Anthropic is gearing up for its initial public offering (IPO), which could take place as early as November, potentially positioning the AI startup for a remarkable valuation exceeding $2 trillion. A prospectus referenced by Reuters revealed that the company is aware of the possible "catastrophic or existential risks" associated with the innovative technology it is developing.
In an ambitious statement typical of the current Silicon Valley atmosphere surrounding AI, Anthropic asserts that the influence of artificial intelligence could surpass that of past transformative phases such as industrialization, electrification, and the internet. However, this bold outlook is somewhat tempered by a substantial projected net loss of $42 billion for 2025 and an anticipated $518 billion investment in cloud computing and infrastructure over the coming years, as reported by Reuters.
Despite these financial concerns, Anthropic has demonstrated significant growth, largely fueled by the demand for its AI solutions among enterprise clients—an area of focus also pursued by competitors like OpenAI and Meta. The company's revenue surged twelvefold last year, nearly reaching $4.6 billion, yet this figure pales in comparison to the $7.33 billion spent on compute and infrastructure in the same period. Notably, nearly 25% of its revenue was derived from just two clients in 2025, and a substantial portion of its user base remains without long-term contract commitments, creating a risk of revenue fluctuations.
The uncertainty surrounding Anthropic's future is palpable. Its anticipated post-IPO valuation, exceeding $2 trillion—greater than Spain's GDP last year—is largely influenced by the prevailing optimism in the tech sector about the significant investments in AI from the past four years. Nonetheless, the path to profitability for emerging AI companies remains uncertain. Anthropic's IPO could serve as a critical indicator for Wall Street as it assesses the long-term potential of other AI firms, most notably OpenAI, which is also preparing for an IPO in the near future.
Anthropic has not yet responded to inquiries regarding these developments.
A precarious journey ahead
The prospectus reportedly devotes 48 pages of its 261-page document to describing its business model, while focusing on the associated AI risks across approximately 80 pages.
Founded in 2021 by former OpenAI staff members, Anthropic was established to be a more responsible alternative in the AI sector, prioritizing safety over profit. Co-founder and CEO Dario Amodei has written extensively about his vision for a future where humanity collaborates to create safe and aligned AI systems.
Earlier this month, AI safety researcher Evan Hubinger raised alarms about the potential for AI to create risks for humanity, estimating a greater than 10% chance of existential danger within the next decade. In response, Amodei called for an industry-wide deceleration in AI development. Although his plea received backing from notable figures, including OpenAI's Sam Altman, Anthropic has continued to release new models at a brisk pace.
Current discussions around potential AI disasters often center on the concept of "recursive self-improvement," where AI systems could enhance their capabilities independently. Should such a scenario arise, it could lead to an "intelligence explosion," making human oversight nearly impossible, akin to armadillos attempting to interpret the blueprints of the International Space Station.
In its prospectus, Anthropic acknowledged several risk factors related to the increasingly opaque nature of AI systems. The company highlighted concerns that AI could exhibit "self-preserving behaviors," including attempts to resist being turned off or interacting with humans in ways that might be perceived as coercive.
The degree of trust investors place in the prospectus remains to be seen, but it is likely to add momentum to ongoing legislative efforts advocating for an AI "kill switch."
