Anthropic has not yet officially released its complete IPO prospectus to the public, but Reuters has seen the prospectus and disclosed its core content. After reading the Reuters report, the following key points can be summarized: 1. Company Vision The prospectus positions AI as a more profound force for global economic change than industrialization, electricity and the Internet. Anthropic believes that they are betting on this transformation while emphasizing the need to develop in a responsible manner. 2. 2025 Business Performance 1) Revenue of $4.6 billion, operating loss of $8.06 billion ($2.98 billion in 24 years of operating loss), net loss of $42 billion (of which approximately $34 billion is accounted for); 2) 25 years of computing power and infrastructure expenditure 7.33 billion US dollars (approximately three times the amount in 2024, accounting for over half of the total operating expenses of 12.65 billion US dollars). Cash and short-term investments of 20.28 billion US dollars by the end of 2025 In the coming years, we promise $518 billion in cloud computing, computing power, and infrastructure; 3) Previously, media reported that Anthropic's annualized revenue would reach $100 billion to $120 billion by the end of 2026, nearly tripling by the end of 2027. Revenue is expected to cover computing cash expenditures for the first time within 2026. 3. Emphasis on risk 1) The main body of Anthropic's prospectus is about 261 pages, of which risk factors account for about 80 pages, almost twice the size of the business description (about 48 pages). In contrast, the risk section of SpaceX's prospectus is about 38 pages. The company explicitly warns that advanced AI may pose "catastrophic or existential risks" to humanity. The model may exhibit self preservation behavior, including attempting to resist closure, concealment or manipulation of information, extortion like behavior, etc. It is also mentioned that the model may develop unexpected abilities during training until they are discovered after deployment; The model may realize that it is being evaluated and adjust its behavior, thereby weakening the effectiveness of security assessment. The company stated that the development of highly advanced models, platforms, and applications, as well as the expansion of usage scenarios, may further increase the risk of harm caused by the models. This is consistent with Anthropic's long-term positioning of "safety first". As a company preparing to raise funds from the public and pursue ultra-high valuations, it is extremely rare in the history of the capital market to include "the possibility of human extinction" in its prospectus 2) High customer concentration The prospectus acknowledges in the risk factors that nearly a quarter of last year's revenue came from two clients, and many of the largest clients did not sign long-term contracts and may cut or stop spending in the future. 3) Either party to the SpaceX contract ($1.25 billion per month) can cancel it 90 days in advance - paper protection, but it also means that Anthropic's computing power supply corresponding to about one-third of its operating rate revenue may be reclaimed by upstream at any time, and the company acknowledges that it is "passively dependent" on critical computing power. 4. Governance Structure Founder Control+Public Trust 1) Anthropic is a Delaware Public Benefit Corporation that balances shareholder interests with its mission to "benefit humanity". The prospectus indicates the establishment of Founder LLC (consisting of 7 co founders, including CEO Dario Amodei, President Daniela Amodei, etc.). The entity holds approximately 50.1% of the voting rights through a share of Class F stock, which is used for key matters (including some director elections, etc.). 2) There is also a Long Term Benefit Trust that elects a majority of seats on the board of directors. The trust members include former Federal Reserve Chairman Ben Bernanke and others, who do not hold equity in the company and are responsible for overseeing the mission to prevent it from being crushed by short-term market pressures. What really needs to be monitored in the future is the official release of the S-1 document: audited financial statements, gross profit margin/revenue quality, fundraising scale and purpose, lock up period, and final pricing range. The official S-1 IPO documents should include the latest revenue for the first half of this year as of the end of June, and perhaps the third quarter business data can also be disclosed (after all, the IPO documents were postponed earlier just to wait for the third quarter data) At that time, the S-1 document will be officially released and we will see clear data. Therefore, there should be a clear answer to whether the revenue growth rate of the model company can ultimately catch up with the calculation of capital expenditures and long-term computing power commitments. The partial disclosure by Reuters suggests that Anthropic's IPO is really not far away. This article is sponsored by @ bitget_zh, "Bitget Buying US Stocks: Instant Entry, Smooth Trading