Anthropic tells IPO investors its AI may "imminently" cross its own threshold for speeding up AI research

Anthropic has told prospective shareholders that it may soon meet its own definition of AI that dramatically speeds up AI research. The sentence is new: it appears in the amended S-1 filed with the SEC on 2 November and is absent from the version filed a week earlier. The same amendment fills in two numbers the first filing left blank: $17.0 billion of revenue in the third quarter, and about $895.4 billion of commitments, almost all for compute.

What did Anthropic add to its amended IPO filing?

One sentence about how close its models are to automating AI research, and the financial figures for July to September. In the risk factor on fast-moving capabilities, the amendment says that "our internal estimates and our expectations regarding future models, rather than our current models, indicate that we may imminently meet our Responsible Scaling Policy's definition of 'dramatically accelerated progress attributable to AI.'" A search of the 26 October filing finds neither "imminently" nor that phrase.

A second change moves a warning into the past tense. The first filing said Anthropic "may not identify a significant acceleration in our rate of development until after it has occurred." The amendment says it has "in the past not identified, and may in the future not identify" such an acceleration until after the fact. The surrounding text explains why: much of the evidence "is derived from lagging indicators."

The amendment also says Anthropic is "considering additional measures, including updates to our Responsible Scaling Policy, to require human review and approval of certain decisions in our research and development processes as our models perform an increasing share of that work." That sentence is new as well.

What is Anthropic's AI research threshold?

It is one of the capability lines in Anthropic's Responsible Scaling Policy, the document that sets out what safeguards the company commits to as its models grow more capable. Version 3.4, in effect since 8 July 2026, says the threshold is met if either "our models would be able to fully substitute for our entire set of Research Scientists and Research Engineers, at competitive costs (i.e., within a factor of 5)," or there is "dramatic acceleration" of the pace of AI progress "for reasons that likely relate to the automation of AI R&D."

The policy puts a number on the second test. It counts as met when Anthropic observes or expects double the rate of progress in aggregate AI capabilities, compared with both the expected rate and the fastest sustained rate seen without significant AI help, and it is plausible that the doubling is substantially due to automated research. A footnote spells out the measure: "as much progress in one year as one would see in two years at baseline," which it says "is not the same idea as 'doubling researchers' productivity.'"

The filing's sentence points at the second test, not the first. The phrase it quotes, "dramatically accelerated progress attributable to AI," is the filing's own paraphrase; the policy text itself says "dramatic acceleration."

Has Anthropic crossed the threshold?

No, according to the filing. "We have not designated any of our current models as exceeding this threshold, and we continue to assess the underlying data," it says. The expectation rests on internal estimates and on "future models, rather than our current models." Anthropic adds that it is "evaluating the implications of meeting the AI research and development capability threshold and will release further reports in the future as we set forth in our Responsible Scaling Policy."

The filing gives no date and names no model. Elsewhere it repeats a line from the first filing, that "our models are contributing to the pace of our own research," and warns that the risk grows "as AI systems substantially automate AI research and development." We covered the earlier risk language in our item on the S-1 and the leaked draft in our item on the prospectus.

How much revenue did Anthropic make in the third quarter?

$17.0 billion in the three months to 30 September, against $1.4 billion in the same quarter of 2025. Second-quarter revenue was $11.6 billion, so the quarter grew by about 47% on the one before, by our arithmetic from the filed figures. Revenue for the first nine months of 2026 was $33.2 billion, up from $2.5 billion a year earlier.

The quarter turned an operating profit into a loss. Anthropic reported an operating loss of $567 million, against operating income of $974 million in the second quarter, "primarily driven by $5.0 billion of compute under short-term commitments." Short-term compute, the filing says, "currently costs significantly more than comparable compute procured under our multi-year commitments." On its own adjusted measure, which strips out some costs, Anthropic reports a 15% operating margin for the quarter. The Next Web led its coverage of the amendment with the revenue and commitment figures.

How much has Anthropic committed to spend on compute?

About $895.4 billion. "As of and subsequent to September 30, 2026, we have entered into purchase commitments and lease agreements," the amendment says, "with an aggregate committed spend of approximately $895.4 billion, substantially all of which is related to compute and infrastructure." About $447.5 billion falls before the end of 2030 and nearly all the rest between 2031 and 2045. About $69.8 billion is due in the twelve months after 30 September.

In the 26 October filing the same sentence read "an aggregate committed spend of more than $" followed by a blank. For scale, the commitment is roughly 27 times the revenue Anthropic booked in the first nine months of 2026, again by our arithmetic.

When is the Anthropic IPO?

Still not set. The amendment leaves the price range and the number of shares blank, and the shares cannot be sold until the SEC declares the registration effective. The listing is planned on the Nasdaq Global Select Market under the ticker ANT. Reported dates and valuations come from people familiar with the matter rather than from the filing; the price range would normally appear in a later amendment. This item reports what the filings say and is not investment advice.