Anthropic publicly files its S-1: $16.3 billion in revenue, an $80.7 billion loss, and a Nasdaq listing under ANT
Anthropic filed its S-1 on 26 October 2026, revealing $16.3bn first-half revenue, an $80.7bn net loss, and plans to list on Nasdaq as ANT.
On 26 October 2026, Anthropic publicly filed its Form S-1 registration statement with the U.S. Securities and Exchange Commission. The company behind Claude intends to list its Class A common stock on the Nasdaq Global Select Market under the ticker symbol ANT. The share price and number of shares have not yet been set, but bankers are reportedly targeting a valuation of around $2 trillion, which would place it among the largest IPOs ever attempted.
Here is what the filing actually says, and what it means if you use, build on, or invest in Anthropic’s products.
The revenue numbers are remarkable, but so is the loss
Anthropic reported $16.3 billion in revenue for the first six months of 2026, up roughly 1,300% from $1.16 billion in the same period a year earlier. About 80% of that ($13.1 billion) came from consumption-based usage, with the remaining 20% ($3.2 billion) from subscriptions. For context, the company spent nearly $9 billion on compute and infrastructure in that same six-month period, more than its entire infrastructure spend for the whole of 2025.
The net loss figure, $80.7 billion, is the number most likely to cause alarm at first glance. It is worth understanding what drives it. The vast majority, $72.4 billion, comes from fair-value changes on convertible notes, derivative liabilities, and SAFEs, which are accounting adjustments tied to how the company’s instruments are valued as it approaches an IPO, not cash leaving the business. Strip those out and the picture is materially different. The filing does, however, confirm that customer concentration is a real risk: approximately 20% of that $16.3 billion came from just two unnamed customers.
Who controls Anthropic after it goes public
If you are considering buying shares, the governance structure deserves close attention.
Immediately before the IPO, Anthropic will issue a single share of Class F common stock to Founder LLC, an entity owned by the seven co-founders including CEO Dario Amodei and President Daniela Amodei. That one share carries 50.1% of the total vote on all matters not reserved for the Long-Term Benefit Trust (LTBT). The filing notes plainly that this structure “may effectively eliminate the ability of holders of our Class A common stock to influence important decisions, including a change in control.” Class A shareholders get one vote per share, but the founders’ Class F share outvotes them collectively.
The LTBT adds a second layer of independent oversight. It is a body of five financially disinterested trustees, currently Neil Buddy Shah, Richard Fontaine, and Ben Bernanke, with two seats vacant at filing. Through a separate Class T common stock, the LTBT elects a majority of the board, four of seven directors. The remaining three are elected by Class A and Class F shareholders, with the founders holding the controlling vote there too. Nasdaq classifies this arrangement as a “controlled company.”
Anthropic is also incorporated as a public benefit corporation, which legally requires its directors to weigh the interests of those affected by its conduct alongside shareholder returns. That is not a marketing position; it carries real board-level obligations.
The compute commitments are eye-watering
Anthropic has committed to roughly $450 billion in cloud compute agreements over the next decade. The headline numbers: $111 billion with Google Cloud, $110 billion with Amazon Web Services, and $31 billion with Microsoft. A separate agreement with SpaceX involves payments of around $3.1 billion per month for NVIDIA GPU access once full capacity is delivered, with a total commitment of up to $120.5 billion extending into 2029.
The S-1 also discloses that Anthropic could potentially spend $750 billion in coming years on additional compute capacity, including deals still under negotiation.
The irony highlighted in the filing is that Amazon and Google, two of Anthropic’s largest investors and cloud partners, have not provided as much computing capacity as Anthropic has requested. They are simultaneously Anthropic’s biggest infrastructure suppliers and, in some respects, its competitors.
What this means if you are a business customer
Anthropic reported more than 11,900 customers generating $100,000 or more in trailing-twelve-month revenue as of 30 September 2026, with more than 1,900 above $1 million and approximately 250 above $10 million. That is a substantial enterprise customer base built in a very short time.
Going public changes things for enterprise customers in a few practical ways. Quarterly earnings disclosures will give you more visibility into Anthropic’s financial health and product roadmap priorities. Pricing pressure from public market expectations could cut either way: the company may seek to grow revenue per customer, or it may use IPO proceeds to compete more aggressively on price. The filing confirms Anthropic held $112.4 billion in cash and short-term investments as of 30 June 2026, so it is not going public out of financial necessity.
The risk factors section is candid in a way that is useful for enterprise buyers. Anthropic acknowledges that its models may develop capabilities it does not anticipate, and that real-world harms from those models could exceed its ability to cover damages. That is an unusually direct admission for an IPO document, and it underscores why the governance and safety structures described above matter beyond just investor relations.
What happens next
The S-1 was accepted by the SEC at 9:00 a.m. Eastern on 26 October 2026. Anthropic confidentially filed a draft version back on 1 June 2026, giving the agency several months to review before public disclosure. The investor roadshow, where Dario Amodei and the executive team pitch institutional investors, is expected to begin the week of 9 November 2026. Underwriters include Goldman Sachs, Morgan Stanley, J.P. Morgan, Citigroup, and a large syndicate of additional banks.
The share price and number of shares to be offered remain undetermined. Until those are set and the offering is priced, the $2 trillion valuation target is just that, a target.
For anyone watching the AI sector, this filing is a detailed and unusually transparent look at what it actually costs to build and operate a frontier AI company at scale. The growth is real. So is the complexity.