TL;DR: Redmond-based orbital compute startup Starcloud has raised $170 million in Series B funding at a $1.1 billion valuation to build space-based AI data centers. Backed by Benchmark and EQT Ventures, the company plans an 88,000-satellite constellation to bypass terrestrial power grid limitations. This capital injection highlights growing investor confidence in orbital infrastructure as an alternative to land-based facilities.

The race to build AI data centers in orbit reached a new milestone on March 30 when Redmond, Washington-based Starcloud reached a $1.1 billion valuation. As global data center power consumption threatens to outstrip terrestrial grids by 2026, tech companies are exploring space to access continuous solar energy and natural vacuum cooling. Starcloud's $170 million Series B round, led by Benchmark and EQT Ventures, establishes orbital infrastructure as a major sector for venture investment. See our Full Guide to understand the technical and financial forces driving this orbital expansion.

Why are companies building AI data centers in space?

Space-based data centers solve the dual terrestrial bottlenecks of electrical grid capacity and land availability by utilizing near-continuous solar energy in orbit. Power limits are real. On Earth, AI training models consume vast amounts of electricity, competing with municipal grids and driving up utility costs. In contrast, satellites in low Earth orbit receive constant solar exposure, providing a reliable power source that does not draw from civilian energy networks.

Starcloud is addressing this opportunity with a planned constellation of 88,000 data center satellites. By offloading complex machine learning workloads to orbit, enterprise users can run resource-heavy calculations without carbon footprint penalties. This infrastructure is particularly valuable for processing data generated directly in space, such as high-resolution planetary imaging.

Solving the Grid Capacity Crisis

Hyperscale data centers on Earth face severe power constraints as utility companies struggle to upgrade grid infrastructure. By relocating computational workloads to orbit, companies bypass terrestrial regulatory delays and grid saturation. Orbital data centers utilize direct solar radiation, eliminating the intermediate transmission losses and grid fees associated with traditional power delivery systems. This setup provides an independent power source dedicated entirely to AI model generation and batch inference.

Technical Specifications of Starcloud's Constellation

To prove the feasibility of orbital computing, Starcloud launched a satellite in November carrying Nvidia's H100 graphics processing unit (GPU). This mission successfully ran AI training and inference tasks in orbit, demonstrating that high-performance hardware can operate under space conditions. The company's next step is an October launch carrying Amazon Web Services' (AWS) Outposts hardware to integrate orbital servers directly with terrestrial cloud services. These tests validate the thermal management systems required to cool GPUs in a vacuum.

How does Starcloud compete with SpaceX and Blue Origin in the orbital compute market?

Starcloud competes by securing early commercial partnerships with Nvidia, Google Cloud, and Amazon Web Services, while its rivals build proprietary infrastructure. While SpaceX and Blue Origin focus on vertically integrated ecosystems, Starcloud operates as an open-access infrastructure provider for existing cloud platforms. This positioning allows developers to deploy orbital workloads using familiar enterprise tools.

In February, Elon Musk's SpaceX acquired xAI and announced plans for a network of one million orbital data center satellites. Jeff Bezos' Blue Origin is pursuing similar space-based compute infrastructure. Starcloud bypasses this proprietary rivalry by acting as a neutral hardware partner, allowing AWS and Google Cloud to offer space-based computing to their existing enterprise clients. This approach avoids direct competition with established cloud providers, turning potential rivals into distribution channels.

Proprietary Ecosystems vs Open Cloud Partners

SpaceX aims to capture the entire value chain by combining its Falcon 9 launch vehicles with xAI software and proprietary satellite networks. Starcloud, conversely, relies on third-party launch contracts and focuses entirely on the compute payload and thermal management. This strategy attracts hyperscalers who want to avoid lock-in with a single launch provider or competitor ecosystem. Open models invite collaboration. By separating the launch service from the computing platform, Starcloud offers greater architectural flexibility to its enterprise partners.

Near-Term Commercial Contracts

Starcloud is generating immediate revenue through dedicated payloads for Earth observation and Deep Orbit Weather (DOW) satellites. CEO Philip Johnston confirmed that the company is negotiating binding energy offtake agreements with major hyperscalers, expected to launch in mid-2026. These agreements allow cloud providers to pre-purchase computational capacity before the full constellation is active. This pre-construction commitment secures the long-term cash flow required to fund ongoing satellite manufacturing and deploy consecutive orbital fleets.

High launch costs are the primary barrier to space-based AI commercialization

High launch costs are the main obstacle preventing space-based data centers from achieving economic parity with terrestrial facilities. Transporting heavy computing hardware, cooling systems, and power generation equipment into orbit requires substantial capital before generating any operational revenue. To compete with terrestrial data centers, the cost per kilogram of orbital payloads must decrease.

Starcloud expects launch costs to drop sufficiently by 2028 or 2029 to make orbital computing cost-competitive with land-based facilities. This projection depends on the deployment of next-generation, fully reusable heavy-lift launch vehicles from providers like SpaceX and Blue Origin. As launch capacity increases, the cost of deploying Starcloud's planned 88,000 satellites will drop proportionally. The company must sustain its development pace while waiting for launch market dynamics to mature.

The Timeline for Launch Cost Parity

The economic model for orbital data centers relies on the commercial maturity of heavy-lift rockets. Industry estimates suggest that launch costs must fall below $100 per kilogram to make orbital AI training cheaper than terrestrial grid power. Starcloud's road map targets this cost threshold in 2028, coinciding with the scaled production of its proprietary satellite hardware. Achieving this target will allow the company to offer competitive per-megawatt pricing to enterprise cloud customers.

Venture Capital Backing and Financial Runway

The latest $170 million Series B round brings Starcloud's total funding to $200 million. Prior rounds raised $34 million from investors including Andreessen Horowitz (a16z) and In-Q-Tel, the venture capital arm of the Central Intelligence Agency. This diverse investor base provides the financial runway needed to fund hardware research, manufacturing expansion, and launch contracts through 2026. Securing sovereign-linked capital from In-Q-Tel also highlights the potential national security implications of secure orbital processing nodes.

Key Takeaways

  • Orbital compute addresses power grid limits: Space-based data centers utilize near-continuous solar radiation, bypassing the energy and land constraints currently stalling terrestrial hyperscale developments.
  • Partnership strategies differentiate market entry: Starcloud’s open-access model secures early cloud integration with Nvidia, AWS, and Google Cloud, contrasting with the proprietary, vertically integrated systems proposed by SpaceX.
  • Launch cost reduction remains the critical metric: Commercial viability of orbital data centers depends on reusable heavy-lift launch vehicles lowering transport costs to competitive levels by 2028 or 2029.