Anthropic’s IPO prospectus lays bare the contradiction at the heart of the AI boom
The company promises to help solve humanity’s most difficult problems while warning that its technology could threaten humanity itself. Its financial disclosures offer a glimpse of the stakes as AI investment reaches unprecedented levels and questions about the industry’s sustainability intensify.
The word “humanity” appears 21 times in the prospectus Anthropic filed on Monday ahead of its anticipated Nasdaq IPO. Most references concern the company’s ambitious goals. In its opening lines, Anthropic speaks of the “responsible development of advanced AI for the long-term benefit of humanity.” Elsewhere, it promises to help with “solving humanity’s most difficult problems,” a phrase that recurs throughout the document.
But some of the most striking references appear in a less reassuring context. In the “Risk Factors” section, on page 46, Anthropic warns that its technology could “pose a catastrophic or existential risk to humanity.”
The prospectus is more than a snapshot of the company’s finances. It captures the defining contradiction of the AI era: promises of extraordinary progress alongside warnings of potentially unprecedented dangers, with enormous financial rewards at stake for those positioned to benefit.
Anthropic is reportedly preparing for an IPO that could value it at $2 trillion and raise as much as $100 billion, potentially challenging Elon Musk’s SpaceX for the record for the largest public offering in history. The prospectus arrives at a moment when expectations for the AI revolution are running high, even as concerns about an AI bubble grow, particularly around the vast sums being poured into computing infrastructure.
How investors respond to Anthropic’s disclosures could offer a glimpse of which narrative currently carries more weight. The figures provide an unusually detailed look at the financial health of one of the world’s most closely watched technology companies, as well as the question hanging over the entire industry: can the business of building ever more powerful AI systems become sustainable?
According to the prospectus, Anthropic generated $4.59 billion in revenue in 2025, an increase of 1,087.8% from 2024. Revenue had already risen 841.5% in 2024 compared with the previous year. The growth reflects surging demand for AI products, particularly among enterprise customers, which form the core of Anthropic’s business.
The company’s customer base has expanded just as quickly. The number of enterprise clients rose from approximately 1,500 at the end of 2025 to more than 11,900 by the end of September, an increase of 693.3%.
That expansion, however, has come at a substantial cost. Anthropic’s largest expense category in 2025 was computing and infrastructure, which totaled $7.33 billion, up 190.5%. Personnel expenses climbed 305.1% to $1.84 billion, reflecting the cost of competing for highly sought-after AI talent. Marketing expenses reached $568 million, an increase of more than 1,220%.
The losses remain substantial, although their growth rates tell a more complicated story. Anthropic’s operating loss widened 170.5% in 2025 to $8.06 billion, while its net loss increased 405.1% to $41.97 billion. Both grew more slowly than in 2024, when the operating loss rose 440.8% and the net loss increased 442.7%.
The first half of 2026 brought further rapid growth in revenue. Sales surged 1,300.3% from the same period a year earlier to $16.29 billion. At the same time, the operating loss narrowed 67.5%, from $2.14 billion to $693 million. The net loss, however, rose 522.6% to $80.75 billion.
Despite the revenue growth and improvement in operating performance, the figures suggest that reaching break-even, let alone sustained profitability, remains a formidable challenge. Still, Anthropic says it is generating more revenue relative to the computing power it consumes. In the second quarter of 2025, the company generated $5.60 in revenue for every dollar spent on computing. A year later, that figure had more than doubled to $11.60.
Yet Anthropic’s story, like that of the broader AI industry, cannot be understood through financial statements alone. The history of technology offers plenty of examples of seemingly prohibitive engineering and cost challenges being overcome through innovation. Computing power may become cheaper, models may become more efficient, and new techniques may change the economics of AI. Whether those advances arrive quickly enough to justify today’s investments is another matter.
The larger question is what Anthropic’s technology will mean for the world.
The prospectus begins with a statement that captures the scale of the company’s ambition. “We believe that AI is one of the most consequential technologies in human history,” its seven founders write in a letter to investors. Few would dispute the significance of the technology, even if they disagree sharply over whether it will usher in a historic improvement in human well-being or expose humanity to dangers it cannot control.
Anthropic’s founders emphasize the potential benefits. “AI can profoundly accelerate the rate of progress in biology and medicine,” they write, before outlining a broad range of possible applications.
They also acknowledge that the technology could be used for devastating purposes. The prospectus warns of AI-assisted cyberattacks and the development of biological weapons, as well as the unpredictable behavior of autonomous systems.
Other risks include authoritarian governments using AI to build increasingly sophisticated surveillance systems, suppress dissent and consolidate power. Anthropic also warns that AI could cause significant economic disruption, potentially reshaping labor markets and industries faster than societies can adapt.
The prospectus devotes nearly 100 pages to the risks facing the company. Among them is the possibility that AI could threaten humanity itself.
“AI could, if not developed, deployed, and operated with appropriate safeguards, oversight, and security measures, pose catastrophic or existential risks to humanity, particularly if such systems were misused, replicated, weaponised, or allowed to operate with a high degree of autonomy,” Anthropic warns.
The company says the danger could intensify as AI systems become capable of automating more of the research and development process that produces increasingly advanced models. That could create a feedback loop in which progress accelerates beyond researchers’ ability to evaluate new systems and develop adequate safeguards.
The conclusion Anthropic draws from this warning is revealing. The company says such risks “could adversely affect our business .”
In other words, the prospectus frames a potential existential threat to humanity, at least in this passage, partly in terms of the damage it could do to Anthropic’s business. The warning is real, but so is the gulf between the scale of the danger described and the corporate consequences emphasized in the disclosure.
The juxtaposition is difficult to ignore. A company promising to help solve humanity’s greatest problems is also warning investors that its technology could pose a catastrophic risk to humanity, while treating the resulting damage to its own business as one of the possible consequences.
Humanity has rarely faced a technological crossroads quite like this one. AI could help accelerate scientific discovery, improve living standards and solve problems that have resisted generations of human effort. It could also amplify cyber and biological threats, enable new forms of authoritarian control and create risks that are difficult to predict, let alone contain.
Such a powerful technology demands extraordinary care in its development and deployment. That means weighing potential benefits against risks, building safeguards that keep pace with technical progress and being prepared to slow down when the dangers warrant it.
Instead, the prospectus offers a glimpse of an industry caught between its own warnings and the commercial incentives driving it forward. Anthropic and its competitors acknowledge the possibility of severe harm, yet continue to race to build and deploy more capable systems. The prospect of enormous returns makes restraint difficult: a company that slows down may fear that a rival will move ahead and capture the market.
That tension is at the heart of the AI boom. The companies developing the technology may genuinely believe it can transform the world for the better. But they also have powerful financial reasons to keep pushing forward, even as they warn that the consequences could extend far beyond their balance sheets.
Anthropic’s prospectus makes both sides of that equation unusually clear. It asks investors to believe in AI’s extraordinary promise while disclosing risks that, by the company’s own account, could be existential. The question is whether financial markets, regulators and the companies themselves can take those warnings as seriously as they take the opportunity.
