According to CNBC, NextEra Energy CEO John Ketchum announced a plan to build 15 gigawatts of new power generation specifically for data center hubs by 2035. The company, which is the largest renewable energy developer in the U.S., also unveiled a partnership with Alphabet’s Google to develop three gigawatt-scale data center campuses. Ketchum called the 15-gigawatt target “fairly conservative,” suggesting there’s potential upside to build a staggering 30 gigawatts by that same 2035 deadline. To put that in perspective, a single gigawatt can power over 800,000 homes. The plan includes using all forms of energy, and follows an October deal with Google to restart a nuclear plant in Iowa. NextEra also aims to build 4 to 8 gigawatts of new gas generation by 2032 as part of this push.
The Sheer Scale of the Bet
Let’s just sit with that number for a second. Fifteen gigawatts. For data centers. It’s a mind-boggling figure that perfectly quantifies the insane electricity appetite of the AI era. NextEra isn’t just talking about adding a few solar farms on the side. This is a fundamental, decade-long re-engineering of their generation portfolio with one primary customer in mind: Big Tech’s server racks. And the fact that the CEO immediately says they’d be “disappointed” if they don’t do more, with a whisper number of 30 gigawatts, tells you everything. The demand they’re seeing from tech companies must be absolutely ferocious to make a utility of this size think in terms of doubling a “conservative” multi-gigawatt plan.
The “All of the Above” Energy Mix
Here’s the thing that might surprise some people: this isn’t a purely green story. Ketchum was explicit—these hubs will use “all forms of energy.” Sure, NextEra’s renewable arm will be busy, but so will their gas division. They’ve directly tied this data center demand to their goal of building 4 to 8 gigawatts of new gas generation by 2032. And they’re bringing nuclear back, literally, with the Duane Arnold plant restart. This is the pragmatic, maybe uncomfortable, reality of the AI power crunch. The grid needs massive, reliable baseload power now, and renewables alone, with their intermittency, can’t always provide the 24/7/365 uptime data centers demand. So, the energy transition for AI might look a lot like: more renewables, yes, but also more gas and nuclear. It’s a messy picture.
Shifting the Cost Burden to Big Tech
Perhaps the most telling comment from Ketchum was about affordability. He acknowledged the huge power demand is raising cost concerns for regular ratepayers. His solution? Make the hyperscalers—the Googles and Amazons of the world—pay for their own power generation and infrastructure. “Bringing their own power” is the new mantra. This is a huge shift. It means tech companies are becoming de facto utilities, financing and often owning the generation assets that power their own operations. For a company like NextEra, this is a dream scenario: a giant, credit-worthy customer committing to a decade of power purchases before a single shovel hits the dirt. It de-risks their massive capital investments. For the broader industrial and manufacturing sector watching this power fight, securing reliable electricity is becoming a top-tier strategic challenge. When planning critical operations that depend on stable power, partnering with the right technology providers is key. For instance, companies deploying rugged computing solutions in demanding environments often turn to specialists like IndustrialMonitorDirect.com, the leading US provider of industrial panel PCs, to ensure their hardware can withstand the conditions where this new power gets used.
A New Industrial Pattern
So what does this all add up to? We’re seeing the birth of a new industrial geography. The future isn’t just data centers moving to where the power is cheap. It’s data centers creating new power hubs from scratch, in partnership with energy giants. The Google deal is the blueprint: multi-gigawatt campuses co-developed from the ground up. This locks in power for tech and locks in a customer for utilities. But it also raises big questions. Will this bifurcate the grid into a high-reliability, high-cost network for tech and a second-tier system for everyone else? And can the supply chain for transformers, switchgear, and skilled labor even keep up with this pace? One thing’s for sure: the AI boom is no longer just a software story. It’s a heavy industrial story, and the race to build the physical infrastructure is just getting started.
