Starcloud Raises $250M for Orbital Data Centers as Launch Options Tighten and Nvidia Joins

Starcloud Secures $250 Million to Advance Orbital Data Centers

Starcloud announced a $250 million Series A extension at a $2.3 billion valuation, bringing total capital raised to $450 million since its 2024 founding1

1

2

2

. Manhattan West led the round with participation from Nvidia, which contributed $25 million, and Cisco Investments1

1

3

3

. The funding will support manufacturing expansion at a new 100,000-square-foot facility in Woodinville, Washington, engineering collaboration with Nvidia, and procurement of future launch slots2

2

. CEO Philip Johnston emphasized that securing launch capacity has become one of the biggest costs as SpaceX plans to phase out Falcon 9 by 20281

1

.

Launch Capacity Constraints Drive Strategic Planning

Starcloud faces mounting pressure to secure launch capacity as the rocket transportation market tightens. Johnston told TechCrunch that "launch is pretty constrained right now because Falcon 9 is scheduled to end in 2028"1

1

. The company has filed an FCC application to operate 88,000 satellites1

1

2

2

, requiring enormous launch allocation. SpaceX is transitioning to Starship, which has not yet flown twice, creating uncertainty for satellite operators3

3

. Competing rockets like Blue Origin's New Glenn and ULA's Vulcan are not flying regularly, while Rocket Lab's Neutron has not reached the pad1

1

. Johnston acknowledged the challenge: "Obviously if we can't book any SpaceX launch capacity in 2029, that will be challenging for us"3

3

. The company is considering buying dedicated Falcon 9 launches and signing contracts with multiple providers to mitigate risk1

1

.

Nvidia Partnership Validates Space-Based AI Infrastructure

Nvidia's investment signals strong confidence in Starcloud's approach to orbital AI data centres. In November 2025, Starcloud became the first company to fly an H100 GPU to orbit aboard Starcloud-1, achieving roughly 100 times the GPU compute previously available in space2

2

3

3

. The company successfully trained the first AI model in space, a large language model called NanoGPT, and demonstrated inference using a version of Google's Gemini in orbit2

2

3

3

. Johnston emphasized that Nvidia "did way more technical duty on this than anybody else," basing their investment on flight data from Starcloud-11

1

3

3

. The companies are collaborating on the Nvidia Space-1 Vera Rubin Module, designed specifically for orbital conditions and expected to deliver 25 times the computing capacity in space compared to the H1002

2

4

4

. Starcloud's satellites will serve as early flight platforms for these space-rated chips, anticipated to launch in late 20281

1

.

Near-Term Deployment and Technical Challenges

Starcloud plans to launch two Starcloud-2 satellites with 8-kilowatt compute capacity each on rideshare flights in 20271

1

3

3

. These satellites will perform AI inference in orbit for customers including US government agencies1

1

5

5

. The company's next-generation Starcloud-3 spacecraft, designed for SpaceX's Starship rocket, will feature 200-kilowatt power capacity5

5

. Johnston stated on LinkedIn that the company intends to ramp production to 100 satellites per week3

3

. Engineers are addressing three critical design challenges: determining radiator size based on chip operating temperature, positioning radiation shielding, and ruggedizing hardware to survive launch1

1

3

3

. Unlike terrestrial data centers, orbital systems cannot use conventional air cooling and must transfer heat through radiators while protecting components against radiation4

4

.

Addressing Terrestrial AI Energy Bottleneck

Starcloud positions orbital compute power as a solution to terrestrial constraints limiting AI infrastructure expansion. The AI energy bottleneck stems from grid connection delays, land scarcity, and water consumption for cooling2

2

3

3

. In orbit, satellites access continuous sunlight and avoid grid dependencies5

5

. Solar panels can face the sun constantly without nighttime energy reduction or adverse weather interference5

5

. The company envisions a constellation providing 20 gigawatts of computing capacity in space through 88,000 satellites2

2

4

4

. This contrasts sharply with current deployment: Starcloud has booked two Starcloud-2 satellites delivering 16 kilowatts against the 20-gigawatt target3

3

. The company's 25 employees are scaling manufacturing capabilities to bridge this gap1

1

3

3

. Starcloud also enables faster data analysis by processing sensory measurements in orbit rather than transmitting raw datasets to ground stations5

5

.

Market Competition and Long-Term Viability Questions

Starcloud is not alone in pursuing data center satellite network opportunities. SpaceX has filed plans for up to one million data center satellites through its Starmind project2

2

. SpaceX is developing the AI1 satellite with a 230-foot wingspan and computing capacity similar to Starcloud-3, with mass production planned as early as late 20275

5

. TechCrunch reported in May that one rival concluded there were not enough rockets available and raised $275 million to build its own launch vehicles3

3

. SoftBank's Masayoshi Son dismissed space data centers in June as "a bet on the AI race rather than a business"3

3

. Industry observers have called the economics of orbital AI "brutal"3

3

. Reaching the 88,000-satellite constellation would require massive expansion of satellite manufacturing and launch capabilities far beyond current industry capacity4

4

. Starcloud has hedged by signing a deal in May to equip satellites with SpaceX Starlink mini laser terminals for data transmission3

3

. The company's valuation jumped from $1.1 billion in March to $2.3 billion, reflecting investor confidence despite execution risks3

3

.