SpaceX is quietly moving into the power business, and the target is one of AI’s biggest bottlenecks. The company is laying the groundwork for a turbine-blade factory aimed at easing the data center power crunch, according to an exclusive report from The Information. That’s a striking pivot for a company known for rockets and satellites, and it tells you just how valuable electricity has become in the AI race.
What stands out here is who’s making the move. Elon Musk’s rocket company isn’t a utility or an energy startup. When a firm like SpaceX starts building the hardware that generates power, it’s a signal that the shortage is real and that existing suppliers can’t keep up.
⚡ Why this matters
AI runs on compute, and compute runs on power. Training and serving large models pulls enormous amounts of electricity, and the new wave of data centers going up around the U.S. needs gigawatts that the grid simply doesn’t have ready.
The squeeze has three parts:
- Demand is exploding. Every major lab and cloud provider is racing to build bigger clusters, and each one needs its own power supply.
- Grid connections are slow. Hooking a new data center to the grid can take years, so operators are turning to on-site generation instead.
- Turbine supply is tight. Gas turbines, the fastest way to spin up serious power, are back-ordered well into the back half of the decade.
That last point is the key. Turbine blades are precision components, and only a handful of manufacturers make them at scale. If SpaceX can produce them, it attacks the exact chokepoint that’s holding builders back.
🏭 The status quo before this
Until now, data center operators had a short list of options, and none were fast. They could wait in line for grid power, buy scarce turbines from established players like GE Vernova or Siemens, or strike deals for nuclear and renewable capacity that take years to come online. Microsoft, Amazon, Google, and OpenAI’s backers have all been scrambling to lock up electricity, sometimes reviving old nuclear plants or signing multi-year supply contracts.
SpaceX entering turbine manufacturing changes the math. The company has deep experience machining high-performance metal parts under extreme conditions, which is basically what rocket engines demand. Turbine blades sit in a similar engineering world: high heat, high stress, tight tolerances. The skills transfer better than they might first appear.
🔧 What the technology actually does
A gas turbine burns fuel to spin a shaft connected to a generator, and the blades are the parts that capture that energy as the hot gas rushes through. They have to survive brutal temperatures without warping or cracking. Making them well is hard, which is why so few companies do it. Solve the blade problem and you solve a big piece of the turbine supply problem.
For AI builders, that could mean on-site power arriving in months rather than years, without waiting on the grid or the existing turbine backlog.
🔭 What to watch next
This is early. The Information frames it as SpaceX laying groundwork, not flipping a switch, so treat the timeline with patience. A few things worth tracking:
- Whether SpaceX sells turbines to outside data center operators or keeps the capacity for Musk-linked projects like xAI’s Colossus supercomputer.
- How incumbents like GE Vernova and Siemens respond to a new entrant with serious manufacturing muscle.
- Whether other tech giants follow by vertically integrating their own power supply chains.
The bigger takeaway is that the AI power crunch is reshaping who builds what. Compute was the first constraint everyone talked about. Power is the one that actually caps how fast this industry can grow, and now one of the world’s most capable manufacturers is treating it as a problem worth solving in-house.
If you want the full details on SpaceX’s plans, they’re in the original report from The Information.