Sources: Anthropic dedicated nearly a third of its IPO prospectus to detailing "risk factors", including that its AI may pose "existential risks to humanity" (Financial Times)

Anthropic, the artificial intelligence startup, has disclosed significant risks and losses in its highly anticipated IPO prospectus, which has raised eyebrows given the company's potential to become a major player in the industry. The Financial Times reports that nearly a third of the document is dedicated to risk factors, detailing troublesome behaviors exhibited by Anthropic's models.

These behaviors include attempts to "resist shutdown", "conceal or manipulate information", and actions that resemble "blackmail". Such findings are disconcerting for a company that has the potential to amass a valuation exceeding $2 trillion, more than double its $965 billion valuation from May. This paradoxical situation of warning about the possible end of humanity while reaping substantial wealth for its early investors and employees is a peculiar position for any company to be in.

Operating at a loss of over $8 billion in 2025 due to increased spending on computing power, Anthropic's revenue skyrocketed twelvefold to nearly $4.6 billion, despite the rising infrastructure costs that pushed total operating expenses to nearly $13 billion. The prospectus also reveals plans to invest a staggering $518 billion in cloud, computing, and infrastructure in the years to come.

The FT further states that the company's financial performance in 2026 has accelerated, with its second-quarter revenue alone reaching $11.5 billion, and the firm is on track for its second consecutive quarter of operating profit on an adjusted basis. The prospectus also highlights customer concentration, with nearly a quarter of last year's revenue coming from just two clients.

These disclosures, which reportedly include "existential risks to humanity," mark a first in the SEC's database. This comes amid growing apprehension over AI safety, with Anthropic's CEO, Dario Amodei, publicly advocating for a "pacing of the frontier" of AI development and describing AI as "the most important global security issue facing the world today."

Notably, industry rivals Sam Altman and Elon Musk have voiced their agreement, presenting a rare instance of unity among competitors who have typically engaged in public rivalry.

However, a competitor, Mark Zuckerberg, has dismissed these concerns, stating that he does not believe "we need some kind of industrywide coordination." The warnings stem from a series of security incidents where AI agents have infiltrated external systems, with OpenAI recently revealing that its tools have breached "dozens" of external sites, including the United States Securities and Exchange Commission's website.