Google’s $4.75 billion purchase of data center and energy developer Intersect marks one of its boldest bets yet on the infrastructure needed to keep the artificial intelligence boom running. The all‑cash deal is designed to secure both computing capacity and electricity at a time when power‑hungry A.I. systems are straining grids and forcing tech giants into increasingly creative financing and construction schemes.

Google is acquiring Intersect, a privately held San Francisco company that both operates data centers and develops energy projects, in a transaction valued at $4.75 billion in cash. The companies were hardly strangers: Google had already invested in Intersect a year earlier and was using it as a partner to build new data center sites before deciding to bring much of that work in‑house.

The deal is explicitly framed as part of Google’s push to expand the infrastructure behind its artificial intelligence services, which rely on massive, power‑intensive computing clusters. As competition with Amazon, Meta, Microsoft and OpenAI accelerates, securing reliable power and data center capacity has become as strategically important as developing the A.I. models themselves.

Intersect occupies a rare niche: it develops data centers and the energy generation needed to power them, tightly linking two sides of the A.I. infrastructure problem that are usually handled by separate companies. By acquiring Intersect, Google is effectively buying a toolkit for building new facilities and their accompanying power supplies in tandem, rather than stitching together multiple outside partners.

Sundar Pichai, Google’s chief executive, said Intersect would help the company “expand capacity” and “operate more nimbly in building new power generation in lock step with new data center load,” underscoring how critical power planning has become to the A.I. race. He also cast the deal as part of a broader push to “reimagine energy solutions” and bolster U.S. technological leadership.

A.I. arms race and new financing tricks

Across the industry, the rush to dominate A.I. has triggered a global building spree, with Google, Amazon, Meta, Microsoft and OpenAI committing billions of dollars to new data centers. These facilities can cost tens of billions collectively, and the companies have experimented with new financial structures that shift debt onto smaller partners to keep the largest obligations off their own balance sheets.

Against that backdrop, an outright acquisition of a data center and energy developer stands out. Rather than relying only on partners and creative financing, Google is choosing direct ownership of key assets, signaling that it sees control over infrastructure as a long‑term strategic advantage worth the upfront cost.

Regulatory shadows, but limited penalties

Google’s merger activity has been restrained in recent years as regulators in the United States scrutinized whether it was illegally maintaining monopolies in areas like internet search. That pressure appeared to dampen its appetite for large acquisitions even as rivals pursued high‑profile deals of their own.

Although courts have found that Google violated antitrust laws, the resulting penalties so far have been relatively light, clearing some of the uncertainty that had hung over big transactions. The Intersect acquisition suggests Google is more confident that it can navigate regulatory review when the deal is framed as an infrastructure investment rather than a move to buy a direct competitor in advertising or search.

What Google gets – and what it doesn’t

Under the terms Google outlined, the company will acquire some of Intersect’s employees, a set of data center projects and “multiple gigawatts” of energy capacity. That scale is significant: gigawatts of power are what underpin A.I. clusters capable of training and running frontier‑scale models.

Yet Intersect will not disappear entirely into Google. The company will continue to operate as an independent business and will keep data centers in Texas and California that serve other customers, preserving a portion of its pre‑existing client base outside of Google.

Texas as a testbed

Google singled out Haskell, Texas, as a major beneficiary of the deal. There, the company is investing $40 billion through 2027 to build out A.I. infrastructure, and it said Intersect’s assets and expertise would help accelerate that timeline.

Texas has already become a magnet for data center and energy projects, thanks to abundant land and a power grid that, while sometimes strained, is relatively open to large industrial users. For Google, demonstrating that it can integrate Intersect’s capabilities quickly in Haskell will be an early test of whether the acquisition pays off in the form of faster, more reliable A.I. build‑outs.

A rare kind of deal in the A.I. era

In today’s A.I. boom, many of the most talked‑about deals involve software, model licensing and partnerships between tech platforms and model developers. Google’s move for Intersect is different: it is a bet on shovels and power lines rather than just algorithms and cloud contracts.

The acquisition underscores a simple but often overlooked reality of generative A.I.: behind every conversational chatbot and image generator are immense industrial systems that need electricity, land, and hardware on a scale more familiar to utilities than app developers. Owning a larger slice of that physical infrastructure could shape how quickly — and how sustainably — Google can grow its A.I. business in the years to come.