and security measures. The filing says the company has in the past failed to spot a significant acceleration in its pace of development until after it happened. It adds that the same thing could happen again in the future.

A threshold that may be close

The filing says the company's internal estimates and expectations about future models indicate it may soon meet its policy's definition of "dramatically accelerated progress attributable to AI." Anthropic said those estimates concern future models, not current ones. It also said it has not designated any current model as exceeding the threshold and is still assessing the underlying data.

Keeping humans in the loop

Anthropic said in the filing that it does not expect to be able to safely let Claude fully design and build later generations of frontier models on its own "without meaningful human supervision or understanding." That holds even if such autonomy becomes technically possible, the company said.

For that reason, the company said it is considering updates to its Responsible Scaling Policy "to keep humans in control of frontier model development at Anthropic," TechCrunch reported. TechCrunch said the company is promising to keep people involved in its research process for the foreseeable future.

Why RSI draws attention

TechCrunch describes RSI as the point where AI becomes so good at training itself that it triggers a dramatic acceleration in AI capabilities. People who fear AI see it as a point of no return, where AI could move beyond human control. Supporters see some version of it as the technology's promise, a source of major gains in productivity and scientific discovery.

According to TechCrunch, Anthropic appears to view RSI as both a promise and a danger. The outlet noted that when the term caught on earlier this year, many people dismissed it as a buzzword or a goal too hard to reach, define or prove.

Questions about the disclosure

TechCrunch noted that skeptics could argue an S-1 works as both a disclosure and a marketing document. In that view, claims about being close to a self-building model could be partly promotional. TechCrunch also reported that Claude's abilities in coding, math and cybersecurity have grown sharply this year. More coverage of artificial intelligence and the companies building it is in our AI and tech section, including a look at how OpenAI may rebrand its data centers.

What to watch next

• Whether Anthropic updates its Responsible Scaling Policy to keep humans in control of frontier model development, which the filing says the company is considering.
• Results of Anthropic's ongoing assessment of the data behind its estimates, since it has not designated any current model as exceeding its threshold.
• Whether future Anthropic models meet the policy's definition of "dramatically accelerated progress attributable to AI," which the filing says may happen imminently.