Plans to put data centers in orbit, pitching space as the next frontier for cloud computing and AI, are running into a new kind of resistance: environmental and scientific groups demanding the U.S. government slow down.
A coalition led by Earthjustice is asking the Federal Communications Commission, the agency that licenses satellite systems and manages spectrum, to halt reviews of several proposed “orbital computing” constellations until the FCC completes a sweeping environmental analysis. Their warning is blunt: combined filings from SpaceX, Blue Origin, Starcloud, and Cowboy Space could add up to more than 1 million satellites in low Earth orbit, a scale they say makes the FCC’s current case-by-case approach obsolete.
The petition leans on the National Environmental Policy Act (NEPA), the bedrock federal law that requires agencies to study environmental impacts before taking major actions. The groups argue the FCC has been treating most satellite licenses as effectively exempt from detailed environmental review, an assumption they say no longer holds in the era of mega-constellations.
Table des matières
- 1 Earthjustice wants a program-wide environmental review before the FCC approves more filings
- 2 Why “orbital computing” is suddenly part of the AI infrastructure race
- 3 More than 1 million satellites: the number driving the political fight
- 4 Debris, ozone, rocket emissions, and light pollution: what the groups say is at stake
- 5 The FCC is already rethinking its environmental rules, and that could slow approvals
- 6 Key Takeaways
Earthjustice wants a program-wide environmental review before the FCC approves more filings
Earthjustice filed the petition on behalf of DarkSky International, Environment America, and Public Employees for Environmental Responsibility (PEER). Their request: the FCC should prepare a “Programmatic Environmental Impact Statement,” or program-wide EIS, before approving any of the orbital data-center constellations currently under review.
Unlike a project-specific study, a programmatic EIS is designed to evaluate cumulative impacts across multiple related actions, risks, alternatives, tradeoffs, and long-term effects. The petition argues that putting hundreds of thousands, or potentially a million, new objects into low Earth orbit isn’t just a technical licensing matter. It’s a major public-policy decision with environmental consequences that should be analyzed as a whole.
The groups also take aim at the FCC’s long-standing reliance on “categorical exclusions,” a regulatory shortcut that presumes certain actions don’t significantly affect the environment. That logic, they argue, breaks down when the proposal isn’t dozens of satellites but fleets so large they would require frequent launches, constant replacements, and routine deorbiting.
Why “orbital computing” is suddenly part of the AI infrastructure race
Companies pushing orbital data centers frame the idea as a workaround for Earthbound constraints: crowded power grids, land-use fights, and local opposition to new data centers. With AI driving demand for computing and storage, the pitch is that moving some infrastructure into space could relieve pressure on the ground.
But critics say the public has seen few concrete details about how these systems would reduce, or even measure, their environmental footprint. The petition argues the companies have not publicly laid out robust mitigation plans for launch emissions, pollution from atmospheric reentry, or the optical impacts of large satellite fleets on the night sky.
The groups want a process that forces comparisons among alternatives: smaller architectures, different orbital bands, stricter deorbit requirements, and limits on how densely satellites can be packed at certain altitudes.
More than 1 million satellites: the number driving the political fight
The petition names SpaceX and Blue Origin, two of the most prominent U.S. space companies, as well as Starcloud and Cowboy Space. Taken together, the filings cited by the groups could lead to more than 1 million satellites in low Earth orbit.
That figure is doing real work in this debate. It shifts the conversation from the FCC’s usual territory, spectrum allocations and orbital slots, into a broader fight over environmental planning and oversight. The groups argue that at this scale, satellite turnover becomes a permanent industrial cycle: manufacturing, launching, operating, replacing, and deorbiting on repeat.
They also point to other industry experiments, like floating data centers explored by companies including MOL and Hitachi, as evidence that space is not the only path forward, and that NEPA requires agencies to consider reasonable alternatives.
Debris, ozone, rocket emissions, and light pollution: what the groups say is at stake
The petition lays out a list of concerns that span the atmosphere, orbital safety, and scientific research. On the atmospheric side, the groups cite rocket emissions and pollutants created when satellites burn up during reentry. With massive constellations, they argue, end-of-life reentries could become a steady stream rather than occasional events, potentially increasing the amount of material deposited in the upper atmosphere.
They also raise the specter of ozone impacts, arguing that without a program-wide review, regulators may fail to capture how increased launch cadence and reentry byproducts interact over time.
On orbital safety, the groups warn that more objects in orbit mechanically increases collision risk, and the chance of debris-generating events that can cascade. Even if individual satellites meet deorbit rules, they argue, density itself changes the risk profile.
And for astronomers and dark-sky advocates, the concern is visibility. DarkSky International has long criticized satellite constellations for bright streaks across telescope images and broader light pollution. The petition argues that when multiple operators fill similar orbital bands, the combined effect, not any single license, determines what researchers and the public actually see in the night sky.
The FCC is already rethinking its environmental rules, and that could slow approvals
The petition lands as the FCC is separately reviewing its own environmental procedures, acknowledging how dramatically commercial space activity has grown over the past decade. The groups are urging the agency to use that moment to reset how it evaluates satellite licensing at scale, and to pause certain approvals until the new approach is clear.
For companies, the immediate risk is timing. A programmatic EIS can add months, or longer, through data collection, public comment, and formal responses. That uncertainty can ripple through manufacturing schedules, launch contracts, ground infrastructure planning, and investor expectations.
For regulators, a broader review could create a sturdier legal and scientific foundation for decisions that may shape near-Earth space for decades. For the industry, it could mean redesigns, dimmer satellites, stricter tracking and reporting, different orbits, or tighter deorbit standards, that raise costs but potentially reduce long-term harm.
The bigger question now is whether the FCC treats orbital data centers as just another licensing queue, or as the start of a new, planet-scale infrastructure that demands a new level of scrutiny.
Key Takeaways
- Earthjustice is asking the FCC to conduct a programmatic EIS before approving any orbital data centers.
- Filings from SpaceX, Blue Origin, Starcloud, and Cowboy Space would exceed one million satellites in low Earth orbit.
- NGOs cite rocket emissions, reentry pollutants, ozone impacts, orbital debris, and light pollution.
- The FCC is reevaluating its environmental review rules, raising the risk of regulatory delays for the industry.
- Critics denounce the lack of public environmental mitigation plans in current applications.
- Un million de satellites, FCC sous pression, centres de données en orbite, l’obstacle inattendu venu des ONG environnementales - 16 juillet 2026
- GPT-5.6, 12 juillet 2026, 5 outils concrets montrés à Sam Altman, du dressing IA à Pokémon Go pour chats, inattendu - 16 juillet 2026
- 16 pouces, RTX 5090 à 175 W, Ryzen 9 9955HX3D, écran OLED 2560 x 1600, ce portable surprend sur les compromis - 15 juillet 2026
