Muon Space CEO on data center satellites, the launch cost curve, and why software is the hidden bottleneck in space

Summary

The Muon Space CEO makes a case that the space industry's real constraint is no longer launch cost — it's software.

The launch cost curve

About 15 years ago, getting a small satellite to orbit meant traveling to southeast Russia and launching on converted ICBMs. Today, the same payload rides a SpaceX Transporter mission at roughly one-tenth the cost. That's at least one, arguably two, orders of magnitude of improvement in a decade, driven almost entirely by SpaceX. Another order of magnitude is plausible — but getting there requires launch cadence, not just hardware capability. The Starship analogy is the airline model: a 737 flown five times a year makes tickets unaffordable; flown ten times a day, the fixed-asset economics work. Starship needs a demand anchor the way Starlink was for Falcon 9, something that pulls the vehicle to three, four, or five launches a day.

Space-based data centers

The CEO is broadly constructive on the concept. Space offers near-limitless solar power in sun-synchronous orbits where eclipse is avoidable, and a three-Kelvin cosmic background that functions as a free thermal sink for compute-heavy hardware. The physics are favorable. The blocker is getting mass to orbit cheaply enough that multi-ton, power-hungry payloads become economically rational. If Starship hits the next cost threshold, deploying AI training infrastructure in orbit becomes a credible first-order demand driver — and that demand, in turn, is what funds the launch cadence that brings costs down further. The loop is self-reinforcing if it starts.

Where the supply chain still breaks

Satellite buses are increasingly commoditized, drawing heavily on EV electronics — batteries, semiconductors — rather than traditional aerospace supply chains. The payload layer is the remaining hardware bottleneck: mission-specific instruments built in small quantities at high cost. But the CEO argues the less-discussed constraint is software. Satellites are complex autonomous robots operating in global networks, and the software that integrates avionics, ground stations, communications, and mission logic is still largely bespoke for every new program. The industry keeps rebuilding the same integration from scratch rather than treating the stack as a platform. His pitch for Muon Space is exactly that platform: a hardware-software foundation designed to interoperate from firmware to application layer, the way a data center rack does, so a customer building a data business doesn't have to reconstruct the full vertical for each new use case. That lesson, he says, came directly from Skybox, where the team effectively ran two companies — a satellite company and a data company — under one roof, when the data company should have been able to stand alone.