Investors Raised $36 Billion in Debt for Google Chips for Anthropic

Anthropic finances the purchase of critically expensive servers through third-party investors, who have attracted $36 billion in borrowed funds. This demonstrates how AI companies competing with OpenAI are managing astronomical infrastructure costs.

The Circular Financial Scheme

According to Bloomberg, a group of investors is taking out a $36 billion loan to purchase Google TPU (Tensor Processing Units) — specialized chips used to train and run modern AI models. These chips are then transferred to Anthropic. This approach allows Anthropic itself to avoid being the borrower and not bear direct legal responsibility for the use of financing. This financial architecture resembles structures used by OpenAI. Money passes through multiple hands, companies receive the equipment they need, but formally distance themselves from debt obligations.

Why Investors Take This Risk

AI infrastructure requires capital investments that go beyond traditional financing:

• A single global data center costs $10–20 billion
• Google TPUs are supplied in limited quantities and only under long-term contracts
• Training a modern LLM requires millions of computing hours
• Competition for chips has intensified — capacity must be reserved years in advance

For investors, this is a risky bet: if Anthropic grows in valuation and captures a significant share of the AI market, their investment will pay off many times over. If not, they are left holding expensive equipment.

What This Means

Fintech and traditional financiers are becoming woven into the core of the AI industry. The race for computing power may soon hit borrowing limits — banks will begin imposing strict conditions or refuse outright. This could slow the growth of new AI startups and concentrate power in the hands of large companies with their own cash flows.