Starcloud Adds $250M to Series A as Orbital Compute Meets Launch Reality

Starcloud has added a $250 million extension to the $170 million Series A it announced in March, bringing the combined Series A to $420 million and total capital raised to approximately $450 million.

Manhattan West led the extension at a reported $2.3 billion valuation. NVIDIA, Cisco Investments, Cedar Capital, Goanna Capital, and Standard Capital joined the financing, alongside returning investors Benchmark, EQT, Soma Capital, NFX, and 776.

The extension is not a separate new Series A. It is additional capital attached to an already large financing as Starcloud moves from a first demonstration toward manufacturing, launch procurement, and a much more difficult version of the orbital-compute thesis.

What Starcloud has already placed in orbit

Starcloud launched its Starcloud One spacecraft in November 2025 carrying an NVIDIA H100 GPU. The company says the mission ran Google's Gemma model and trained OpenAI's NanoGPT in orbit.

The demonstration established that modern AI hardware and workloads could operate aboard the spacecraft. It did not establish that orbital data centers can compete economically with terrestrial infrastructure.

That next question is larger. A commercial system must survive radiation, launch vibration, thermal management, communications latency, repair constraints, and the cost and availability of reaching orbit. Hardware that becomes more valuable every quarter on Earth can become obsolete before a satellite architecture finishes moving through manufacturing and launch.

Capital for manufacturing and launch access

Starcloud plans to use the extension to expand manufacturing, support a 100,000-square-foot facility in Woodinville, Washington, advance its Starcloud-3 spacecraft, and secure launch capacity.

The launch market has become a strategic constraint. Satellite operators can design around compute density and power, but they still need reliable access to a rocket, an integration schedule, and an orbital slot. The additional capital gives Starcloud more capacity to reserve that supply before its larger systems are ready to fly.

The company also plans to fly NVIDIA's Vera Rubin Space-1 platform in late 2028. That target remains a future mission, not hardware already deployed.

Starcloud has filed with the Federal Communications Commission for a constellation that could include up to 88,000 spacecraft. The filing is a request, not regulatory approval or a deployed plan.

The economics behind the ambition

The company's argument is that abundant solar power and radiative cooling can eventually support large-scale computing in space. If the model works, orbital infrastructure could relieve some pressure on terrestrial power grids, water use, land, and permitting.

The capital intensity arrives before that economic case is proven. Starcloud must show that the gains in energy availability can outrun launch costs, replacement cycles, communications requirements, and the difficulty of maintaining hardware that cannot be serviced like a conventional data center.

The $250 million extension buys manufacturing capacity and access to the next experiments. It also makes the eventual comparison less theoretical. The first spacecraft showed that compute can run in orbit. The next ones have to show why enough of it should stay there.