Google is committing up to $40 billion to Anthropic in cash and compute, according to TechCrunch AI, citing a Bloomberg report. The Alphabet subsidiary is putting $10 billion in now at a $350 billion valuation, with another $30 billion tied to Anthropic hitting specific performance targets. TechCrunch AI also notes that Google Cloud will provide a fresh 5 gigawatts of capacity over the next five years, with room to scale further.
This lands in a week where Anthropic is already swimming in capital and compute deals. The company just secured an additional $5 billion from Amazon as part of a broader arrangement that could see Anthropic spending up to $100 billion for roughly 5 gigawatts of compute over time. Earlier this month it locked in data center capacity with CoreWeave, and a separate partnership with Google and Broadcom maps out 3.5 gigawatts of TPU-based capacity starting in 2027.
Why Google is writing this check
Google is a direct competitor to Anthropic in frontier models. It’s also one of Anthropic’s most important suppliers. Anthropic leans heavily on Google Cloud for chips and infrastructure, including Google’s tensor processing units (TPUs), the custom AI chips that have become the most credible alternative to Nvidia’s GPUs.
That dual role explains the size of the investment. Google gets:
- A bigger equity stake in one of the two leading model labs
- A long-term cloud customer locked into TPU-based infrastructure
- A counterweight to OpenAI’s tightening orbit around Microsoft, Oracle, and now Cerebras
For Anthropic, the cash unlocks the compute it desperately needs to keep Claude running at the scale users expect.
The compute squeeze is the real story
The AI race is now defined less by model architecture and more by who can secure the chips, energy, and data centers to train and serve these systems. OpenAI has spent months stitching together multi-hundred-billion-dollar deals across cloud providers, chip suppliers, and power generation. Anthropic has been racing to do the same.
That scramble has been visible to users. Claude customers have complained loudly about usage limits in recent weeks, and Anthropic’s wave of infrastructure announcements is a direct response. The Google deal isn’t just a financial event. It’s an attempt to fix a capacity problem that’s costing Anthropic mindshare with developers.
Mythos in the background
The investment news arrives just after Anthropic released Mythos, which the company calls its most powerful model to date, to a limited group of partners. TechCrunch AI reports Mythos has significant cybersecurity applications, and Anthropic has restricted broader access while it evaluates misuse risks with select organizations. The model has reportedly already leaked into unsanctioned hands. Running it at scale is also expected to be expensive, which makes the timing of a fresh 5 gigawatts of Google capacity look very deliberate.
Valuation and an IPO on the table
Anthropic’s valuation sat at $350 billion as recently as February. Per Bloomberg’s reporting cited by TechCrunch AI, investors are now eager to back the company at $800 billion or more. Anthropic is reportedly weighing an IPO as soon as October.
What stands out here is the velocity. Anthropic has more than doubled its private valuation in months, layered on tens of billions in fresh capital, and is now openly flirting with the public markets while shipping its most capable model yet.
What to watch next
- Whether Anthropic actually triggers the second $30 billion tranche, and on what timeline
- How Google balances its dual role as competitor and supplier as Gemini and Claude keep shipping
- Whether the new 5 GW of Google Cloud capacity is enough to ease Claude’s rate limits
- Movement toward an October IPO filing
The practitioner takeaway is straightforward. If you’re building on Claude, expect more headroom on capacity in the coming months but also expect Anthropic to push hard on monetization as it heads toward public markets. The model labs aren’t slowing down. They’re getting more expensive to compete with by the week.
Full details on the deal are at the original TechCrunch AI report.