Anthropic’s initial public offering (IPO) filing includes a notably candid risk disclosure that highlights the potential for its artificial intelligence (AI) products to pose an existential threat to humanity. While this stark admission is unusual in IPO documents, experts warn that it may obscure more critical financial and operational details investors need to assess the company’s prospects.
Mandatory since the 1960s in the United States, risk factors sections have historically aimed to provide transparency for potential investors by outlining challenges and uncertainties. However, these disclosures have grown increasingly lengthy and, at times, somewhat perplexing. For example, Microsoft’s 1986 filing contained fewer than two pages of risks, while more recent companies such as SpaceX and Uber have submitted listings with 38 and 50 pages of risk factors, respectively.
Anthropic’s warning about the possibility of human extinction is distinctive but not entirely unprecedented. Some firms occasionally include unconventional risks—such as laboratory testing companies acknowledging pathogen transmission or media enterprises referencing risks tied to leadership changes due to incarceration. Energy firms also routinely address catastrophic threats tied to their operations.
The justification for highlighting an apocalyptic risk in Anthropic’s filing may extend beyond disclosure itself. If regulatory bodies impose stringent restrictions on AI development due to fears of existential harm, such warnings might support legal defenses against shareholder claims. However, for most investors, the pressing concern remains whether Anthropic will generate sustainable profits.
On that front, the IPO filing—and subsequent annual reports—offer limited guidance. Typically, companies avoid sharing forward-looking revenue and profit forecasts in pre-IPO documentation due to legal liability considerations. Instead, they rely on historical figures, which may be less informative for fast-growing companies like Anthropic. Last year, Anthropic reported $4.6 billion in revenue, yet by July of this year its run-rate revenue was estimated at $65 billion, reflecting rapid expansion but little visibility into future profitability.
Academic research indicates that roughly one-third of companies compensate for limited written forecasts by providing verbal guidance during pre-IPO roadshows, though these updates tend to be vague and focus on expected profit margins rather than precise financial outcomes. Investors seeking more detailed intelligence, such as the identity of major customers representing significant revenue shares or Anthropic’s own quantitative assessment of the likelihood of AI-related catastrophic scenarios, are unlikely to receive it.
Instead, Anthropic is expected to present broad estimates of its total addressable market—figures that often verge on the speculative. This mirrors tactics used by firms like SpaceX, which identified a market opportunity of $28.5 trillion in its IPO filings, far exceeding analysts’ near-term revenue expectations. The implication for investors is clear: the most prominent risk may be overestimating the company’s value based on optimistic but vague projections, leaving shareholders vulnerable to significant disappointment.