Google's acquisition of Intersect Power for $4.75 billion marks a fundamental shift in how hyperscalers approach their most pressing constraint: electricity. Rather than simply buying power from the grid, tech giants are now building their own energy infrastructure alongside their data centers. This vertical integration reflects a hard reality facing the AI industry: there is no other way to secure reliable power in the quantities needed to fuel the next generation of artificial intelligence systems.

Why Are Tech Giants Suddenly Building Power Plants?

The AI infrastructure boom has created an unprecedented demand for electricity. In the first quarter of 2026 alone, wholesale power prices on the PJM grid (which serves much of the eastern United States) rose 76 percent year over year, while US residential electricity rates climbed to 17.45 cents per kilowatt-hour, up 9.5 percent. Meanwhile, the interconnection queue for new power projects has stretched to approximately 54 months, meaning companies can wait over four years just to connect to the grid.

The Big Five tech companies (Microsoft, Alphabet, Amazon, Meta, and Oracle) are guiding for $660 to $690 billion in capital expenditures in 2026, with roughly 75 percent of that amount directed toward AI infrastructure like servers, graphics processing units (GPUs), and data centers. That level of spending creates an urgent problem: the power grid cannot keep pace. Rather than wait for utilities to build new capacity, hyperscalers are taking matters into their own hands.

Google's strategy with Intersect reflects this urgency. The company announced the Meitner Energy Center in Gray and Roberts Counties, Texas, which will feature a data center co-located with over one gigawatt of new wind, solar, and battery storage capacity, plus on-site natural gas generation to ensure reliable operations. The dedicated power generation will serve the data center directly and reduce strain on the local grid.

"Five years ago, vertical integration didn't make sense. Now it does, because there is no way to procure the option on power in this country unless you own it," noted Caroline Golin, Google's former global head of energy.

Caroline Golin, Former Global Head of Energy at Google

How Are Different Hyperscalers Securing Power for AI Data Centers?

Each major tech company is pursuing a distinct strategy to lock in the electricity their AI operations require:

• Google's Approach: Acquiring renewable energy developers like Intersect to gain development capabilities and claim uncommitted power resources where and when needed, as demonstrated by the Meitner Energy Center project in Texas.
• Amazon's Strategy: Acquiring data centers already positioned near energy assets or purchasing the energy assets themselves outright, such as the Sunstone solar and storage project in Oregon with 1.2 gigawatts of capacity in each category.
• Microsoft's Method: Leveraging capital to fund exceptionally large corporate power purchase agreements, most notably a $16 billion, 20-year deal with Constellation Energy to restart a nuclear reactor at Three Mile Island in Pennsylvania.

These divergent approaches reflect different philosophies about control and risk. Google's acquisition of Intersect gives it the flexibility to develop new projects. Amazon's direct ownership of energy assets ensures long-term supply security. Microsoft's massive power purchase agreements with nuclear operators provide stable, carbon-free electricity at scale.

What Does This Mean for the AI Industry's Energy Crisis?

The shift toward vertical integration signals that hyperscalers view power as a strategic asset, not a commodity they can simply purchase from the market. The physical grid is beginning to pass AI's power demand directly to consumers and voters, and the response has not been uniformly positive. Rural American communities, where the majority of planned AI infrastructure is set to be built, have begun organizing protests against new data center projects.

Google's emphasis on advanced air-cooling technology to limit water consumption at the Meitner Energy Center reflects another mounting concern: data center water use. The company publicized responsible stewardship practices around water earlier in June 2026, an attempt to address mounting community pushback against data center expansion across the country.

The capital intensity of AI infrastructure has reached unprecedented levels. The largest tech platforms now allocate 45 to 57 percent of their revenue to capital expenditures, a level previously unthinkable for software businesses. This spending is not sustainable through traditional financing alone. Hyperscalers raised approximately $108 billion in debt during 2025, with projections of roughly $1.5 trillion in issuance over the coming years.

Interestingly, the circular financing among vendors, AI model creators, integrators, and chip manufacturers has reached new heights. Nvidia committed up to $100 billion to OpenAI, while OpenAI builds data centers capable of holding as many as 50,000 Nvidia GB200 chips each, with Oracle Cloud infrastructure powering flagship sites and CoreWeave providing infrastructure rental services. These interlocking agreements exceed $800 billion in total commitments.

The Intersect acquisition represents more than a single transaction. It signals that the era of hyperscalers as pure software companies is ending. They are becoming energy companies, infrastructure developers, and power generators. For the AI industry to continue its explosive growth, it must solve the power problem. Google's bet suggests that owning the solution is the only viable path forward.