Starcloud’s FCC paperwork covers 88,000 spacecraft. The company employs 25 people.
That mismatch explains most of what matters about the startup’s fresh funding. Starcloud said it has attached a $250 million extension to the $170 million Series A it wrapped up in March, a round that values the company at $2.3 billion. The money funds a larger factory and accelerates Starcloud-3, the biggest orbital data center spacecraft in the company’s lineup, which is designed to ride SpaceX’s upcoming Starship rocket.
The more revealing reason CEO Philip Johnston is hoarding capital, though, has nothing to do with manufacturing. It’s a mounting fear that no ride will be available.
The launch squeeze is the real story
“We can see what’s coming—we’re going to need to book an enormous amount of launch,” Johnston said.
Falcon 9, the workhorse that turned cheap smallsat deployment into a routine affair, is slated for retirement in 2028. SpaceX intends to swap in Starship, a far bigger vehicle that remains unproven. Anyone assembling a satellite constellation across a multi-year horizon is left with a planning problem that has no tidy solution.
“As soon as we can, we want to get under contract with things like Starship,” Johnston said. “One of the biggest costs is now on securing your launch capacity….launch is pretty constrained right now because [SpaceX’s] Falcon 9 program is scheduled to end in 2028.”
Rival options remain immature. Neither Blue Origin’s New Glenn nor ULA’s Vulcan has settled into a regular cadence. Rocket Lab’s Neutron has yet to reach the pad. Launch economics were already the thorniest piece of the orbital data center argument — thorny enough that one startup in the sector opted to build rockets of its own instead of waiting around.
Starship’s timeline, meanwhile, keeps sliding. SpaceX CEO Elon Musk said this week that the company is pushing back an attempt to catch a returning Starship by a few months, with a first re-flight of the vehicle now targeted for the end of the year or early 2027.
Publicly, Johnston isn’t doing much hedging. He says he stays confident that SpaceX can demonstrate rapid, repeated reuse of the most powerful rocket ever built. He’s equally clear-eyed about the fallout if that doesn’t materialize.
“Obviously if we can’t book any SpaceX launch capacity in 2029, that will be that will be challenging for us,” he said.
What actually flies first
Ahead of any Starship-class hardware, Starcloud plans to loft a pair of 8 kw compute satellites — dubbed Starcloud-2 — aboard rideshare missions in 2027. They’ll handle orbital inference workloads for customers that include US government agencies. The company is also mulling a dedicated Falcon 9 purchase to orbit additional spacecraft, and is exploring contracts with other providers for missions further out.
That covers the near-term business. Everything about the long-term thesis hinges on Starship pushing launch costs low enough that an orbital inference layer can undercut, or at least match, ground-based data centers on price.
Nvidia wrote a check, and Johnston thinks he knows why
Manhattan West Ventures headed up the extension. Nvidia and Cisco took part, and a person familiar with the deal said Nvidia contributed $25 million. Benchmark, EQT, Soma, NFX, 776, Cedar Capital, Goanna Capital and Standard Capital rounded out the participants.
Johnston reads the Nvidia investment as an endorsement of something quite specific rather than a broad vote of enthusiasm. Starcloud is the only known operator of an Nvidia H100 — a terrestrial data center GPU — in orbit, and the first to train a model on one. Most other GPUs flying today are edge processing parts, which serve an entirely different purpose.
What Starcloud has learned is now flowing back to Nvidia as the chipmaker develops its first GPU purpose-built for space, the Vera Rubin Space-1.
“The reason they’ve chosen to do this investment now is because of all of this data that we got from Starcloud One,” Johnston said. “They, more than any other VC, did way more technical duty on this than anybody else.”
The chip doesn’t exist yet
Vera Rubin Space-1 has yet to be built. Starcloud is aiming to fly it in late 2028, and its engineers are tracking a short list of design choices that will decide whether that holds: the relationship between the chip’s operating temperature and the size of the radiators shedding that heat, the placement of radiation shielding, and how much ruggedizing the components require to survive a launch.
None of that is glamorous, yet those constraints determine whether a data center GPU can survive above the atmosphere long enough to be worth anything.
The company’s 25 employees are standing up production lines inside a 100,000 square foot facility in Woodinville, Washington — just down the road from the sites where SpaceX and Amazon assemble satellites for their own communications networks. Hiring continues. Access to that talent pool is precisely why the address matters.
If you’re looking for a number to track, ignore the valuation. Watch instead for whether Starcloud inks a Starship contract before Falcon 9 is retired in 2028. Every other part of this plan flows from that signature.