Why Elon Musk is making turbine blades for AI data centres
SpaceX is reportedly building a Texas foundry for turbine parts as AI data centres drive demand for more reliable power.
AI is hungry for power, and the race to build more data centres is creating a surprising new bottleneck: turbine parts. That is where Elon Musk’s SpaceX comes in.
According to a Mint report, the company is setting up its own foundry in Texas to make vaned and high-spec gas turbine blades, key parts that help turbines generate electricity.
The move may sound far removed from rockets and AI, but there is a simple reason behind it. AI data centres need enormous amounts of reliable electricity, and getting new power infrastructure online is becoming increasingly difficult.
The power problem behind AI
Musk has said that one of the biggest bottlenecks for natural gas turbine production is casting the blades and vanes. Casting involves shaping metal in moulds, but turbine components require highly specialised manufacturing because they have to withstand extreme heat and pressure.
For AI companies, this matters because data centres cannot simply run on occasional power. They need a steady supply of electricity around the clock. Renewable energy is growing, but Musk has said natural gas will still be needed for several years while cleaner power capacity catches up.
SpaceX believes making turbine components itself could help move projects faster. Musk has claimed the approach could accelerate the arrival of new turbines by as much as 18 months, potentially cutting through one of the bottlenecks slowing new power capacity.
Why SpaceX is getting into turbines
The planned foundry will be built in Bastrop, around 48 kilometres east of Austin, close to an existing Starlink manufacturing facility. SpaceX reportedly acquired about 830 acres in the area between March and June 2026.
There is also a surprising connection between the company’s rocket business and turbine manufacturing. Morgan Stanley analyst Adam Jonas has pointed out that some of the alloys and furnaces used for SpaceX’s Raptor rocket engines are similar to those needed for data centre turbine components.
The Bastrop facility is expected to make single-crystal parts from advanced nickel-based alloys. These specialised components are designed to survive temperatures of up to around 1,980°C inside the hottest sections of turbines.
AI is pulling tech companies into energy
The bigger story here is not really about turbine blades. It is about how AI demand is starting to reshape the energy supply chain. Specialised foundries capable of producing these components are reportedly booked years in advance, with some capacity tied up through 2029 or 2030.
That leaves companies building new power projects waiting for parts before they can bring turbines online. SpaceX’s answer is to make the critical components itself. If the plan works, the company could gain more control over a part of the power infrastructure needed to support the wider AI boom.
The project is still being developed, and SpaceX has not announced an official operating date. Environmental regulatory filings are also pending. But the direction is already interesting: the AI race is no longer just about better models and faster chips. Increasingly, it is also about who can secure enough electricity to run them.