Key Points

• SpaceX filed with the FCC on January 30, 2026, to deploy up to one million solar-powered satellite data centers in low Earth orbit — 150 times larger than its existing Starlink network — describing the project as "the most efficient way to meet the accelerating demand for AI computing power"
• The in-orbit data center market was valued at $500 million in 2025 and is projected to reach $39.09 billion by 2035, with Google (Project Suncatcher), Starcloud, Lonestar Data Holdings, and China all pursuing active programmes
• The US Government Accountability Office published a formal assessment on April 28, 2026, confirming space-based data centers could reduce land, electricity and water demands — but flagging that heat dissipation in space remains a major unsolved engineering challenge

On January 30, 2026, SpaceX filed an application with the Federal Communications Commission that stopped the data center industry in its tracks. The company requested authorisation to deploy up to one million solar-powered satellite data centers in low Earth orbit — a constellation 150 times larger than its existing Starlink network — describing the project in blunt terms as "the most efficient way to meet the accelerating demand for AI computing power."

Five days later, the FCC's Space Bureau formally accepted the filing for review. The most ambitious data center project in human history had entered the regulatory process.

The answer to both questions is the same: electricity.

As explored in our guide on data centers and energy, global data center electricity consumption reached 565 terawatt-hours in 2026, up 27% from 2025. The four largest hyperscalers — Amazon, Google, Microsoft, and Meta — are spending a combined $725 billion on AI infrastructure this year. In growing numbers of markets, the constraint is no longer chips or capital. It is power. Grid operators in Ireland, Singapore, the Netherlands, and parts of the United States have imposed restrictions on new data center connections.

Orbital data centers offer a theoretical solution. Satellites in sun-synchronous orbits receive near-continuous solar energy, unaffected by weather, nighttime cycles, or atmospheric filtering. According to space infrastructure analysts, orbital systems can generate up to 40 times more solar energy than equivalent ground-based installations. They require no land permits, no water cooling, and generate no local emissions. Excess heat radiates directly into space.

The US Government Accountability Office took the concept seriously enough to publish a formal Science and Technology Spotlight on April 28, 2026, concluding that space-based data centers "could reduce the land, electricity, and water needed for data centers on Earth" — while cautioning that "engineering and economic barriers to deployment" remain significant.

No proposal has captured the scale of the opportunity — or the ambition — more dramatically than SpaceX's FCC filing.

The January 30, 2026 application describes an Orbital Data Center System of up to one million satellites operating between 500 and 2,000 kilometres altitude in sun-synchronous orbits. The satellites are designed as compute nodes, not communications relays — their primary function is running AI workloads, not routing internet traffic. The constellation would connect to Starlink via optical intersatellite links capable of one terabit per second throughput.

SpaceX stated in the filing that the system would achieve "transformative cost and energy efficiency while significantly reducing the environmental impact associated with terrestrial data centers." At projected scale, launching one million tonnes of satellites annually would generate 100 gigawatts of AI compute capacity — equivalent to approximately 20% of current US electrical consumption dedicated entirely to artificial intelligence.

The filing arrived five days before SpaceX's acquisition of xAI on February 2, 2026, creating a combined entity valued at $1.25 trillion. The merger integrates three critical capabilities under one corporate structure: SpaceX's launch infrastructure, Starlink's laser-mesh satellite network, and xAI's Grok AI models. On June 8, 2026 — timed deliberately to the week of SpaceX's anticipated IPO — the company unveiled AI1, the first physical satellite prototype in its orbital data center constellation.

Tim Farrar, President of TMF Associates, characterised the original FCC filing as "quite rushed" and likely a narrative tool for SpaceX's upcoming IPO, noting that the proposal has no guarantee of being approved. The filing is a spectrum and orbital-slot reservation, not a launch commitment.

SpaceX is not operating in isolation. A parallel commercial ecosystem is forming rapidly.

In November 2025, Google unveiled Project Suncatcher, a research initiative for scalable space-based AI infrastructure developed in partnership with Planet Labs. The project is equipping satellite constellations with tensor processing units — the same chips that power Google's AI training workloads on Earth — connected via optical links. Prototype launches are planned for early 2027.

Starcloud, a startup backed by Nvidia's Inception programme and Y Combinator, achieved a pivotal milestone in September 2024 when it successfully trained an AI model in orbit using an Nvidia H100 processor. In February 2026, Starcloud filed with the FCC for a constellation of 88,000 satellites focused on distributed AI inference, with each satellite handling workloads in parallel.

