Anthropic files S-1 with 94 pages of risk factors and existential warnings

Anthropic filed its S-1 prospectus on Monday, submitting a 335-page document that blends aggressive financial projections with stark warnings about the future of humanity. The filing includes 94 pages dedicated to risk factors, a volume that contrasts sharply with earlier tech IPOs. The company’s founders argue that this level of disclosure is essential to attracting long-term investors who understand the scope of the technology.

The document marks a shift in how Silicon Valley pitches its vision to Wall Street. Jay R. Ritter, a professor at the University of Florida’s Warrington College of Business, noted that Apple’s 1980 prospectus was only 47 pages long and contained no risk factors. Ritter identified Facebook’s 200-page filing as a turning point, which included 22 pages of risk factors. He described the current trend as a bloat that makes it difficult for readers to find critical information.

Anthropic’s founders explicitly reject a purely transactional view of their stock. In a letter included in the filing, they urge investors to see their purchase as an opportunity to be part of something greater. They describe the march toward advanced AI as one of the seminal moments in human history. The company aims to help humankind chart a safe course through this technological evolution and bring forth an age of miracles.

The filing does not shy away from the potential downsides. On page 46, the document states that concerns have been raised that the models pose catastrophic or existential risks to humanity if not developed with appropriate safeguards. It warns that the development of highly advanced models could further increase the risk that these systems cause harm. CEO Dario Amodei has previously stated there is an up to 25 percent chance that further AI development could lead to human extinction.

The company calls this approach radical honesty. The founders explain that oversharing is part of an anti-sell strategy designed to avoid buying based on unfulfillable expectations. They cite a guiding principle of underpromising and overdelivering. The filing also includes admissions that the company may make forward investments that seem irresponsibly large. It notes that model safety and alignment may lead to limiting or delaying the deprecation of older models.

Anthropic will go public as a public benefit corporation. This entity type requires the company to prioritize a social mission alongside shareholder interests, unlike standard C-corporations that must maximize shareholder value. The filing presents powerful AI as inevitable. It argues that Anthropic should lead the development of this technology to ensure it is safe and secure. The company states that the value generated by the AI exponential will far outweigh any shorter-term economic sacrifices made as part of its mission.

Financial details in the S-1 show a company with high growth and high costs. Anthropic saw explosive revenue growth in the second quarter, but operating losses narrowed only slightly. The document reveals hundreds of billions of dollars in financial obligations to secure computing power. The company also announced it will start building its own data center infrastructure, a costly move that will impact profit. Third quarter results are currently missing from the filing.

Ritter expects Anthropic to share its third quarter results as the IPO date approaches. He compared the company to Apple, which underperformed the market in its first three years but is now worth $5 trillion. He distinguished Anthropic from WeWork, which he described as a money-losing real estate company pretending to be a tech company. Anthropic includes a risk factor on page 63 stating it has a history of net losses and may not generate sufficient revenue to achieve or sustain profitability in the future.