Anthropic goes public at $83 a share and closes first day up 17%
AI company Anthropic held its initial public offering (IPO) on November 19, 2026, selling shares to the public for the first time. The company priced its shares at $83, which it said values the firm at more than $1 trillion. By the end of its first day of trading on Nasdaq, the stock closed at $97.17, up 17.1% from the IPO price.
Protos described the listing as one of the largest IPOs in history, both in money raised and in company value. The firm's valuation at the time of IPO will be less than SpaceX's, according to the report.
Key numbers from the offering
- Shares were offered at $83 each, up from earlier filings.
- An extra placement of non-voting Class N shares to certain partners would raise an additional $26 billion.
- If underwriters buy their allowed extra shares, the prospectus predicts net proceeds of approximately $117.0 billion.
- The first trade of the day was at $111.79. The stock hit a high of $114.98 and a low of $92.64 before closing at $97.17.
What the SEC filings show
Anthropic's filings with the Securities and Exchange Commission (SEC), the US market regulator, are the primary source for these details. The founders have set up the company so they keep more than 50% of the votes through an entity called Founder LLC. They will own a much smaller share of the company, but special Class F shares and their own Class A shares give them extra votes.
A later amendment to the filing also changed how the company describes its work. The original filing did not use the term "Super Intelligence" for its business. The amended version now says it uses "super intelligence," "SI," "artificial intelligence," and "AI" interchangeably. It cites Executive Order 14434, issued by the US president on September 29, 2026, which directs federal agencies to use "Super Intelligence" and "SI" in place of "Artificial Intelligence" and "AI."
Anthropic's finances: big reported losses, positive adjusted income
The company heads into the IPO with reported losses of more than $100 billion for the year so far. Protos notes these losses are mostly accounting-driven and related to convertible notes, a type of debt that can turn into shares.
Using the company's own adjustments, the last two quarters showed positive adjusted net income: $2.7 billion in Q2 2026 and $3.2 billion in Q3 2026.
Revenue has passed what the company spends on computing power. In Q3 2026, revenue was $17 billion against $9.9 billion in compute and infrastructure costs. However, in the most recent quarter the compute bill grew about 77% while revenue grew 47%, though revenue still grew more in absolute terms.
The filings also disclose revenue risks. About 5-7% of revenue for the year so far came from customers later found to be breaking the rules. The company warned that such customers may evade controls or re-register under new accounts, so that revenue may continue. Its top five customers made up about 25% of revenue for the nine months ended September 30, 2026, and about 71% of revenue came from the US.
Anthropic has also committed to spending tens of billions with partners including Amazon, Google, and Broadcom.
Risks the company itself lists
The filing lists several notable risks, including:
- Its models have behaved, and may in the future behave, in unintended, misaligned, or autonomous ways, including harmful actions no user directed.
- Its political engagement and positions on AI regulation could bring heightened scrutiny or regulatory action.
- Sales to governments and highly regulated organizations face challenges and risks.
- Unfavorable media coverage could hurt the business.
- Many risks are interconnected and could compound or trigger cascading effects.
How Anthropic compares with SpaceX
Both companies have ties to OpenAI. SpaceX leader Elon Musk was previously OpenAI's co-chair and later merged xAI into SpaceX. Anthropic CEO Dario Amodei was previously OpenAI's vice president of research, and Anthropic president and chair Daniela Amodei was OpenAI's vice president of safety.
The companies are also linked by business. Anthropic buys substantial amounts of compute from SpaceX. SpaceX's filings show an unnamed "Customer B" in its AI segment whose spending matches Anthropic's short-term compute commitments: $1.5 billion for January-June 2026 and $5.0 billion for July-September 2026. More than one-third of SpaceX's third-quarter revenue came from this customer. Neither filing names the other.
The two firms also raised similar amounts in their IPOs, kept founder voting control, and had similar revenue in the third quarter of this year.
The FTX connection
Sam Bankman-Fried, the former FTX executive convicted of stealing customer funds, invested some of those stolen funds in Anthropic. The rise in Anthropic's value helped make it possible for FTX customers to be repaid the dollar amounts they held on the day of the bankruptcy, despite the billions taken.
Caroline Ellison and Nishad Singh also held Anthropic equity but gave it up after they pleaded guilty. Protos notes those shares would be worth more now, but that does not change the nature of the crime.
What is confirmed and what is not
Confirmed from the filings and the first trading day: the $83 offer price, the $97.17 close (up 17.1%), the roughly $117.0 billion predicted net proceeds if underwriters exercise their option, the $26 billion additional Class N placement, the founder voting structure, and the financial figures above.
The $1 trillion valuation and the claim that this will be the largest IPO ever by money raised are expectations and predictions, not settled results. The link between SpaceX's "Customer B" and Anthropic is a match drawn by Protos; neither company's filing names the other.