Amazon and Alibaba pursue divergent AI strategies, with very different implications for crypto infrastructure

Two of the world’s largest tech companies are making billion-dollar bets on artificial intelligence. They just happen to be betting on completely opposite things. Amazon is shutting down its in-house AI model development. Alibaba is doubling down on it. The divergence tells us something important about where AI infrastructure is heading, and why crypto markets should be paying attention. Amazon: the landlord play Amazon closed its AGI Lab and began phasing out most of its in-house Nova AI models in late July 2026. The San Francisco research unit saw layoffs, and the company redirected those resources toward what it does best: infrastructure. The price tag for this pivot is not small. Amazon plans capital expenditures approaching $220 billion for 2026, with the bulk of that aimed at scaling AI infrastructure. For context, that’s roughly the GDP of Greece being funneled into servers and data centers in a single year. The strategy positions AWS as a multi-model hosting platform, essentially becoming the grocery store of AI compute. Enterprises that don’t want to be locked into a single model provider can shop around, running different models for different tasks, all on Amazon’s infras...