Orbital Data Centers

Orbital data centers are computing systems placed in Earth orbit, from a processor that filters its own satellite's sensor data to a proposed optical network of server spacecraft. The gap between proposals and flown hardware is wide: filings before the US Federal Communications Commission (FCC) propose systems of up to 1,000,000 satellites,^[26] while the flown hardware includes a single NVIDIA H100 GPU, launched on November 2, 2025,^[8][9] and a Google test satellite carrying TPUs, launched on October 1, 2026.^[47][48]

A useful definition depends on architecture and workload, not on whether a company calls a payload a data center. Peer-reviewed literature distinguishes orbital edge data centers, which process data where it is collected, from orbital cloud data centers intended to provide shared compute and storage across a network.^[1] The distinction matters because the evidence is uneven. Onboard inference, high-performance computing, commercial servers, and limited model training have all been demonstrated in orbit. Dedicated commercial compute payloads have also launched. As of November 1, 2026, however, no megawatt-class orbital computing facility, tightly coupled AI-training cluster, or general-purpose cloud service for terrestrial users had been demonstrated. The peer-reviewed literature still treats systems at those scales as frameworks to evaluate, while the dated evidence below consists of studies, filings, and development programs.^[1]

Status as of November 2026#

Orbital computing remained at experiment scale in early November 2026. Google's first Project Suncatcher prototype, built with Planet, launched on SpaceX's Transporter-18 rideshare on October 1.^[47][48] On October 30, Google said the satellite's onboard Gemma 4 model had answered its first prompt, which Google called the first AI response from a data-center-class TPU in orbit; thermal and radiation tests are next, before a two-satellite laser-link test in 2027.^[49][81] On October 14, a LandSpace Zhuque-2E launched eight computing satellites that Chinese state media called the second batch of ADA Space's "Xing Suan" project.^[50][51] NVIDIA said on August 24 that the first-generation Starmind satellite would use an optimized Vera Rubin NVL72 system, which Elon Musk said was designed for launch in the fourth quarter of 2027.^[52][53] Starcloud, which raised $250 million in August,^[54] has vibration-tested Starcloud-2 and plans to launch it in January 2027, chief executive Philip Johnston said in an interview published October 31.^[86] The FCC sent information requests on August 28 to SpaceX and four other orbital-data-center applicants; none of the five had been granted by November 1.^{[36][55][56][57][58][59]}

Definitions and taxonomy#

Peer-reviewed work sorts orbital computing into five categories, each with a defining function and a boundary that separates it from the next larger claim.^[1]

Conventional spacecraft computers are not automatically orbital data centers. Navigation computers, instrument controllers, and processors that execute fixed mission software are ordinary spacecraft avionics. The term becomes useful when the payload provides substantial data reduction, reprogrammable acceleration, storage, or shared computing as a mission function. Lunar surface archives are another adjacent idea, but they are off-world storage rather than orbital data centers.^[1]

The clearest near-term category is edge computing. ESA's PhiSat-1, launched on September 3, 2020, used an Intel Movidius Myriad 2 processor to classify cloudy hyperspectral images and avoid downlinking unusable data.^[2] That was AI inference on data collected by the same spacecraft, not AI training and not a cloud service.

What has been demonstrated#

The HPE Spaceborne Computer program established that recent commercial hardware can perform useful high-performance computing in a protected orbital environment. Spaceborne Computer-1 reached the International Space Station in 2017 as an unmodified commercial system. NASA reports that software monitored abnormal behavior and adjusted performance, and that the system completed more than 18 months of operation before returning to Earth. Because the computer was hosted on the station, the trial did not validate a standalone spacecraft's power, cooling, communications, or servicing architecture.^[3]

Spaceborne Computer-2 moved from endurance testing toward applications. An upgraded third configuration arrived in 2024 with more than 130 TB of flash storage. HPE reported one 2022 life-sciences workload in which 2.8 GB of source data, normally requiring about 18 hours to transmit, was processed into a 92 KB result that took about two seconds to downlink.^[4] The example shows why reducing data before transmission can be valuable; it does not establish the economics of a standalone satellite.

