The Orbital Data Center Race in 2026: Cowboy, Starcloud, SpaceX, Anthropic, and Why Capital Is Pouring Into AI Compute in Orbit

The orbital data center category went from speculative concept to one of the most heavily capitalized space-industry segments of 2026 in a span of weeks. Cowboy Space closed a $275 million Series B at a $2 billion valuation on May 11, 2026, just over a month after Starcloud crossed unicorn status with a $170 million Series A at a $1.1 billion valuation. Starcloud is reportedly pursuing at least $200 million more in a deal that would double that valuation to $2.2 billion, according to a source close to the situation cited in SpaceNews coverage. SpaceX has publicly articulated its own orbital data center ambitions and has reportedly discussed deployment of up to one million data center satellites over time. AI company Anthropic last week expressed interest in using SpaceX's proposed orbital data centers as part of a broader agreement for access to SpaceX terrestrial compute capacity. Blue Origin and others are also pursuing the emerging orbital compute market. Two-month-cumulative private capital flowing into the orbital data center category now exceeds $445 million in announced equity (Cowboy plus Starcloud's first round), with substantially more pending.

Why Capital Is Pouring In: The AI-Infrastructure Constraint

The investor thesis underlying the orbital data center category is, at its core, a structural argument about terrestrial AI infrastructure constraints. AI compute demand has scaled faster than terrestrial data center deployment can keep up, particularly because terrestrial data centers are constrained by power availability (utility-grid capacity, baseload generation, and grid interconnection timelines), water availability (cooling water in regions with strained water supply), and land permitting (community opposition, zoning, environmental review). Each of those constraints is compounding: power-grid interconnection queues in major U.S. data center regions extend multiple years; water-supply constraints in arid data center regions are pushing operators toward less-suitable geographies; and land permitting timelines are extending as community concerns intensify. Orbital data centers offer a structurally different constraint envelope: solar power is freely available in continuous sunlight in many orbital regimes; cooling is handled by radiating waste heat to deep space rather than competing for terrestrial water; and the deployment regime is regulated by spacecraft licensing rather than terrestrial land-use review. Whether the orbital architecture wins economically against terrestrial alternatives depends on launch cost trajectories and on the ability to industrialize compute hardware that survives the orbital environment — but the structural argument is compelling enough to mobilize substantial growth capital.

Cowboy: Vertically Integrated Rocket-Plus-Data-Center

Cowboy Space's $275 million Series B is the largest single financing of the orbital data center category to date. The architectural distinction is the integrated rocket-upper-stage-as-data-center design, which contrasts with the conventional satellite-on-launch-vehicle approach taken by competing developers. Cowboy is building a launch vehicle larger than Falcon 9 (70m) but smaller than Starship (>120m), targeting 20,000–25,000 kg of payload to LEO, with first proprietary launch carrying a 1-megawatt orbital data center planned for end of 2028. Earlier mission milestones include a 2026 power-beaming demonstration with manufacturer Apex and a 2027 deployment of the first Galactic Brain orbital data center node using NVIDIA Space-1 Vera Rubin modules. The $2 billion valuation places Cowboy at the high end of orbital-data-center-category private valuations alongside Starcloud's pursued $2.2 billion.

Starcloud: Y Combinator's Fastest Unicorn, Starship-Dependent

Starcloud is the principal direct competitor and the company that was first to unicorn status in the category. Y Combinator's fastest unicorn raised a $170 million Series A at a $1.1 billion valuation co-led by Benchmark and EQT Ventures with NFX, Nebular, Adjacent, 776 Ventures, Fuse Ventures, Manhattan West, and Monolith Power Systems. The company is reportedly pursuing at least $200 million more at a $2.2 billion valuation. Starcloud's plans target an 88,000-satellite orbital data center constellation, with a 3-ton Starcloud-3 spacecraft and a Starship deployment dependency. The architectural bet is that SpaceX's Starship economics — once industrialized at full operational cadence — will deliver dollars-per-kilogram-to-LEO low enough to make the conventional satellite-on-launch-vehicle approach economically dominant, even if the per-satellite mass overhead is higher than Cowboy's integrated architecture would deliver.

SpaceX: Internal Compute and the Anthropic Signal

SpaceX has its own orbital data center ambitions, reportedly discussed at deployment scale of up to one million data center satellites over time — a number that, if even approximately operationalized, would dwarf the Cowboy and Starcloud constellation plans by orders of magnitude. The internal SpaceX thesis appears to combine launch-economics dominance (Starship-class cadence and cost), in-house compute hardware development, and an internal customer base that includes xAI and Tesla compute requirements. The most concrete external commercial signal is Anthropic's publicly expressed interest in using SpaceX's proposed orbital data centers as part of a broader agreement for access to SpaceX terrestrial compute capacity — the first major AI-customer endorsement of the orbital data center thesis that includes a named tier-one AI lab. The Anthropic signal underscores that AI customer demand for orbital compute is not speculative but is being actively planned for by the AI labs that consume the largest fraction of compute industry-wide.

What the Capital Concentration Means for the Industry

The two-month accumulation of growth-stage capital into orbital data centers is structurally significant for the broader space industry. First, it validates orbital infrastructure as a credible AI-infrastructure category — institutional venture capital at growth scale (Index Ventures, IVP, Benchmark, EQT, Andreessen Horowitz, NEA, and many others) is collectively committing capital that exceeds the early-stage funding totals of many adjacent commercial space sub-sectors over multi-year periods. Second, it accelerates the launch-economics race — both Cowboy and Starcloud business cases depend on dramatic per-kilogram launch cost reductions, which puts substantial market-pull behind Starship industrialization (for the rideshare-architecture approach) and behind Cowboy's own clean-sheet vehicle development. Third, it pulls AI-customer commercial relationships into the orbital category — the Anthropic-SpaceX signal is the first major operational link, and additional AI-customer commitments are likely to follow as the technical demonstrations from 2026 through 2028 validate or recalibrate the architectural thesis. The orbital data center race in 2026 is not just a competition among orbital startups; it is a structural shift in how the AI compute industry and the launch industry are jointly evolving.

Frequently Asked Questions