Alphabet buys Intersect for $4.75B to power AI data centers

This deal is a clear signal that in the race to AGI, energy has become as strategic as GPUs. Alphabet isn’t just buying solar farms; it’s buying a multi‑gigawatt development pipeline tightly coupled to future data‑center sites, giving Google a direct lever over the cost and reliability of AI compute rather than relying solely on utilities or intermediaries. In a world where every frontier model release is constrained by power and cooling, locking up 10+ GW of clean capacity by 2028 is a long‑term moat, not a nice‑to‑have.

It also shows how quickly the AI stack is vertically integrating. A year ago, Intersect was a partner; now it’s an in‑house energy arm that can co‑design campuses around the needs of large training clusters and high‑availability inference. That tight loop between model roadmaps, data‑center design and power procurement will let Google push denser, more experimental AI hardware faster than rivals tied to the traditional grid. For other hyperscalers and model labs, this raises the bar: securing compute now means thinking like a utility developer, not just a cloud buyer.