The earliest well-documented onboard machine-learning training occurred before the commercial orbital-data-center wave. In autumn 2022, University of Oxford researchers uploaded code to D-Orbit's ION SCV004 spacecraft and trained a small cloud-detection classifier in orbit. An encoder had been trained on the ground, while the classifier was trained onboard from compressed image representations.^[5] The associated 2023 paper is a preprint, not a peer-reviewed journal article; its authors describe the result as the first few-shot training of a machine-learning model onboard a satellite to their knowledge.^[6] This makes the later claim that Starcloud-1 trained the first AI model in space too broad.

Axiom Space describes several progressively more independent demonstrations. An AWS Snowcone device operated on the station during Ax-1 in 2022, and its AxDCU-1 prototype was deployed there in fall 2025. Axiom says two dedicated data-center nodes launched on January 11, 2026, together with Kepler Communications' first tranche of optical relay satellites. Its public description does not identify the nodes as independent spacecraft and does not provide enough telemetry to establish uptime, customer use, or delivered computing performance independently.^[7] On October 6, 2026, Axiom said it was operating two units of its Axiom Resilient Compute (ARC) platform, ARC 1 and ARC 2, which had launched earlier that year on Kepler's optical relay network, and that it had confirmed links between the two units and with the ground. The update gave no performance figures beyond the 2.5 Gbps rate of the optical links, and the post-quantum encrypted workload transfer it described had been demonstrated only on the ground.^[68] Axiom has also announced a station-hosted node, AxODC Node ISS, developed with Spacebilt for the International Space Station in 2027.^[69]

Lonestar Data Holdings has tested data storage off Earth rather than orbital computing. Its Freedom payload, an 8 TB solid-state storage unit of about 1 kg carrying data for the Florida state government and the Isle of Man, rode Intuitive Machines' IM-2 lander to the Moon.^[74] The lander touched down in a crater near the lunar south pole on March 6, 2025 but came to rest on its side, and Intuitive Machines declared the mission concluded the next day.^[75] Lonestar's first StarVault orbital storage payload is to fly on Sidus Space's LizzieSat-4 satellite,^[76] which Sidus said on September 28, 2026 would launch on SpaceX's Bandwagon-6 rideshare no earlier than April 2027.^[77]

Starcloud launched Starcloud-1 on November 2, 2025. The company identifies it as the first spacecraft to carry an NVIDIA H100 GPU, a narrow and verifiable hardware first.^[8][9] Starcloud reports that it later ran the Gemma language model and trained NanoGPT onboard.^[8][85] That was an early language-model demonstration, not the first training of any machine-learning model in orbit. In an interview published October 31, 2026, Starcloud's chief executive, Philip Johnston, said the satellite's processors were now running trial inference jobs on synthetic-aperture-radar and hyperspectral images for government customers, and that the company would announce results later.^[86]

In China, 12 computing satellites developed by the startup ADA Space and Zhejiang Lab launched on May 14, 2025 as the first batch of two overlapping constellations: the Three-Body Computing Constellation, led by Zhejiang Lab, and Star-Compute, ADA Space's own program.^[61][62] Zhejiang Lab plans about 1,000 satellites by 2032, and its director, Wang Jian, has said the finished constellation would have a total computing power of 1,000 peta operations per second (POPS). ADA Space plans 2,800 satellites for Star-Compute.^[62][70][71] ADA Space later reported holding a continuous laser link between two of those satellites for eight days at separations of up to 1,000 km.^[62] Eight more ADA Space computing satellites launched on October 14, 2026, on a Zhuque-2E carrying ten satellites in all. CGTN, China's state-run international broadcaster, described them as the second batch of ADA Space's "Xing Suan" (Star-Compute) project, did not mention Zhejiang Lab, and said they would test liquid cooling at the 10-kilowatt scale, large deployable radiators, a laser communication network, and in-orbit training of foundation models.^[50][51] Orbital Chenguang's Chenguang-1, the first of a planned series that is meant to test the technology needed for orbital data centers at gigawatt scale, reached orbit on a Lijian-1 rideshare from the Jiuquan Satellite Launch Center on July 24, 2026, Beijing time.^[62] The same launch carried a demonstrator on which Zhejiang Lab planned to validate new GPU computing technology,^[62] and Xinhua reported on July 28 that two new satellites had joined the Three-Body constellation, one of them the Yunjian Muxi computing satellite.^[72] According to Cailian Press, a Shanghai financial news service, Yunjian Muxi was developed by the chipmaker MetaX with Zhejiang Lab and other partners, reached orbit on July 24, and carries an intelligent-computing payload built on MetaX GPU chips to test a Chinese-made GPU's reliability in orbit.^[73] The performance figures reported for these satellites come from the operators and Chinese state media.

