Anthropic IPO: What Investors Need to Know After Draft Prospectus Leak

TradingKey - On September 29, Eastern Time, Reuters reported that a draft of Anthropic's IPO prospectus obtained by the agency disclosed the company's revenue, losses, long-term compute spending commitments, and proposed listing governance arrangements.

Following the leak of the draft prospectus, Anthropic's historical financial data, long-term compute contracts, and shareholder rights have drawn significant market attention. Below is the key information currently known about Anthropic from the leaked draft IPO prospectus.

When Will Anthropic Go Public? IPO Valuation Could Exceed $2 Trillion

Anthropic announced on June 1, 2026, that it has confidentially submitted a draft S-1 registration statement to the U.S. Securities and Exchange Commission (SEC). According to a previous report by Reuters citing sources, the company may go public after the U.S. midterm elections in November, with its valuation potentially exceeding $2 trillion.

As of September 30, the specific listing date, number of shares to be issued, and offering price have not yet been announced.

In May this year, Anthropic completed a $65 billion Series H financing round with a post-money valuation of $965 billion. If the IPO valuation exceeds $2 trillion, it will more than double the valuation from that round.

In accordance with SEC requirements, the company must publicly file its registration statement and previously confidentially submitted drafts and amendments at least 15 days before the roadshow. At that time, investors will be able to review the financial statements, shareholding structure, and offering arrangements.

In addition, the list of brokerages participating in the retail allocation has not yet been announced, and subscription eligibility, minimum subscription amount, and allocation rules remain to be disclosed.

Anthropic Revenue and Loss: 2025 Compute Spend Reaches $7.33 Billion

According to Reuters, key financial data for 2025 disclosed in the draft are as follows:

Compute and infrastructure expenditure accounted for approximately 58% of full-year operating expenses, exceeding revenue over the same period. Of the approximately $42 billion net loss, roughly $34 billion stemmed from non-cash accounting charges arising from prior financing arrangements, primarily reflecting the increased valuation of financing instruments convertible into company shares.

Revenue growth also brought an improvement in profitability. According to a Financial Times report, Anthropic's revenue reached $11.5 billion in the second quarter of 2026, and it recorded an adjusted operating profit excluding stock-based compensation. The company expects its adjusted operating profit to remain positive in the third quarter.

Anthropic IPO Major Risks: Customer Concentration, Compute Contracts and Shareholder Rights

Customer concentration and cloud platform reliance. According to a Reuters report, in 2025, two customers contributed approximately 12% of revenue each, totaling about 24%. The report did not disclose customer names. The draft warned that many large customers have not signed long-term contracts and may reduce or cease purchases.

In the same year, Anthropic generated approximately $2.16 billion in sales through the cloud platforms of Amazon (AMZN) and Google (GOOGL), accounting for 47% of full-year revenue. In return, the company paid about $351 million in distribution fees to these platforms. The platforms are also responsible for collecting payments from certain customers, and payment disputes or settlement delays could affect the company's cash recovery.

Long-term computing power contracts and payment obligations. The draft shows that Anthropic's infrastructure arrangements with six partners involve at least $518 billion in spending over the next decade. Approximately 80% of this amount corresponds to non-cancellable contracts or fees that must be paid regardless of actual usage.

These arrangements could limit the company's ability to cut spending. If revenue growth falls short of expectations, the company must still fulfill the relevant payment obligations. Some partners are simultaneously investors, suppliers, and sales channels, and changes in their pricing, computing power supply, and contract terms could also affect operations.

AI safety and product development.Anthropic warned in the draft that more advanced AI models could pose catastrophic or existential risks to humanity, and might exhibit behaviors such as resisting shutdown, concealing actions, or manipulating users. The company stated that models might recognize they are being evaluated, and some capabilities and risks might not be discovered until after deployment.

Anthropic also disclosed that it had limited or delayed certain features due to safety considerations. The company chose to allocate computing power to research and safety efforts, rather than developing commercially attractive image and video generation products.

Founder control and shareholder interests. Under the proposed arrangements in the draft, Founder LLC, composed of seven co-founders, will hold a special Class F share, entitling it to 50.1% of the voting power on certain director elections and other matters submitted for shareholder vote. The Class A shares proposed to be issued to the public will carry one vote per share.

Class F and Class A shareholders will jointly elect three directors, while the remaining four directors will be appointed by the Long-term Benefit Trust. Anthropic intends to continue operating as a Delaware public benefit corporation, balancing the public interest in its decision-making.

The draft warned that these governance arrangements could cause certain decisions to conflict with the financial interests of shareholders and adversely affect the value of Class A shares.