The Anthropic IPO is one of the most anticipated stock market listings in the artificial intelligence industry, but the company itself is not shy about warning about the risks of the technology it is developing. According to The Verge, the company's preliminary prospectus filing reveals mounting losses, leadership proposals to maintain control, and an open admission that its AI development plans could "further increase the risk of our models causing harm."
Anthropic , the company behind the popular AI assistant Claude , is on track for a valuation approaching $2 trillion — a figure that puts it in the elite of the most valuable tech companies in the world. Despite that valuation, the company is posting significant operating losses, reflecting the enormous costs of developing and training large language models. The contradiction between the ambition of an IPO and the frank warnings about the risks makes the Anthropic IPO one of the most interesting and controversial events in the tech industry.
The company was founded in 2021 by Dario Amodei and Daniela Amodei , along with other former OpenAI executives , with a central narrative of “responsible AI development.” Since then, it has received billions in investments from companies such as Google and Amazon , while Claude models have gained a significant presence in both the enterprise and consumer markets. The decision to go public marks a new phase for the company, which will now also have to be accountable to public investors.
See also: Anthropic and OpenEvidence offer free medical AI in nearly 100 countries
Anthropic IPO: What the prospectus reveals about AI risks
Anthropic’s IPO prospectus is unusually candid about the risks facing the company — and humanity at large. Anthropic explicitly admits that developing more powerful AI models could increase the risk of harm , even if that’s not the intention. That’s a rare admission for a company that’s also asking investors to invest in it based on the promise of developing just that technology.
Among the risks mentioned in the bulletin are “ catastrophic ” scenarios where AI models could be used to develop biological or chemical weapons, carry out large-scale cyberattacks, or engage in other activities that could seriously harm society. The company acknowledges that even with the best security measures, it cannot completely guarantee that its models will not be used for malicious purposes.
Additionally, the bulletin addresses the risk of “ misalignment ” — the potential for AI models to develop goals that are not aligned with human values and intentions. This is one of the central issues in AI safety research , and Anthropic recognizes it as a real risk that has not yet been fully resolved.
Anthropic IPO: Power Structure and Financial Losses
Beyond the AI, the Anthropic IPO also reveals important details about the company’s power structure. Leadership is reportedly proposing structures that would allow it to maintain control even after going public — a practice that has become common among tech companies but raises questions about accountability to investors. Dario Amodei as CEO and Daniela Amodei as President are expected to retain disproportionate influence through dual-class share structures.
See also: Anthropic & OpenAI: Need for independent AI safety assessors
Financially, the picture is mixed. Anthropic has seen rapid revenue growth, with annual revenue reportedly approaching $4 billion — an impressive number for a company that was founded just a few years ago. But its operating expenses are equally impressive: training large language models requires massive amounts of computing power, and the company continues to post net losses. This means that investors in the Anthropic IPO are betting on future profitability, not present-day profitability.
Comparisons with other AI companies are inevitable. OpenAI , Anthropic 's main competitor , is also in the process of transforming into a for-profit entity, while Google DeepMind and Meta AI are also investing billions in model development. The competition is fierce, and Anthropic needs to prove that its focus on security is not a barrier to its competitiveness.
It’s worth noting that Anthropic has developed a unique approach to AI called Constitutional AI — a method of training models based on a set of principles and values. This approach has received positive reviews from security researchers, but it’s not a panacea. The newsletter itself admits that even with these techniques, risks remain.
For European and Greek investors, the Anthropic IPO is of particular interest in the context of the European AI Act, which sets strict requirements for high-risk systems. Anthropic will have to comply with these requirements to maintain access to the European market, which adds additional cost and complexity to its business prospects. European regulatory pressure is one of the risk factors mentioned in the prospectus.
See also: Artificial intelligence safety: Anthropic proposes "brake"
Overall, the Anthropic IPO represents a critical moment for the entire AI. The company is being asked to balance investor demands for growth and profitability with its commitment to responsible AI. The prospectus’s candor about the risks is remarkable, but it doesn’t answer the fundamental question: can a company that acknowledges that its technology could be “disruptive” continue to develop it at a rapid pace? That contradiction will remain at the center of the debate as Anthropic moves toward its public listing.