Google's first Project Suncatcher prototype, built with Planet and launched on October 1, 2026, answered its first prompt from the ground at the end of that month.^[48][81] Over its first weeks in orbit the team worked through a commissioning checklist that included switching on the host and TPU trays and running early machine-learning tests. In the last week of October, during a ground-station pass, engineers sent up a prompt asking the satellite's Gemma 4 open model for a haiku about being in space, and the poem came back down as the satellite passed by Svalbard, Norway. Google, announcing the result on October 30, called it the first AI response from a data-center-class TPU in orbit.^[81] Starcloud had reported in December 2025 that Starcloud-1 was running Google's Gemma model on its H100 GPU.^[8][85] Follow-on prompts asked for space-themed limericks and for telemetry reports on the status of the payload systems, and Google said it would next run a series of tests on how the TPUs behave under thermal stress and radiation exposure in orbit.^[81] Ars Technica reported before launch that the satellite carries four TPUs and that its cooling system would limit them to runs of about 15 minutes;^[82] after the first prompts, TechCrunch reported that the TPUs' power demands allow only one of them to run at a time, in stretches of about 20 minutes.^[83]

Dated program matrix#

The program matrix below separates flight evidence from plans across orbital computing efforts. "Launched" does not mean that a commercial service or its advertised performance has been independently verified.

ESA's 2022 to 2024 generic study simulated an architecture rather than building one. It found a workload-dependent tipping point at which processing data onboard could be preferable to downloading raw data.^[10] ASCEND studied a much larger European cloud concept. Its official project report defined a 10 MW minimum viable product and a 1 GW objective for 2050, while calling for further work on hardware compatibility, architecture, business planning, launch capacity, and orbital assembly.^[11] Thales Alenia Space also said ASCEND's environmental case required a future launcher about ten times less emissive over its lifecycle than the baseline used in the study.^[12] These were conditional study outputs, not proof that the proposed system would be carbon-neutral or economical.

Google announced Project Suncatcher as a research moonshot, not a product. Its June 2026 revised preprint models one illustrative cluster of 81 satellites at a mean altitude of 650 km, within a 1 km radius and with neighbor separations varying roughly from 100 to 200 m. The same paper reports ground radiation tests and a short-path optical bench test, while identifying thermal control, high-bandwidth ground communications, reliability, and repair as unfinished work.^[14] A peer-reviewed version of the paper was published in Joule on October 1, 2026.^[63] In a preprint of a paper presented at the 2026 AAS/AIAA Astrodynamics Specialist Conference, two Massachusetts Institute of Technology researchers kept the 100 m minimum spacing and 1 km cluster radius of Google's example and found that a hexagonal planar arrangement could hold 367 satellites, more than four times the 81 in Google's layout; a three-dimensional layout they proposed holds more than either planar design once the cluster radius is at least 13.5 times the minimum spacing.^[84] At the program's announcement, Google said it and Planet would launch two prototype satellites by early 2027.^[13] A single prototype went up first: SpaceX's Transporter-18 rideshare carried it to orbit on October 1, 2026, launching from Vandenberg Space Force Base.^[47] Planet made contact and began commissioning the same day, and Google said the satellite was operating as expected.^[47][48] Google planned to gather data over the following weeks on how its TPUs cope with the physical stresses of spaceflight and with radiation and temperature extremes, and said it would put two satellites in orbit in 2027 to try out the laser links between them.^[48][49] On October 30, Google said the satellite had worked through its commissioning checklist and returned its first AI responses from orbit.^[81] Because that first flight is a single satellite, no orbital TPU-cluster result existed as of November 1, 2026.

