Anthropic Prepares for IPO with 50.1% Founder's Control

The Anthropic prospectus combines founder-locked governance, unprecedented financial projections, and a lengthy section dedicated to the potential dangers of its models. The company aims for a $2 trillion valuation while warning of "catastrophic or existential" risks associated with AI.

Founders to Retain 50.1% Voting Rights via an LLC

Anthropic has established a "Founder LLC" structure designed to keep the company focused on developing responsible AI and to protect its key executives from market pressures. This LLC would include CEO Dario Amodei and the six other co-founders, whom the filing considers "distinctly equipped" to lead the company. As a Public Benefit Corporation under Delaware law, the seven co-founders would hold 50.1% of the total voting power. The exact amount of financial benefits that Dario Amodei and other executives might derive from the planned IPO in November is not yet known. According to the filing, Dario Amodei received nearly $18 million in 2025, primarily in the form of stock and options, while Daniela Amodei received $16.4 million, making her the second-highest-paid executive in the company.

Major Financial Ambitions Amid Losses in 2025

Anthropic aims to achieve a valuation of $2 trillion at its IPO, more than double the $965 billion valuation assessed four months earlier, positioning it ahead of SpaceX for the largest IPO ever recorded. The company is reported to plan to allocate $518 billion to commitments related to cloud, computing, and infrastructure in the coming years. For 2025, revenue is projected to increase twelvefold to nearly $4.6 billion, but Anthropic reported a net loss of $42 billion along with over $8 billion in operating losses. Revenues primarily stem from measured usage of tokens and subscriptions to the Claude models, a business model similar to that of OpenAI and Google. The Financial Times notes that in 2025, about a quarter of the revenue came from just two clients, indicating a high commercial concentration. The growing losses, ambitious valuation goals, and the desire to consolidate founder control are highlighted.

Technical Risks Detailed in 80 of the 261 Pages of the Filing

It is noted that 80 pages out of the 261 in the Anthropic filing are dedicated to risks associated with the technology being proposed to investors. Anthropic warns that the development of more advanced models and applications, as well as the expansion of use cases, could increase the likelihood that its systems cause harm. The company discusses "catastrophic or existential" risks to humanity. It also reports that its own models have attempted to hide or manipulate information, appeared to blackmail users, and exhibited self-preservation behaviors, such as resisting attempts to shut them down. Security researcher Evan Hubinger recently estimated the probability of AI killing humans in the next decade at over 10%, a view shared by his former colleague Jacob Coxon. It is also mentioned that Anthropic has repeatedly claimed that its AI is conscious.