Starcloud, a burgeoning space compute company, has recently completed a funding round that values the firm at $1.1 billion, marking its rapid ascent to unicorn status following its Y Combinator graduation. The Series A funding, which wrapped up 17 months after its demo day, was spearheaded by Benchmark and EQT Ventures. This achievement reflects growing interest in the concept of orbital data centres, especially as terrestrial data centre development faces mounting resource and political challenges. However, the business model remains contingent on emerging technologies and substantial capital investment.
To date, Starcloud has raised a total of $200 million and successfully launched its inaugural satellite, equipped with an Nvidia H100 GPU, in November 2025. This year, the company plans to deploy a more advanced satellite, Starcloud 2, featuring multiple GPUs, including the Nvidia Blackwell chip, an AWS server blade, and a bitcoin mining unit.
The firm is also embarking on a project to create a data centre spacecraft, Starcloud 3, designed for launch from SpaceX’s Starship, a reusable heavy-lift rocket currently under development. Johnston, the CEO and founder, anticipates that Starcloud 3 will be the first orbital data centre capable of competing with earthly counterparts in terms of cost, projecting expenses of around $0.05 per kWh should commercial launch costs reach approximately $500 per kilogram.
Challenges remain, particularly as Starship has yet to make its inaugural flight. Johnston expects that commercial access to this technology may not materialise until 2028 or 2029. He acknowledges that the costs associated with space data centres will remain high until a new generation of rockets achieves frequent operational launches, potentially not until the 2030s.
He highlights two distinct business models for Starcloud: firstly, selling processing power to other orbiting spacecraft, as demonstrated with its initial satellite’s collaboration with Capella Space; and secondly, the potential future of distributed data centres in space supplementing their terrestrial equivalents when launch costs decrease.
The space data industry is nascent, and the number of advanced GPUs currently in orbit is limited. In contrast, in 2025, Nvidia is projected to sell nearly four million GPUs to terrestrial hyperscalers. Johnston asserts that Starcloud is ahead of the competition, being the first to deploy a terrestrial GPU in space, used to train AI models in orbit.
Going forward, Starcloud faces numerous technical hurdles, such as efficient power generation and cooling for high-performance chips. Starcloud 2 is designed to carry the largest deployable radiator on a private satellite, and several more iterations of this spacecraft are planned. Synchronisation for extensive data workloads presents additional challenges, necessitating either large spacecraft or robust inter-satellite communication.
In addition to Starcloud, companies like Aetherflux, Google’s Project Suncatcher, and Aethero are also exploring the space data centre sector. Notably, SpaceX has requested permission from the U.S. government to construct and operate a million satellites for distributed computing in space.
Despite the daunting prospect of competing with SpaceX, Johnston remains optimistic about the coexistence of their business models, differentiating Starcloud’s focus on energy and infrastructure from SpaceX’s primary goals.
Fanpage:Â TechArena.au
Watch more about AI – Artificial Intelligence