Power and illumination#

Solar power in orbit begins with the total solar irradiance near Earth, about 1,361 W/m^2.^[16] If a panel were held normal to the Sun and converted 30 percent of that flux to electricity, the ideal photovoltaic area for 1 MW would be:

P / (eta x S) = 1,000,000 W / (0.30 x 1,361 W/m^2) = about 2,450 m^2.^[16]

That is a physics floor, not a spacecraft design. It excludes pointing losses, wiring and conversion losses, payloads other than the computers, end-of-life degradation, temperature effects, shadow periods, battery charging, and redundancy. The 2026 Google preprint's "up to eight times" annual energy comparison is between a panel in a favorable orbit and the same panel on the ground at mid-latitude. It is not an eightfold advantage for the complete power system.^[14]

A dawn-dusk sun-synchronous orbit can maximize illumination and give a comparatively stable thermal geometry, but "sun-synchronous" does not mean "always in sunlight." Eclipse duration depends on altitude, local solar time, season, orbital beta angle, and the size of the formation. ESA's Aeolus occupied a dawn-dusk orbit at about 320 km, yet passed through Earth's shadow for as long as 20 minutes per orbit over the winter hemisphere.^[17] Starcloud, which has applied to operate satellites in sun-synchronous orbits between 600 and 850 km, told the FCC in September 2026 that each spacecraft would be in sunlight for about 97 percent of its operational lifetime, which it said made the battery subsystem less critical than on a mission with substantial eclipse exposure.^[37][87] A real design therefore needs an orbit-specific eclipse analysis and either storage, workload throttling, redundant power, or some combination.

Orbit selection trades power against other constraints:

No orbit is preferred for every design. Edge processing should usually stay close to the sensor it serves. A cluster optimized for solar availability may be poorly placed for low-latency access to a particular ground market.^[14][17]

Heat rejection#

Space is a vacuum, so an external radiator cannot remove heat by convection. Internal heat can move through conduction, heat pipes, or pumped fluid loops, but the spacecraft ultimately rejects it as thermal radiation. For an ideal surface looking only at deep space, where a roughly 3 K sink is negligible beside a 300 K radiator, the rejected heat is approximately:^[18]

Q = epsilon x sigma x A x T_r^4.^[18]

Here Q is waste heat, epsilon is surface emissivity, sigma is the Stefan-Boltzmann constant, A is emitting surface area, and T_r is radiator temperature.^[18] The fourth-power temperature term makes a warmer radiator smaller, but chips and coolant loops limit how warm it may run.

The following values are transparent lower bounds calculated from the cited radiation equation with emissivity 0.90, a 3 K sink, view factor 1, and no absorbed sunlight, Earth infrared, or albedo. Area means active emitting area, not necessarily the projected footprint of a two-sided panel.^[18]

These figures should not be quoted as flight-system sizes. A radiator partly sees warm Earth, or absorbs solar and reflected energy, rejects less net heat. Heat exchangers, pumps, working fluid, pipes, deployment hardware, micrometeoroid tolerance, margins, and redundant loops add mass. Coatings also change during a mission: NASA notes that atomic oxygen in low Earth orbit, ultraviolet exposure, and radiation can alter optical properties and degrade thermal performance.^[19]

A 2026 technical preprint, using a specified high-sunlight 1 MW case rather than the ideal table above, estimated about 2,500 m^2 of radiator and 5,640 m^2 of beginning-of-life photovoltaic area.^[20] Its similarity to the 300 K ideal radiator result should not obscure the assumptions. Different chip efficiency, coolant temperature, radiator coating, geometry, orbit, or eclipse duty cycle changes the answer.

Radiation, faults, and reliability#

Space radiation causes several different failure modes. Total ionizing dose accumulates over time; displacement damage alters semiconductor materials; and a single energetic particle can cause a bit flip, transient, latch-up, or destructive burnout. ESA treats these as distinct effects whose severity depends on orbit, shielding, mission length, component process, and circuit design.^[21]

High-performance commercial chips offer far more compute than traditional radiation-hardened processors, but they need a system-level safety case. Measures can include shielding, error-correcting memory, memory scrubbing, watchdogs, checkpoint and restart, duplicate computation, power cycling, fault isolation, and a simpler radiation-tolerant controller supervising the commercial accelerator. Spaceborne Computer-1 is evidence that software management can work in a particular hosted environment, not proof that every GPU can survive every orbit for years.^[3][14][21]

