Orbital signs a bus partner, but the license is the asset

Orbital Compute has chosen the company that will build its satellites, naming Reflex Aerospace, the German manufacturer, platform partner for Orbital-1—the constellation Orbital Compute says it will operate as a 100,000-satellite orbital data center—and handing Reflex its first US customer.

The instrument is a development agreement rather than a purchase, with Reflex leading integration of the spacecraft from a bus design it already flies and making that design, along with the technologies used across the constellation, exclusive to Orbital. No price appears in the announcement; in this sector position gets disclosed and price withheld, and unpriced announcements are how positions get staked.

What Reflex brings, according to founder and CEO Euwyn Poon, is a proven bus platform and an engineering team able to adapt it for computing in orbit, where a satellite needs hundreds of kilowatts of power, substantial heat rejection, and volume production designed in from the first day.

The specification behind that adaptation is large: Orbital says its satellites will approach 250kW of power apiece and deliver 100kW of compute, which across 100,000 units adds to the 10GW the company cites for the constellation. Engineering teams in Los Angeles and Berlin will work as one integrated group to build it.

The pathfinder demonstration satellite, carrying Nvidia's Blackwell chip to test GPU operation, radiation tolerance, thermal performance, and data downlink, is intended to fly in 2027; Poon has described the demonstrator as perhaps one one-hundredth the size of the Orbital-1 spacecraft it precedes.

Orbital filed with the FCC for the constellation in early July and is awaiting the regulator's approval. The filing went in weeks before the FCC voted to overhaul its licensing process into what it calls an "assembly line" model, with a "presumed acceptable framework" that would grant applications such as Orbital's immediately, without the traditional cross-examination.

Against that ambition, the capital is thin: Orbital closed a $5 million pre-seed in mid-June across 15 investors, which buys engineering and regulatory position rather than production slots and is an immaterial sum against any hardware program of this size.

Data center developers compete for scarce power, land and water, Poon says, and space offers a way past those terrestrial constraints, with abundant solar energy; the spacecraft needed to capture it has not flown.

That framing cuts against the usual reading of a deal like this: on the ground, grid access is the asset and generation and compute are derivatives of connection consent, with developers queueing for a connection they do not control. In orbit the scarce good is the same thing wearing a different form: solar flux is free and effectively unlimited, but orbital slots and spectrum are rationed by a regulator, and Orbital's regulator is midway through rewriting how it hands those rights out.

The Reflex agreement, then, matters less for the bus than for the credential: a named European platform builder with a design it already flies, bound exclusively to Orbital, gives the FCC and any future financier a hardware roadmap behind the filing. Exclusivity is the point of the trade—a claim that the engineering is real and this is not another paper constellation.

For Reflex, the value runs the other way: an experienced European builder gains its first US customer and a marquee program to point at, at the cost of making its bus and constellation technologies exclusive to a startup that has yet to fly anything. That is an unusual concession for a manufacturer with a working product, and it suggests Reflex is pricing the exclusivity as a bet on Orbital's licensing position rather than on near-term orders.

The rest of the calendar is set by other people—the FCC vote already happened, the pathfinder flies in 2027 if it flies, and the 10GW does not exist until 100,000 buses do, which leaves Orbital with $5 million and a partner in Berlin to show for it.

The Reflex agreement puts no hardware in orbit by itself; what it does is keep a 100,000-satellite application legible at the exact moment the regime that will decide whether it is ever granted is being rewritten to favor applications like it. The next test is whether the FCC grants that filing before the pathfinder flies in 2027.