Hyperscaler

The big three clouds occupy every layer of the AI stack: they build and lease the datacenters, buy accelerators at the largest scale (increasingly alongside chips of their own: Trainium, TPUs, Maia), fund frontier labs through multi-billion-dollar partnerships, and sell both raw GPU capacity and managed model APIs.

Their model platforms (Amazon Bedrock, Azure AI Foundry, and Google Vertex AI) resell frontier and open models with enterprise wrapping: unified billing, private networking, regional residency, and compliance certifications. Per-token prices generally track the model creators' own rates, so the draw is contracting and governance rather than discounts, though committed-spend agreements can effectively subsidize usage.

For enterprises already inside one cloud, the hyperscaler route is often the path of least resistance to production AI; the strategic caution is concentration: compute, models, and infrastructure increasingly settle into a handful of vertically integrated stacks.