Google's Trillium tests illustrate both promise and uncertainty. The June 2026 preprint modeled a five-year, 10 mm aluminum-equivalent shielded sun-synchronous mission at about 150 rad(Si) per year. In a 67 MeV proton beam at the University of California, Davis, Crocker Nuclear Laboratory, where the chips ran AI workloads while the team watched for errors such as bit flips,^[49][63] high-bandwidth-memory stress tests began showing irregularities after 2 krad(Si), while other end-to-end and compute tests ran to 15 krad(Si) without a total-dose hard failure;^[14] the peer-reviewed Joule version gives that ceiling as 10 krad(Si).^[63] The tests also observed silent data corruption and system interruptions. Google judged inference errors potentially manageable, but said the effects on long training runs and the effectiveness of system-level mitigations still require study.^[14] A proton-beam test cannot by itself validate the full particle spectrum, every component, manufacturing lot, thermal state, or years of autonomous operation. Google's prototype satellite is meant to measure, in orbit, how the TPUs cope with radiation and thermal extremes.^[49] After its first AI responses in late October 2026, Google said its next tests would examine how the TPUs hold up under thermal stress and radiation in orbit.^[81]

Reliability also includes ordinary hardware failures. Pumps, storage devices, deployable structures, solar arrays, and batteries can degrade or fail. Terrestrial data centers tolerate failed components because technicians replace them. An orbital service must instead carry spare capacity, isolate faults remotely, accept declining performance, or pay for robotic servicing or replacement launches.^[14][19] Starcloud's Johnston described the company's approach in October 2026: redundancy for critical parts such as pumps, extra solar capacity to cover panels that lose a few percent of output a year, chip testing at particle accelerators, and a three-month burn-in of Starcloud-2's chips at full power on the ground to catch early failures. He said the business case would collapse if chips failed 10 or 20 percent more often in orbit than on the ground, since they are by far the costliest part of the system.^[86]

Networks, latency, and workload placement#

Orbit shortens the path to another spacecraft but adds a bottleneck for terrestrial data. A low Earth orbit radio path can have less than 10 ms of one-way propagation delay to a visible ground station, yet that physical floor is not application latency. Ground-station visibility, scheduling, routing, weather, retransmission, queueing, encryption, and the terrestrial path can dominate.^[22]

Optical links provide high rates without consuming the same radio-frequency bandwidth, but they require narrow-beam acquisition, tracking, and pointing. NASA's TBIRD demonstration achieved a 200 Gbit/s space-to-ground optical downlink and sent 4.8 TB error-free in five minutes during a 2023 pass.^[23] That result proves a burst link, not continuous global capacity. Clouds and severe atmospheric attenuation can prevent an optical downlink, so networks need geographically diverse ground stations, storage, alternate routing, or radio fallback.^[24]

Tightly coupled AI training is harder than downloading finished results. Google's preprint estimates that its proposed cluster would need aggregate bandwidth on the order of 10 Tbit/s per link, compared with roughly 1 to 100 Gbit/s for surveyed commercial optical inter-satellite terminals. A bench setup reached 800 Gbit/s in one direction and 1.6 Tbit/s bidirectionally over a short free-space path, but it was not a moving, thermally constrained flight network.^[14] Close formation improves the optical link budget while making relative navigation, collision avoidance, plume impingement, and fault containment harder.

Workload placement determines whether the communications constraint closes:

The strongest demonstrated case is data reduction at its source. General cloud computing for users on Earth gives up that advantage because both inputs and outputs traverse the space-to-ground boundary.^[2][10][20]

