Anthropic S-1 Filing Reveals $1 Trillion Commitment Over 20 Years

According to data from Anthropic's S-1 filing, the company is entering into massive financial commitments ranging from $850 billion to $1 trillion for compute and infrastructure investments over the next 20 years. To cover these costs, the company needs to exhibit rapid growth and rely on recursive self-improvement.

Anthropic's S-1 filing shows that the company provides financial justifications to maximize its compute spending. The company has tied its future to the expectation of dramatic growth; its revenues multiplied 14-fold from the first half of 2025 to the first half of 2026, reaching $16.3 billion. It is noted that the $80 billion loss is largely an accounting matter, and excluding these items and employee compensation, an adjusted income of $1.62 billion was achieved in the first half of 2026.

The company defines only OpenAI and Google as competitors while focusing on enterprise areas such as coding and cybersecurity. In the first half of 2026, approximately three-quarters of customers spent at least $100,000 annually. However, Anthropic faces intense competitive pressure from other pioneer labs and open-source models. Unlike Google, the company does not have multiple revenue streams and carries long-term liabilities totaling between $850 billion and $1 trillion through fixed and non-cancellable purchase commitments, leases that have not yet commenced, and new compute deals. Risk factors also include the warning that using AI to develop more AI could advance model capabilities faster than safety measures.