Alphabet Pledges Up to $40 Billion in Anthropic AI Bet

Alphabet’s planned investment of up to $40 billion in Anthropic marked one of the largest private bets yet on frontier artificial intelligence, and it showed how quickly the industry’s capital demands had outgrown simple startup financing. Google committed $10 billion in cash immediately, with another $30 billion tied to performance targets, valuing Anthropic at about $350 billion and linking the deal to the compute needed to train and run large models at scale.

The money was not just buying influence. It was buying access to infrastructure, the scarce layer now deciding which AI companies can keep growing. Anthropic said the funds would support a major expansion of its computing capacity, a bottleneck that has become central to the race among the biggest AI players. In a market increasingly shaped by chips, power and data-center space, the ability to secure enough hardware has become as important as model quality.

The investment also landed just after Anthropic disclosed that Amazon would invest up to $25 billion in the company, underscoring how intensely the leading cloud and tech companies were competing for a stake in the same fast-growing rival. Anthropic’s annual run-rate revenue topped $30 billion this month, up sharply from about $9 billion at the end of 2025, as demand for its coding-focused tools drew more enterprise customers into the fold. That surge made Anthropic one of the most valuable prizes in the AI market and helped explain why two of the industry’s largest infrastructure providers were willing to keep pouring in capital.

The broader shift is from flashy model launches to an arms race in capacity accumulation. Anthropic has already lined up other infrastructure deals, including work with Broadcom and CoreWeave, and it is expected to secure nearly a gigawatt of capacity through Amazon’s chips by year-end. Investors have also been talking up valuations far above the latest round, a sign that capital is chasing companies that can actually absorb more computing power, not just promise it.

For Google, the strategy cuts both ways. Anthropic’s Claude models compete directly with Google’s Gemini efforts, even as Anthropic relies on Google cloud services. That creates a delicate balance between collaboration and rivalry, and it deepens concerns that frontier AI is concentrating power among a small circle of firms with the cash to lock up chips, cloud capacity and data-center access. The deal signaled that the next phase of AI competition may be decided less by who builds the smartest model first and more by who can buy enough power to keep it running.