Semiconductor Factories in Orbit: Space Forge and the Autonomous Future of Chip Manufacturing

£22.6 million. That's what Space Forge raised to build a semiconductor factory. In orbit.

The idea of manufacturing in space is not new. NASA astronauts have run microgravity experiments on the ISS for decades. The difference now is that private companies are building autonomous platforms that operate without human presence. The company's ForgeStar-1 is a microwave-oven-sized satellite that launched on a SpaceX rideshare mission in June 2025 and spent months testing its onboard furnace. On December 31, 2025, it generated plasma at 1,000 degrees Celsius, the temperature needed for gas-phase crystal growth of semiconductor materials.

The case for orbital manufacturing

Terrestrial semiconductor manufacturing fights gravity at every step. During crystal growth, convection currents and impurities in the growth chamber introduce defects. In microgravity, those convection currents disappear. Atoms align more uniformly, producing crystals with fewer defects and higher purity. It targets wide-bandgap materials: gallium nitride, silicon carbide, aluminium nitride, and synthetic diamond, all critical for power electronics, 5G infrastructure, and quantum computing.

The company estimates that semiconductors grown in microgravity could cut energy use in electronic devices by up to 60%. The economic logic: a single defect-free silicon carbide wafer used in electric vehicle inverters or data center power supplies is worth more than its weight in launch costs. At $25,000 to $100,000 per kilogram to reach low Earth orbit, high-value semiconductor substrates are one of the few products where the orbital manufacturing premium pencils out.

The company's autonomous approach is deliberate. CEO Joshua Western told the Institution of Mechanical Engineers: "The last thing I want to introduce is a human into my manufacturing process." ForgeStar-1 and its successor ForgeStar-2 are designed as fully autonomous platforms that manufacture materials in orbit, deorbit, and return their payloads to Earth using a foldable heat shield called Pridwen. No ISS involvement. No astronaut labor. Just a satellite that flies itself, builds, and comes home.

The case against

The counter-argument is straightforward: orbital manufacturing competes against terrestrial fabs that improve every year. Silicon carbide wafers grown on Earth are getting cheaper and better. The defect density advantage of microgravity is real, but it narrows as Earth-based crystal growth techniques advance. And the orbital side has problems of its own: satellite launch delays, power constraints on a microwave-oven-sized platform, and the difficulty of scaling from a single furnace to industrial throughput.

It launched ForgeStar-1 in June 2025. The satellite generated plasma once in December 2025 and has since been running tests. That is a long way from weekly production runs. Its next satellite, ForgeStar-2, is supposed to bring back actual manufactured materials. There is no confirmed launch date. Competitors like Varda Space have flown multiple missions but focus on pharmaceuticals, not semiconductors. Every in-space manufacturing company is still proving the unit economics.

The broader market may not be as large as proponents claim. The "$20 billion opportunity" often cited for in-space manufacturing includes everything from pharmaceuticals to optical fiber to solar panels. The semiconductor slice is smaller. The company and its partners are positioning for a niche: ultra-high-purity substrates for defense, quantum computing, and specialized power electronics. This is not a replacement for TSMC.

What would need to be true

For orbital semiconductor manufacturing to become commercially significant, three conditions must hold. First, ForgeStar-2 or an equivalent platform must demonstrate repeatable production — not a single plasma strike but dozens of growth cycles with consistent output. Second, the cost per kilogram to orbit must continue falling; SpaceX's Starship and its competitors are on track to reduce launch costs below $1,000 per kilogram, which would change the math substantially. Third, a customer must pay a premium for space-grown substrates that justifies the complexity.

The company has signed partnerships with United Semiconductors for crystal growth processes and with Intuitive Machines for reentry vehicle integration. Those are real commitments. But the distance between a partnership announcement and a revenue-generating production line is measured in years and tens of millions of dollars. Its £10 million ESA grant in June 2026 helps, but the capital requirements for scaling orbital manufacturing are an order of magnitude higher. Compare this to the semiconductor industry's terrestrial capex: a single leading-edge fab costs $10-$20 billion. Orbital manufacturing does not need that scale to be viable, but it does need to prove that space-grown substrates command a price premium large enough to sustain a dedicated launch and operations cadence.

Where the debate lands

Orbital semiconductor manufacturing is not an either-or proposition. The most likely outcome is a hybrid scenario: Earth-based fabs handle the 99% of semiconductor demand that does not require ultra-high purity, while orbital platforms serve a narrow set of defense, quantum computing, and specialty power electronics applications. ForgeStar-1 proved the technology can work. The next two years will determine whether it can work at a commercially relevant scale.