Two Tech Giants Converge on Space Computing

According to The Wall Street Journal, Google is in active negotiations with SpaceX for rocket launch services to deploy data center hardware into Earth orbit . The talks are confirmed by a person familiar with the matter, and Google is also speaking with other rocket launch companies to maintain flexibility .[citation:7]

Google already has skin in the game. The company holds a 6.1 percent stake in SpaceX, and Google executive Don Harrison serves on SpaceX's board of directors .[citation:9] This existing relationship provides a foundation for deeper collaboration on what could become a transformative infrastructure category.

The orbital data center concept is ambitious: Instead of building more power-hungry, land-intensive server farms on Earth, Google would place computing racks aboard satellites powered by solar arrays. The primary drivers are the same factors constraining terrestrial AI expansion — power availability and land costs .[citation:3]

2027 Prototype Satellite Planned

Google launched "Project Suncatcher" in 2025, with plans to launch prototype satellites by 2027 .[citation:1] The company is collaborating with Planet Labs, the satellite imaging firm, to build the necessary hardware for orbital computing .[citation:1][citation:9]

The initial phase involves testing small-scale computing modules on satellites. If successful, Google would scale the architecture to larger orbital deployments — potentially creating a distributed computing network in space .[citation:9]

Scaling Launch Capacity for Thousands of Annual Flights

Musk announced on May 12 that SpaceX is evaluating "several locations domestically and internationally" to build "the world's most advanced space launch site" .[citation:2] The expansion is driven by Starship's full reusability — a system designed to launch thousands of times annually with minimal refurbishment .[citation:4]

According to SpaceX, the fully reusable Starship system is designed to carry more than 100 metric tons of cargo . Current launch capacity at Starbase, Texas, and Florida facilities is insufficient for the company's long-term plans, which include Mars colonization, satellite deployment expansion, and rapid point-to-point Earth transportation.[citation:4]

Solving AI's Energy and Land Constraints

The push for orbital data centers is driven by real constraints in terrestrial AI infrastructure .[citation:1][citation:3]

• Unlimited solar power: Space-based data centers would draw power directly from solar arrays, bypassing grid constraints entirely
• No land acquisition: Orbital real estate is abundant, with no zoning restrictions or environmental impact reviews
• Cooling advantages: Space offers passive cooling, though the vacuum of space presents its own thermal management challenges
• Low-latency potential: For certain applications, orbital proximity could reduce latency compared to distant terrestrial data centers

Industry observers note that orbital computing could fundamentally reshape the economics of AI infrastructure. The primary costs shift from energy and real estate to launch and satellite deployment — costs that SpaceX is uniquely positioned to minimize .[citation:3]

Not Quite Ready for Prime Time

For all the ambition, significant technical challenges remain unresolved :[citation:3][citation:9]

• Heat dissipation: In the vacuum of space, cooling electronics is counterintuitively difficult — there's no air to carry heat away. Radiators are required, but their effectiveness is limited.
• Maintenance and repair: Unlike terrestrial data centers where failed servers can be swapped in hours, orbital failures might require weeks or months to address — assuming on-orbit servicing exists at all.
• Latency for interactive AI: Geostationary orbit introduces significant signal delay. Low Earth orbit (LEO) reduces latency but still exceeds terrestrial fiber optic networks.
• Launch costs: Even with Starship's reusability, deploying thousands of computing nodes in space remains expensive.
• Radiation hardening: Space electronics require shielding against cosmic radiation, adding weight and cost.[citation:3]

Google CEO Pichai acknowledged the timeline: "We will put micro racks into space, attached to satellites, and test them. Then we will scale up." The process will take years, not months .[citation:9]

Why Both Announcements Matter for Investors

SpaceX's summer 2026 IPO is expected to be the largest in history, with a target valuation of approximately $1.75 trillion .[citation:4] Both the orbital data center talks and the spaceport expansion serve as investor narratives — signaling that SpaceX is not merely a launch provider but an infrastructure company for the AI era.

The orbital data center concept differentiates SpaceX from traditional launch competitors. If successful, the company would create a new market category where it controls both the transportation and the computing infrastructure in space.

Recent strategic moves reinforce this positioning:

• Acquisition of xAI (merged into SpaceXAI) valued at $250 billion
• Option to acquire AI coding startup Cursor for $60 billion later this year [citation:1]
• Major AI compute deal with Anthropic: 220,000+ NVIDIA GPUs, 300+ megawatts, with Anthropic expressing interest in orbital computing collaboration