SpaceX filed an FCC application on January 30, 2026 to deploy up to one million satellites designed to function as orbital AI data centers. The satellites would operate between 500 and 2,000 kilometers altitude, connected by optical laser links and integrated with the existing Starlink constellation. SpaceX already operates over 10,000 Starlink satellites—roughly two-thirds of all active satellites in low Earth orbit. The FCC accepted the filing for review on February 4.

1. This Is What AI Has to Do (SpaceX, xAI)

Earth's power grid can't keep up with AI compute demand. Space has unlimited solar power and no NIMBYs.

The math behind the proposal is about energy, not ego. Terrestrial data centers face rising power costs, cooling challenges, and exponential compute demand from AI models. SpaceX's pitch is that satellites harvesting near-continuous solar power in orbit can distribute AI workloads across a million small nodes without overloading any terrestrial grid. SpaceX completed its acquisition of Musk's AI venture xAI in February 2026, creating a vertically integrated space-cloud service that could run Grok models on orbital hardware connected by Starlink's laser mesh network.

Starlink already proved the model works at scale. The constellation serves over nine million paying customers and has demonstrated emergency connectivity in Ukraine and disaster zones. With Falcon 9 reusability driving launch costs down, space-based computing could become economically competitive with terrestrial data centers within years. The January 2026 FCC approval of 7,500 additional Gen2 satellites already committed the U.S. to expanding orbital infrastructure. A million more is a leap, but for SpaceX, it's the same trajectory, accelerated.

2. You'll Destroy The Night Sky (Meredith Rawls, American Astronomical Society, Dark Sky International)

One company's commercial ambition would permanently destroy the night sky for every person on earth.

A million satellites would mean more visible satellites than stars for large portions of the night. Researchers modeling the proposal's impact found that for large stretches of the year worldwide, observers would see more satellites than stars. Meredith Rawls, a research scientist at the University of Washington working on the Vera Rubin Observatory, called it existentially frustrating that commercial launches are interfering with our views of the cosmos.

The damage is already measurable. The ratio of twilight images affected by Starlink's existing constellation rose from 0.5 percent in late 2019 to 20 percent by late 2021. The Vera Rubin Observatory—a next-generation telescope designed to map the entire night sky repeatedly—already suffers satellite streak contamination across a significant share of its observations. A million more would render it partially unusable. The American Astronomical Society issued a formal action alert urging public comment on the FCC filing. Dark Sky International called the proposal a threat requiring immediate action. The International Astronomical Union established a Centre for the Protection of Dark and Quiet Sky in Paris specifically to address megaconstellation impacts. The fundamental argument: the night sky is a global commons, and no company should be able to commercialize it into oblivion.

3. One Company Shouldn't Own Inner Space (China, NASA Debris Office, Progressive Policy Institute, Saarland University)

Space debris, ozone damage, and a monopoly that no regulator on Earth can control.

One near-miss already triggered a formal UN complaint. China filed a protest in January 2026 after a Starlink satellite passed within 200 meters of a Chinese spacecraft with no prior coordination. SpaceX responded by announcing it would lower 4,400 satellites from 550 to 480 kilometers altitude. Low Earth orbit already contains an estimated 120 million debris fragments and over 43,000 tracked objects larger than 10 centimeters. Each collision generates thousands more fragments—the 2009 Iridium-Cosmos collision alone produced roughly 2,000 trackable pieces. The concern is Kessler syndrome: a cascading chain reaction that could make key orbital altitudes permanently unusable.

The environmental costs are real and largely unexamined. Researchers at Saarland University calculated that orbital data centers would produce emissions an order of magnitude higher than land-based data centers when accounting for rocket launches and reentry pollution, which damages the ozone layer and persists longer at high altitude. SpaceX's FCC filing omits any climate impact analysis. The Progressive Policy Institute warned that SpaceX already dominates U.S. launch capacity. Amazon's competing Kuiper project has deployed over 200 satellites across eight missions—a fraction of Starlink's 10,000-plus. At some point, dominance of orbital space becomes a sovereignty question, and the only country with a regulatory veto is the one where SpaceX is headquartered.

Where This Lands

SpaceX's case is that AI compute needs are outrunning Earth's infrastructure and that orbit is the logical next frontier. The astronomers' case is that a million satellites would permanently alter the night sky for every human being, destroying a scientific and cultural commons. The safety and governance camp argues that one company already controls two-thirds of active satellites, and a million more without international coordination or environmental review isn't innovation—it's a monopoly on space itself. Where this lands depends on whether the FCC treats this as a routine spectrum application or as what it actually is: a decision about who gets to decide what orbit looks like for the rest of the century.

Sources