Microsoft and Chevron are Building a Texas Data Center Monster
·2 min read·Beginner
“
By 2028, they'll flip a switch on a plant that generates enough power to fuel over half a million Texas homes. Guess who wants all of it? Microsoft, obviously—because its AI data centers are basically power-eating monsters.
In 30 seconds
01Microsoft and Chevron build 2.67 gigawatt Texas plant coming online in 2028.
02
→
💡
What this means for you
It means the AI we use daily is insanely power-hungry, and tech companies are turning back to fossil fuels to feed it—not because they're evil, but because green infrastructure can't keep up. The energy transition is messier than LinkedIn posts want you to believe.
Smart search engine Perplexity AI needs serious muscle to grow. They just found it in Crusoe, who'll rent them a whole lot of computing power for years.
·2 min·2·Beginner
Power equals 530k homes but goes entirely to Microsoft's AI data centers.
0320-year natural gas deal because AI hunger outpaces renewable energy capacity.
Chevron and Microsoft just signed a 20-year power deal that basically turns West Texas into a dedicated energy plant for AI. The project is called Kilby (yes, named after the chip—tech companies have no creativity with names) and will hit 2.67 gigawatts when fully ramped up. To put that in perspective: it's like powering 530,000 Texas homes at once, but only to run Microsoft's AI models.
The plant runs on natural gas—not exactly the green energy poster child everyone tweets about, but hey, at least it's not coal from the 1950s. Chevron isn't doing this solo either; they brought in Engine No. 1, an investment fund that wants to play both the energy transition game and the follow-the-money game simultaneously.
Why all this infrastructure drama? Because AI data centers have become the infrastructure crisis of the decade. Tech companies keep promising to save the world with their models, but meanwhile they're consuming power like small countries. Microsoft's already had some rough PR moments—nearly got blocked in a few places because of water consumption—so they said: let's take this to the oil heartland and cut a deal with people who actually know energy.
📬 Enjoying this article?
Get the best AI news every week, straight to your inbox.
The plant should start delivering power by 2028 and ramp up gradually after that. Translation: Microsoft gets all the energy it wants to keep training increasingly massive AI models, and Chevron gets a 20-year customer paying premium rates. It's a marriage of convenience, and it's deeply awkward for anyone hoping renewable energy would save the planet in the next three years.
The real kicker? This is just one of many deals big tech companies are cutting with power providers worldwide. Amazon and Google are doing the same thing. The message is clear: GPU-hungry AI wants reliable, cheap, constant power—geothermal sounds nice on paper, but it can't keep up with the ask. So oil and gas, despite everything, are staying in the game.
This isn't a clean-tech success story. It's the story of how the real economy adapts when technology promises the world and consumes twice as much.
Picture an AI that doesn't just reply, but actually thinks like a seasoned sales pro. Salesforce Koa is here, and it’s set to make some big AI labs very uncomfortable.
Imagine boarding your train for work, only to find the tracks are a mess. In the Netherlands, it's not just a technical glitch, but something far more sinister.
