Starcloud secures $250M to build orbital AI data centers amid launch crunch

A growing competition for space-based infrastructure is unfolding as Starcloud, a startup developing satellites capable of performing AI inference in orbit, announces a $250 million extension to its existing $170 million Series A funding round. The new capital brings the company’s total valuation to $2.3 billion and will fund the expansion of its manufacturing facility in Woodinville, Washington, as well as the development of its largest orbital data center spacecraft, Starcloud-3. The company plans to launch this next-generation satellite aboard SpaceX’s Starship rocket, a vehicle that, despite its promise of reduced launch costs, remains unproven and delayed.

The urgency behind Starcloud’s fundraising stems from a tightening launch market. SpaceX’s workhorse Falcon 9 rocket is scheduled to retire in 2028, leaving a gap that other providers like Blue Origin’s New Glenn and ULA’s Vulcan have yet to fill reliably. Rocket Lab’s Neutron and other new vehicles are still in development, creating uncertainty for satellite operators. Starcloud’s CEO, Philip Johnston, has already filed an FCC application to operate 88,000 spacecraft, underscoring the company’s long-term ambitions. In the near term, Starcloud is targeting 2027 for the launch of its Starcloud-2 satellites, which will perform orbital inference tasks for U.S. government customers. The company is also exploring dedicated Falcon 9 launches and alternative providers to secure additional capacity.

Starcloud’s technical edge lies in its ability to run Nvidia’s H100 GPU in orbit, a feat Johnston claims is unmatched by competitors. Most space-grade GPUs are designed for edge processing rather than full-scale AI inference, but Starcloud has already trained models using the H100 in space. This experience is informing Nvidia’s development of its first space-optimized GPU, the Vera Rubin Space-1, which Starcloud hopes to deploy by late 2028. The company is collaborating closely with Nvidia on thermal management, radiation shielding, and ruggedization to ensure the chip can survive the harsh conditions of space and rocket launches.

The investment round was led by Manhattan West Ventures, with participation from Nvidia, Cisco, Benchmark, EQT, and several other venture firms. Nvidia contributed $25 million, a move Johnston describes as a strong endorsement of Starcloud’s technical leadership. The company’s current workforce of 25 employees is expanding as it scales production in its 100,000-square-foot facility near Seattle, positioning itself alongside other major players like SpaceX and Amazon in the satellite manufacturing ecosystem.

However, significant challenges remain. The Starship rocket, which Starcloud is banking on for cost-effective launches, has yet to achieve rapid reusability, a key factor in driving down launch expenses. SpaceX has delayed its first attempt to catch a returning Starship by several months, with the first reflight now expected no earlier than late 2026 or early 2027. Johnston acknowledged that if SpaceX cannot secure launch capacity by 2029, Starcloud’s timeline could face serious disruptions. The company is also navigating the complexities of regulatory approvals, as operating tens of thousands of satellites requires extensive coordination with agencies like the FCC.

Competition in the orbital AI inference space is intensifying. While Starcloud focuses on high-performance computing in orbit, other startups are pursuing alternative approaches, such as developing their own rockets to bypass launch constraints. The broader industry is watching closely as companies race to establish the first commercially viable orbital data centers, which could redefine how AI workloads are processed and reduce reliance on terrestrial infrastructure.

For now, Starcloud is prioritizing near-term missions, including the 2027 launch of its Starcloud-2 satellites. These spacecraft will serve government clients, demonstrating the feasibility of orbital AI inference. Longer term, the company aims to build a scalable orbital inference layer capable of competing with terrestrial data centers, leveraging Starship’s potential to slash launch costs. Whether this vision materializes depends on SpaceX’s ability to deliver on its promises and the broader industry’s ability to overcome technical and regulatory hurdles.

The funding round also highlights the strategic interest of major tech players in space-based computing. Nvidia’s investment signals confidence in the sector’s growth, while Cisco’s involvement suggests broader enterprise interest in distributed, high-performance infrastructure. As Starcloud expands, it will need to balance rapid scaling with the realities of a constrained launch market and the technical challenges of operating in space.

Looking ahead, industry observers will be watching for updates on Starship’s progress, the regulatory approval process for Starcloud’s satellite constellation, and the company’s ability to secure additional launch contracts. The next two years will be critical in determining whether orbital AI data centers can transition from concept to reality.