Google commits up to $40B to Anthropic

If you have run into Claude's stubborn usage limits lately, Anthropic just found forty billion reasons why that might change soon.

Alphabet plans to invest up to $40 billion in Anthropic through a mix of cash and compute, according to reports from Bloomberg and TechCrunch. Google will commit $10 billion immediately at a $350 billion valuation, with the remaining $30 billion tied to specific performance milestones.

The investment comes right as Anthropic is experiencing unprecedented growth. Driven heavily by developer uptake of its Claude Code agent, Anthropic's annual run-rate revenue recently passed $30 billion, up dramatically from $9 billion at the end of 2025. But that explosive demand has pushed its existing infrastructure to the absolute brink.

Cash, compute, and Claude limits

To keep its models running, Anthropic has been on a massive infrastructure shopping spree. Earlier this month, the company inked a data center deal with CoreWeave and locked in fresh commitments from Amazon, where Anthropic could spend up to $100 billion over time for about 5 gigawatts of compute capacity.

Google's new agreement raises the stakes even further. Beyond the funding, Google Cloud is pledging an additional 5 gigawatts of capacity over the next five years, with options to scale upward.

Anthropic has long leaned on Google Cloud and its specialized Tensor Processing Units (TPUs), widely regarded as the most viable alternative to Nvidia's scarce silicon. Regulatory filings from Broadcom, which co-designs Google's custom AI chips, previously revealed a multi-year deal ensuring Anthropic 3.5 gigawatts of TPU capacity starting in 2027. This new injection effectively doubles down on that pipeline.

Hardware needs will only intensify with Anthropic's latest flagship model, Mythos. Released to select partners this month, Mythos boasts major cybersecurity capabilities, but running frontier models at that scale is notoriously punishing on compute.

Frenemies with multi-gigawatt benefits

The dynamic between Google and Anthropic remains one of the weirdest rivalries in tech. In the consumer and enterprise chat arena, Gemini and Claude fight for the exact same customers every single day.

Yet beneath the surface, Google is happy to play arms dealer. By funneling gigawatts of TPU compute directly into Claude's engine, Alphabet gets a guaranteed, massive tenant for its cloud business while riding Anthropic's equity upside.

Anthropic is already looking ahead to public markets. While Google's initial $10 billion anchors at the $350 billion valuation from February, Bloomberg reports some investors are clamoring to value the startup at $800 billion or more, with an IPO potentially targeted as early as October.

Why it matters

For everyday developers, this is immediate breathing room. Claude Code has become an indispensable workflow tool for many engineering teams, but frequent rate limits have made production reliance painful. Google's 5 gigawatts of dedicated compute should finally give Anthropic the headroom to open up heavier models like Mythos without crippling its core services.

For the AI ecosystem, the battle lines are now defined by sheer physical scale rather than just model weights. OpenAI has built an sprawling infrastructure web, and Anthropic is pulling off an impressive balancing act by keeping both Amazon and Google deeply committed to its roadmap. Playing two cloud titans against each other avoids single-vendor lock-in, which we think is a brilliant hedge.

Still, the operational risks are real. Renting your primary weapon from your fiercest competitor creates constant strategic tension. Moreover, multi-gigawatt buildouts take years, and Anthropic is taking on staggering spending commitments. Heading into a rumored fall IPO, it will have to prove that these eye-watering compute bills translate into sustainable margins, not just endless server burn.