Anthropic Says It May “Imminently” Reach a Self-Improving AI Threshold, and Investors Now Know
Anthropic’s amended IPO filing, published Monday, says the company may “imminently” meet its own definition of “dramatically accelerated progress attributable to AI,” a threshold widely read as a nod to recursive self-improvement (RSI). The company says none of its current models has crossed it and that humans will stay in the loop.
Date & Time: Tuesday, November 3, 2026 · 5:06 PM UTC
Anthropic’s amended IPO filing, published Monday, says the company may “imminently” meet its own definition of “dramatically accelerated progress attributable to AI,” a threshold widely read as a nod to recursive self-improvement (RSI). The company says none of its current models has crossed it and that humans will stay in the loop.
What Is RSI, in Plain English?
Recursive self-improvement is the point where AI gets so good at training itself that progress speeds up sharply, as TechCrunch describes it. Optimists see a tool for faster science and productivity. Doomers fear systems slipping beyond human control. It means AI doing a growing share of the work of building the next AI.
What the Filing Actually Says
The warnings
- Anthropic says using AI in its own research could pose a considerable safety risk if it sharply speeds up progress.
- It says capabilities may advance faster than its ability to build matching safety, alignment, and security measures.
- It says it has failed to spot a major acceleration in the past, and may again, until after it has happened.
- Its internal estimates and expectations for future models, not current ones, suggest it may soon hit the threshold. Per Gizmodo’s reading of the company’s policy, that means doubling its baseline rate of improvement, with AI plausibly responsible.
The reassurances
Anthropic says it has not designated any current model as exceeding the threshold and is still assessing the data. It also says it does not expect to safely let Claude build successive frontier models without meaningful human supervision, and is considering updates to its Responsible Scaling Policy to keep humans in control. The language appears in the amended SEC filing.
What Anthropic’s Leaders Have Said
In a June 2026 Reason interview, Anthropic co-founder Jack Clark described a future where humans step back and “Claude 10” builds “Claude 11.” He said that is not happening today, and that if he had to bet on a year, it would be late 2028. He also said the Anthropic Institute is working on ways to spot RSI and that the world should have the option to slow progress. That interview predates this filing.
Why It Matters for U.S. Readers
Anthropic is heading toward a public listing, as covered in our look at its IPO plans. Investors read risk factors closely, and this one is unusually blunt. It follows the other risk warnings in the same filing. TechCrunch argues that if Anthropic believes RSI is near, regulators and the public should take that seriously. Skeptics note that a filing is both disclosure and marketing, so bold claims deserve scrutiny.
Key Takeaways
- Anthropic says it “may” reach the threshold soon, not that it already has.
- Read it as a risk disclosure with marketing value, and watch for independent verification.
Conclusion
Anthropic’s filing does not announce self-building AI. It warns that AI-assisted research may soon outrun its safeguards. For U.S. readers, the debate over self-improving AI has moved from theory into securities filings.
Frequently Asked Questions
What does RSI mean?
RSI stands for recursive self-improvement: AI that helps build better versions of itself, speeding up progress.
Has Anthropic built a self-improving AI?
No. The company says none of its current models has been designated as crossing the threshold, though it may soon.
Does this mean AI is out of human control?
No. Anthropic says it plans to keep humans supervising frontier model development, and the statement is a risk disclosure, not a confirmed loss of control.
Author’s Review
What stands out to me is the tone. This is not a founder hyping a demo; it is a risk factor written for investors. Still, an IPO filing is also a sales document, so I would treat “close to self-building AI” as a serious signal, not a countdown clock. The part that matters most is the admission that Anthropic has missed acceleration before and could miss it again.
Elena Rostova
B2B Operations Consultant & Business Editor
Elena Rostova is a B2B strategy consultant specializing in digital operations, regulatory risk management, and scaling business models. She writes on corporate licensing, operational frameworks, and enterprise growth.
