Google commits $10B immediately plus $30B contingent on performance targets, valuing Anthropic at $350B and adding 5GW of Google Cloud compute over five years.
Key Takeaways
Investors are now eager to back Anthropic at $800B or more; an IPO is reportedly being considered as early as October 2026.
Anthropic’s Mythos model – its most powerful to date with significant cybersecurity applications – has already leaked despite a limited partner-only release.
Google Cloud deal adds 5GW of new capacity over five years on top of a separate Broadcom/Google TPU partnership delivering 3.5GW from 2027.
Amazon closed an additional $5B investment this week, part of a broader arrangement where Anthropic will spend up to $100B for roughly 5GW of compute over time; CoreWeave data center deal also closed this month.
Recent Claude use-limit complaints are a direct symptom of the infrastructure crunch Anthropic is now racing to close through these stacked compute deals.
Hacker News Comment Review
Discussion is minimal; one commenter flagged this as a duplicate of an earlier thread, suggesting most substantive debate happened there rather than under this submission.
The one analytical comment focuses on OpenAI’s political positioning – specifically Altman aligning with Hegseth – as a reputational inflection point that may have shifted investor and enterprise sentiment toward Anthropic, though the commenter stops short of calling it the primary cause.
Notable Comments
@JumpCrisscross: argues Altman’s Hegseth alignment was “a pivotal moment in both companies’ fortunes” from investor and urban-consumer perspectives, even if not the root cause of OpenAI’s relative slide.
