The Self-Licking Ice Cream Cone
Google is dumping up to $40 billion into Anthropic, with a floor of $10 billion and escalators tied to performance milestones. This follows Amazon’s $5 billion injection just days ago, valuing Anthropic at a staggering $350 billion. The core mechanism here is an incestuous loop: Google and Amazon invest cash, Anthropic uses that exact cash to buy cloud compute and TPU chips from Google and Amazon. It is a self-licking ice cream cone designed to inflate valuations while masking the fact that no one has figured out how to make generative AI sustainably profitable at scale.
Claude’s Growing Pains and the Convenience of a Crisis
Anthropic’s Claude models and tools like Claude Code are seeing real demand, driven partly by OpenAI’s ongoing reputation fires. But success has brought outages and capacity crunches, forcing Anthropic to throttle peak usage and consider stripping compute heavy features from cheaper tiers. This is framed as a scaling problem, but it is really a signal that the underlying economics are broken. Pouring $40 billion into a company that cannot yet afford to run its own inference at scale is not a vote of confidence. It is a bailout dressed up as a strategic partnership.
The Conflict That Dare Not Speak Its Name
Google and Anthropic are supposed to be competitors. Google has Gemini. Anthropic has Claude. Yet Google is now Anthropic’s largest patron. This is not synergy. It is a cartel style arrangement where the dominant cloud provider hedges against its own regulatory risk by owning a piece of the opposition. The AI industry has become a closed loop of hyperscalers funding each other’s rivals to maintain the illusion of competition while the real prize is locking developers into a single cloud ecosystem. If this were antitrust law, it would already be a test case.
Source: Arstechnica