In April 2026, Lonestar Data Holdings announced StarVault, which it describes as "the world's first commercially operational space-based sovereign data storage platform." The service is scheduled to launch in October 2026 aboard Sidus Space's LizzieSat-4 mission. StarVault is not a full data center — it is data storage with advanced cryptographic key escrow capabilities — but it represents the first commercial space data service that enterprises can actually purchase. Lonestar says demand from governments, financial institutions, and critical infrastructure operators has already exceeded expectations.

Axiom Space deployed initial orbital data center nodes in low Earth orbit in January 2026 for cloud and AI workloads, partnering with Kepler Communications for optical relay infrastructure.

The in-orbit data center market was valued at approximately $500 million in 2025. It is projected to reach $39.09 billion by 2035, according to industry analysts, fueled by the convergence of AI compute demand and rapidly falling launch costs driven by SpaceX's reusable Starship rocket.

The race to build orbital data centers is not confined to American companies.

China's National Space Administration deployed three experimental AI satellites in 2024 equipped with domestic AI accelerators. Chinese state media confirmed plans for a 50-satellite constellation by 2028, focusing on Earth observation, autonomous systems, and AI model training — with substantial government funding as part of China's broader space ambitions.

A separate Chinese technology firm announced plans for a constellation of 2,400 inference satellites and 400 training satellites, representing one of the most detailed sovereign AI compute plans disclosed by any nation.

Europe is pursuing its own path. The ASCEND programme, backed by the European Space Agency and the Technical University of Munich, has planned a demonstration mission in 2026 to deploy a small-scale orbital data center module validating European technologies. The programme reflects the strategic importance governments place on not depending entirely on American orbital infrastructure for future AI compute needs.

The GAO's April 2026 report identified the central engineering challenge directly: "Data centers generate excess heat, but space does not cool computing hardware efficiently."

On Earth, data centers use air cooling, water cooling, or liquid immersion cooling to remove heat from servers. In space, the only mechanism available is thermal radiation — emitting heat as infrared energy into the void. This works, but it requires large radiator panels that add weight, cost, and complexity to every satellite. The more compute power packed into a satellite, the larger the radiators required.

A second challenge is latency. Orbital data centers in low Earth orbit at 500 kilometres altitude introduce a round-trip signal delay of approximately 10–15 milliseconds. For AI training workloads that can be batched and processed asynchronously, this is manageable. For real-time applications requiring instant responses, it is not.

A third challenge is servicing. Terrestrial data centers can swap failed hardware within hours. A satellite with a failed GPU cannot be repaired. The entire compute node must be replaced — meaning deorbited and relaunched — at significant cost.

These are engineering problems, not fundamental physical barriers. But they explain why the GAO's assessment was cautious: the concept is technically sound, the market opportunity is real, and the barriers are substantial.

Three developments will determine whether orbital data centers become infrastructure reality or remain a compelling vision.

First, SpaceX's Starship rocket must achieve reliable operational launch cadence. The orbital data center thesis depends entirely on dramatically lower launch costs — the same physics that makes orbital solar power attractive makes it expensive if each kilogram to orbit costs thousands of dollars. Starship targets below $100 per kilogram at scale. Nothing at this scale has been demonstrated.

Second, the FCC and international regulators must navigate the spectrum and orbital debris implications of constellations numbering in the hundreds of thousands or millions of satellites. There are approximately 14,000 active satellites in orbit today. One million more would represent a fundamental change to the near-Earth environment.

Third, the economics must close. The in-orbit data center market is currently a $500 million industry. For it to reach the projected $39 billion by 2035, orbital compute must achieve cost parity with — or undercut — terrestrial alternatives on workloads where latency is not critical. The investors backing Starcloud, Lonestar, Google's Project Suncatcher, and SpaceX itself are betting it will. The GAO is reserving judgment.

What is not in doubt is the direction of travel. The terrestrial grid constraints driving this investment are real, documented, and worsening. If those constraints persist — and every current projection suggests they will — the economic case for moving AI compute into orbit strengthens with every passing quarter.

Sources: SpaceX FCC Filing / GAO / Google / Starcloud and other international news outlets.

Share this article

Disclaimer: This article was produced with AI assistance based on publicly available news sources. While we strive for accuracy, NewsOracle makes no warranty as to the completeness or accuracy of the information. Errors and omissions may occur. Readers should independently verify all information before acting on it. NewsOracle does not intend to defame any individual or organisation and accepts no liability for any loss or damage arising from reliance on this content. Content is for informational purposes only and does not constitute legal, financial, medical, or professional advice. All rights reserved. Unauthorised reproduction prohibited.

More from NewsOracle

politics

Brent Crude Rises 1% as Iran Demands US Concessions for Strait of Hormuz

sports

Rahm Wins LIV Golf Season Crown Third Straight Year After Niemann Wins New York

finance

Brent Crude Rises 1.09% to $84.46 Amid Iran-U.S. Strait Deal Uncertainty