Starcloud Raises $250 Million to Take AI Data Centers Into Orbit

Starcloud has raised a $250 million extension to its Series A at a $2.3 billion post-money valuation, more than doubling its valuation in less than five months. The financing places a very large private-market bet on a still-unproven proposition: that energy-hungry artificial-intelligence computing can eventually be operated economically in orbit.

Manhattan West led the round. Nvidia and Cisco Investments joined as new strategic backers alongside Cedar Capital, Goanna Capital and Standard Capital, while existing investors including Benchmark, EQT, Soma, NFX and 776 participated again. Starcloud says the extension brings total capital raised since its 2024 founding to $450 million.

The company plans to use the money for manufacturing capacity, engineering work with Nvidia and future launch procurement. It is building a 100,000-square-foot facility in Woodinville, Washington, for its next-generation Starcloud-3 spacecraft. The scale of the capital and the investor list make the round material, but neither the valuation nor the technical milestones prove a commercial market at comparable scale.

From demonstration satellite to industrial ambition

Starcloud's technical credibility rests on Starcloud-1, launched in November 2025 with an Nvidia H100 graphics processor. The company says it subsequently trained an AI model in orbit, ran a version of Google's Gemini and demonstrated inference and fine-tuning on flight hardware. Those achievements are company-reported, although the successful deployment established that a data-centre-class chip could operate in space.

The next step is more demanding. Starcloud and Nvidia are developing the Space-1 Vera Rubin Module, a system adapted for radiation exposure and the unusual thermal environment of orbit. On Earth, data centres move heat through air and water systems. In space, heat must ultimately be radiated away, requiring large surfaces and careful power management. Components also face radiation, launch vibration and limited opportunities for physical repair.

Starcloud's long-range plan calls for as many as 88,000 satellites and 20 gigawatts of orbital computing capacity. Those figures describe an ambition, not deployed infrastructure or contracted revenue. They would imply a manufacturing and launch programme on a scale far beyond today's demonstration missions.

The investment thesis is really about energy and launch costs

The appeal of orbital computing begins with abundant solar energy and the difficulty of building large terrestrial data centres. AI infrastructure increasingly competes for grid connections, water, land and political consent. A satellite in the right orbit can receive long periods of sunlight without drawing on a local electricity system.

That advantage comes with a formidable transport bill. Moving hardware into orbit remains expensive, and compute equipment depreciates quickly as new chips arrive. Earlier this year, Starcloud's chief executive said orbital systems could become competitive with terrestrial data centres if launch prices fall toward $500 per kilogram, a threshold he associated with high-cadence heavy-lift rockets around 2028 or 2029. That is a forecast dependent on launch providers delivering capacity, reliability and price reductions that do not yet exist at industrial scale.

Latency and data movement are additional constraints. Some workloads can run where data are generated or where immediate terrestrial response is unnecessary. Other applications require rapid, high-volume links to users and storage systems on Earth. The economic case therefore depends not only on power and cooling but on optical communications, networking and the mix of workloads customers are willing to place off-planet.

Cisco's participation is relevant in that context. The company brings networking and optical infrastructure experience, while Nvidia provides both capital and the core compute platform. Strategic investment can deepen technical collaboration, but it does not by itself guarantee that either company will become a large customer.

A valuation built on execution milestones

Starcloud was valued at $1.1 billion when it raised $170 million in March. The new $2.3 billion valuation reflects both investor appetite for AI infrastructure and the scarcity value assigned to companies with hardware already operating in orbit. It also raises the bar for future execution.

At this stage, the business must translate demonstrations into repeatable manufacturing, launch access and customer contracts. Each stage introduces different risks: spacecraft production can run late, launch schedules can slip, radiation can shorten hardware life, and cheaper terrestrial energy or more efficient chips can change the comparative economics.

The funding gives Starcloud more room to absorb those risks. It can reserve launch capacity, build production systems and gather flight data before public-market pressures demand predictable margins. The round also signals that investors increasingly view space infrastructure as part of the AI capital-expenditure cycle rather than as a separate aerospace niche.

Why it matters

The significance of the round is not that space data centres have become commercially proven. They have not. It is that leading infrastructure companies and private investors are financing the attempt at a scale large enough to move from laboratory demonstration toward industrial production.

For terrestrial data-centre operators, the project highlights how severe the power and permitting bottleneck has become. For chip and networking suppliers, it creates a new environment in which hardware must be redesigned rather than merely installed. For investors, the valuation concentrates several uncertain assumptions—launch-cost compression, reliable thermal control, customer demand and mass manufacturing—into one company.

Starcloud now has capital to test those assumptions. The next evidence should come from flight performance, production cadence and binding customer economics, not from a larger constellation target.

Sources: Starcloud's funding announcement and Reuters' independent funding report, with technical and cost context from TechCrunch's March reporting.