Google to invest up to $40bn more in Anthropic, deepening TPU partnership

Google agreed to invest up to $40 billion in Anthropic, combining cash and compute credits, in a deal that valued Anthropic at $350 billion. Of the total, $10 billion arrived as an immediate cash investment; the remaining $30 billion was contingent on Anthropic’s compute consumption and unspecified milestone triggers, rather than committed outright.

The deal deepened an existing relationship rather than starting a new one: Google Cloud committed 5 gigawatts of additional TPU capacity for Anthropic, coming online from 2027, on top of TPU commitments the two companies had already announced earlier in April. Anthropic has relied on Google’s tensor processing units, alongside Nvidia GPUs, as one of two main compute sources underpinning its model training and serving.

The announcement came four days after Amazon, Anthropic’s other major cloud and compute backer, said it would invest a further $5 billion immediately with up to $20 billion more tied to commercial milestones — meaning Anthropic secured roughly $65 billion in committed capital and compute across a 17-day span in April 2026. The pace of investment reflected both the scale of capital frontier AI labs required to keep training and serving ever-larger models, and continuing reports that investors were seeking to back Anthropic at valuations well above the $350 billion figure struck in this round, with some discussions reportedly reaching $800 billion or higher.