SpaceX is laying the groundwork for a factory that will cast natural gas turbine blades in-house, according to an exclusive report from The Information, as Elon Musk moves to secure power for the AI data centers that his companies are racing to build.
Musk Confirms the Plan For more context on this story, see our ongoing AI trends.
Musk confirmed the reporting in comments to the publication, saying SpaceX is now "doing in-house casting" for the blades and vanes used in natural gas turbines. According to Musk, bringing that manufacturing in-house can accelerate new gas turbines "coming online by up to 18 months" — a striking claim at a time when turbine deliveries have become a chokepoint for anyone trying to stand up new generation capacity quickly.
The Information characterized the effort as an attempt by Musk to bypass the power supply chain for AI data centers "in a way others assumed was impossible, by making highly complex components himself."
Why Blades and Vanes Matter
In a post on X responding to questions about the strategy, Musk laid out the logic. "SpaceX and Tesla are each building 100GW/year of solar production capacity as fast as possible, but natural gas will still be needed to supplement and bootstrap solar for several years," he wrote. "The limiting factor for nat gas turbine production is casting the blades & vanes."
The argument is straightforward: solar and storage can be manufactured at enormous scale, but the gas turbines needed to fill the gaps are constrained by a small number of precision-cast components. Turbine blades and vanes are made from superalloys shaped into complex aerodynamic profiles and must survive extreme temperatures and stresses for years. Casting them well is hard, and casting them at scale is harder still.
By standing up its own casting capability, SpaceX is betting it can shave as much as 18 months off the delivery timeline for new gas-fired generation — the difference, in practice, between an AI data center sitting idle for years versus coming online while demand is still climbing.
The Power Crunch Behind the Move
The AI industry's appetite for electricity has collided with the slow reality of grid expansion. Musk himself has pointed to a projected shortfall of around 15 gigawatts in AI compute power availability, framing infrastructure — not chips — as the binding constraint on the industry's growth.
The problem compounds across the supply chain. Data centers need chips, chips need power, and power depends on turbines, transformers, and transmission lines that take years to procure. Hyperscalers have responded with everything from revived nuclear plants to on-site generation, but Musk's approach is unusual in its ambition: rather than waiting in line for turbine manufacturers, SpaceX intends to make one of the hardest components itself.
Reports from Tom's Hardware and other outlets noted that the in-house manufacturing strategy is aimed specifically at cutting generator delays, with Musk's 18-month figure based on the time typically lost waiting for cast components in the turbine production cycle.
A Rocket Company in the Power Business
There is a certain logic to SpaceX doing this itself. The company already operates some of the most advanced casting and metallurgy operations in the aerospace industry, built to produce engine components that face conditions comparable to — or harsher than — those inside a gas turbine.
Secondary reports, including coverage from industry publications, suggest the foundry work is expected to be based in Texas, where SpaceX already maintains major manufacturing and launch operations. Neither SpaceX nor Musk has disclosed a timeline, budget, or expected production capacity for the facility.
The move also fits the broader arc of Musk's AI push. SpaceX's recent acquisition of the AI coding startup Cursor put the rocket company in direct ownership of a major AI product, and Musk has said his companies are building solar manufacturing capacity at a scale of 100 gigawatts per year each — figures that, if achieved, would rival the entire output of the global solar industry today.
What Comes Next
Musk responded to The Information's report directly, suggesting the company wanted the story accurate rather than secret — an indication that SpaceX sees its power manufacturing push as a strategic message aimed at the AI industry.
What comes next will determine whether this is a prototype or a product line. Casting turbine blades that meet the reliability standards of power generation is a certification-heavy discipline where tolerances are measured in microns and a single defective part can sideline a generating unit, and utilities and turbine OEMs will need convincing before flying SpaceX-cast hardware in critical generation assets. The superalloys involved — typically nickel-based compositions that resist creep and oxidation at temperatures well beyond those a rocket engine nozzle sees in sustained operation — demand process control that takes incumbent suppliers decades to refine. Until the company discloses capacity targets or customers, the factory remains a promise — but one aimed directly at what Musk sees as the single most restrictive bottleneck on AI growth: keeping the lights on.
For an industry that has spent two years treating compute as the scarce resource, the message from SpaceX is that watts, not FLOPs, are now the constraint that matters.
---
Stay Ahead of AIGet the latest AI news, analysis, and breakthroughs — all in one place.
Read more AI news →
