A $17 billion quarter, now in the prospectus
Anthropic’s working headline is no longer a leak or a banker’s teaser: in an amended Form S-1/A filed with the SEC on November 2, 2026, the company reported $16.956 billion in revenue for the three months ended September 30, 2026 . That compares with $11.555 billion in the second quarter of 2026 and $1.382 billion in the third quarter of 2025, putting the latest quarter at roughly twelve times the year-earlier period . Fortune framed the same update as rapid growth with a visible slowdown, noting that quarterly revenue rose about 47% sequentially after a much faster second-quarter jump .
The timing matters because Anthropic is in the middle of an IPO process. The amended filing says the company has applied to list its Class A common stock on the Nasdaq Global Select Market under the ticker “ANT,” while the proposed share count and price range remain blank . AP reported that the filing still did not specify how much Anthropic expects to raise or where its shares will price when it goes public . In other words, investors now have a richer operating record, but not yet the valuation math that will define the offering.
Growth is huge; so are the losses
The headline revenue number is only the first layer. Anthropic reported a third-quarter net loss of $23.291 billion, even as revenue reached nearly $17 billion . AP said the company lost $23.3 billion on quarterly revenue of $17 billion and recorded a $104 billion net loss for the first nine months of 2026, much of it tied to accounting adjustments related to financing agreements . The SEC filing shows that a large portion of the third-quarter loss came below the operating line: a $21.936 billion negative change in the fair value of convertible notes, derivative liabilities and SAFEs .
That distinction matters. At the operating level, Anthropic swung to a $567 million loss in the third quarter after a $974 million operating profit in the second quarter . The company attributed the swing primarily to $5.0 billion of short-term compute that it said currently costs significantly more than comparable compute secured under multiyear commitments, plus $2.0 billion of non-cash charges tied to its donation matching program . On an adjusted basis, Anthropic still reported $2.463 billion of adjusted operating income for the third quarter, but its adjusted operating margin fell to 15% from 19% in the previous quarter .
The market question is therefore not whether Claude demand exists. It plainly does. The question is whether Anthropic can turn that demand into durable margins while racing to secure the infrastructure required to serve it.
The compute bill is the boss battle
Anthropic’s amended filing puts a number on the scale of that race: approximately $895.4 billion in purchase commitments and lease agreements as of and after September 30, 2026, substantially all related to compute and infrastructure . The filing says about $69.8 billion of those commitments are due in the 12 months after September 30, 2026, and about $447.5 billion are expected before December 31, 2030 . AP summarized the same point bluntly: Claude’s maker has committed to spend $895.4 billion on AI computing and infrastructure, including $70 billion due over the next 12 months .
This is where Anthropic’s story stops looking like a classic software IPO and starts looking like a hybrid of software company, cloud tenant, chip buyer and infrastructure financier. The company says it is shifting toward a hybrid model that includes direct chip leasing and dedicated facilities operated on its behalf, a strategy it calls Anthropic Direct . The rationale is straightforward: short-term compute is expensive, while long-term dedicated capacity could improve economics once it comes online .
But the risk is equally straightforward. Commitments made before capacity is operational can reshape cash flow long before the revenue arrives. Anthropic reported $146.6 billion of liquidity as of September 30, 2026, including cash, short-term investments and availability under a credit facility . That is a large cushion by almost any corporate standard, but it sits against a future compute obligation that is several times larger than the company’s current annualized revenue.
Claude is also changing the company that sells it
The amended filing is not just a financial document. It is also a capability disclosure. Anthropic says Claude has become central to the development of future Claude models, including experiments, training-data work, evaluations, safety audits, code writing and code review . In September 2026, more than 90% of contributions to Anthropic’s codebase were written by Claude, and the average daily volume of code contributed per engineer in the third quarter was more than 45 times the average daily volume in 2024 .
That is powerful evidence of internal leverage, but it is not the same as proving autonomous recursive self-improvement. Anthropic itself draws the line carefully. The filing says the company’s internal estimates and expectations for future models indicate that it may imminently meet its Responsible Scaling Policy definition of “dramatically accelerated progress attributable to AI,” but it also says no current model has been designated as exceeding that threshold . Wccftech highlighted the same caveat, noting that Anthropic’s disclosure does not necessarily mean fully autonomous recursive self-improvement has arrived .
The safety concern is that AI-assisted development can compound both progress and mistakes. Anthropic warns that as reliance on Claude deepens, systemic flaws, biases or errors in current models could propagate into later versions and become harder to detect or fix . The company also says it does not anticipate safely allowing Claude to fully autonomously design and build successive generations of frontier models without meaningful human supervision and understanding .
The IPO test
The IPO will test three investor beliefs at once. First, that enterprise and developer demand for Claude can remain strong enough to support a revenue base that already looks enormous. Second, that compute spending can become more efficient as long-term capacity replaces expensive short-term supply. Third, that Anthropic can keep accelerating model development without letting automation outrun oversight.
That combination is why the quarter is so striking. Anthropic did not simply post a $17 billion quarter. It posted a quarter that forces public markets to price frontier AI as a business model: astonishing revenue velocity, thin or volatile operating results, massive forward infrastructure obligations, and a product increasingly used to build the next version of itself. Wall Street just found a new boss battle, and it is measured in tokens, chips, safety thresholds and billions of dollars.
Comments
Be the first to comment.