Anthropic revises S-1/A: Q3 revenue of $17 billion, net loss of $104 billion in the first nine months
Anthropic filed a revised IPO prospectus (Form S-1/A) with the SEC on November 2, 2026. The filing shows that the company's revenue in the third quarter of 2026 was about $17 billion, up about 1,127% year over year; revenue for the first nine months (through September 30) totaled $33.24 billion, up about 1,200% year over year.
Net loss for the first nine months in the same period was $104 billion, with the vast majority coming from changes in the fair value of convertible notes, derivative liabilities and other items — these were non-cash accounting charges brought on by an increase in valuation, rather than actual cash burn. Net loss for the third quarter alone was about $23.3 billion. On an operating basis, the company's operating loss for the first nine months has narrowed significantly.
The prospectus also disclosed that the company's purchase commitments and lease agreements already signed totaled about $895.4 billion, the vast majority of which are related to computing power and infrastructure. As of September 30, cash and short-term investments were substantial, indicating strong financing capacity. The company plans to list on Nasdaq under the ticker symbol “ANT”; the prospectus has not yet filled in the offering price or number of shares.
Original sources:
Title: Anthropic reports $17 billion in Q3 revenue, purchase commitments of $895.4 billion in revised S-1 Publisher: CNBC Publication date: 2026-11-02 URL: https://www.cnbc.com/2026/11/02/anthropic-ant-s-1-quarter-financials.html Source nature: financial media report (directly citing SEC S-1/A) Verification status: full text read in this instance (verified on 2026-11-02) Supported content: supports all core financial figures and commitment scale in the body, including about $17 billion in Q3 revenue, $33.24 billion for the first nine months, $104 billion in net loss and $895.4 billion in purchase commitments
ABAB AI Insight
Anthropic's pace of commercialization is extremely strong among frontier AI companies. Revenue rose from near zero at the start of commercialization in 2023 to about $4.585 billion for full-year 2025, then to $33.2 billion in the first nine months of 2026; in the third quarter alone it reached about $17 billion, extending the strong momentum of about $16.3 billion in the first half. Net revenue retention of about 900% shows rapid expansion in spending by existing customers. The company emphasizes that its “improvement loop” (customer feedback accelerating model iteration) and “investment loop” (revenue feeding back into computing power and research) form a positive feedback cycle.
The main driver of the net loss is non-cash accounting items. Of the $104 billion net loss in the first nine months, the vast majority (about $94.3 billion) came from changes in the fair value of convertible notes and derivative liabilities, which are directly tied to the company's rising valuation and did not represent operating cash outflows. In the first half (through June 30), it had already recorded about $80.7 billion in net loss, and the revised filing updates that same logic. Operating losses have narrowed significantly, with the main costs including computing power and infrastructure (about $18.8 billion for the first nine months) and personnel expenses.
Capital intensity is extremely high. The scale of about $895.4 billion in computing power and infrastructure commitments is enormous, with about $70 billion of that due within the next 12 months, highlighting long-term lock-in risk. The company's cash and short-term investments totaled about $131.9 billion as of September 30, providing a buffer for near-term performance, but if model iteration or customer demand falls short of expectations, the long-term commitments may still create pressure. Amazon is one of the main holders of convertible notes, and the related fair value changes are closely tied to the arrangement capping Amazon's shareholding.
This revised S-1/A is an important milestone in Anthropic's public financial disclosures, and it is also the first time a frontier AI lab has disclosed audited figures to the public on such a scale. It not only shows Claude's pace of commercialization, but also puts computing-power lock-in, valuation volatility and non-cash losses in the spotlight at the same time. Follow-up points to watch include the offering size and valuation range in the formal roadshow, as well as the potential impact of OpenAI's possible IPO timing in 2027 on the market.
ABAB News · Cognitive Law
The main source of the net loss is non-cash expenses arising from changes in the fair value of convertible notes, rather than operating cash burn.
The $895.4 billion in computing power commitments alongside $131.9 billion in cash reflects capital lock-in and financing capacity under high growth.
Q3 revenue for a single quarter has already exceeded half of the first nine months, showing that commercialization is still accelerating rather than slowing.