Economics and lifecycle#

SpaceX's June 5 EU prospectus said orbital AI-compute satellites could begin deploying as early as 2028. SpaceX said it believed about 10 GW of annual deployment could support a commercially attractive business and described up to 100 GW per year as a long-term ambition. The 100 GW case assumes 100 kW of compute power per metric ton and 100 metric tons to orbit per Starship launch, requiring thousands of launches each year.^[44] These are forward-looking goals, not demonstrated performance. The same filing says the initiative is early-stage, unproven at commercial scale or at all, may fail, and has a timeline and launch cadence that may be difficult or impossible to determine.^[44] Musk forecast on August 14, 2026 that orbital compute would become "the only way to scale AI" probably in 2029 because of power availability and permitting problems on land.^[43] GAO reported the Department of Energy's projection that data centers could use up to 12 percent of US electricity in 2028. It also found that large centers would need solar arrays larger than any launched and assembled as of April 2026, that cooling at this scale remained unproven, and that more advanced data-transfer systems might be needed.^[45] The peer-reviewed Nature Electronics paper provides architectural and lifecycle frameworks, not a parity date.^[1] Under its simplified roofline model, a July 2026 preprint finds single-satellite inference potentially feasible but frontier-scale LLM training unlikely to compete with terrestrial data centers.^[46] A June 2026 IEEE Spectrum analysis put the cost of launching and running one GPU in orbit for a year at no less than ten times the terrestrial cost, even with an optimistic launch cost of $44 per kilogram.^[42] None of these sources establishes that orbit will be the only scalable option in 2029. AI Wiki's article on space-based data centers follows the computing side, including the AI workloads, chips and business cases behind these proposals.

Launch price is only one part of delivered compute cost, but current public pricing shows the scale of the gap. As of August 2026, SpaceX advertised a Falcon 9 rideshare minimum of $350,000 for 50 kg to sun-synchronous orbit, with additional mass at $7,000 per kilogram.^[25] That is a purchasable rideshare price for compatible payloads, not SpaceX's internal cost, a universal dedicated-launch price, or a forecast for Starship. Availability is a separate constraint. Starcloud's Johnston told TechCrunch in August 2026 that launch capacity was tight because of SpaceX's planned shift from Falcon 9 to Starship,^[54] and he said in October that part of the company's funding was going to reserve launch slots, with capacity especially tight from 2028 onward.^[86]

Google's preprint does not claim that launch already costs $200 per kilogram. It applies a learning-curve model and says prices might fall below that level in the mid-2030s if its assumptions, including high launch cadence, hold.^[14] The separate 2026 Turyshev analysis is also a preprint. For its representative 1 MW system it estimated 34 to 59 kg of total spacecraft mass per delivered kilowatt. With about 40 kg/kW and a terrestrial infrastructure benchmark of $10,000 to $40,000 per kilowatt, it derived only $250 to $1,000 per kilogram for spacecraft construction and launch together, before communications, operations, utilization, and finite lifetime.^[20] The result is a necessary-condition model, not a price forecast. Starcloud's Johnston has put the threshold for his company's design at about $500 per kilogram. In March 2026 he told TechCrunch that, at that launch price, Starcloud-3 could supply power for around $0.05 per kilowatt-hour, and said the company would not be competitive on energy costs until Starship flew frequently.^[89] Johnston repeated the figure in October 2026, arguing that solar power in orbit avoids the two largest costs of a new solar project built to power a data center on the ground (permits and land, then batteries and backup power) and needs roughly an eighth as many solar cells, so that launch is the main cost it adds; below about $500 per kilogram, he said, orbital computing would be much cheaper than terrestrial computing. He also said Starcloud's present edge-processing business was already economic at 2026 launch prices.^[86] The one-eighth figure compares solar cells alone, like the "up to eight times" panel comparison in Google's preprint,^[14] and Starcloud's own FCC response still gives each satellite a battery subsystem.^[87] The threshold is a company estimate, not a demonstrated cost.

A serious comparison needs at least:

• spacecraft development and qualification;
• solar, storage, conversion, radiator, propulsion, shielding, and communication mass;
• integration, launch, insurance, ground stations, and network operations;
• useful compute after radiation margins, redundancy, and thermal throttling;
• utilization and revenue during each ground or customer contact window;
• hardware lifetime, obsolescence, replacement cadence, and failed launches;
• disposal, collision avoidance, and any servicing system;
• terrestrial alternatives using low-carbon power, dry cooling, or better chips.^[14][20]

Avoided land and cooling water are real benefits only if the comparison uses the same service boundary. Moving compute to orbit does not remove manufacturing, ground terminals, terrestrial networking, launches, or replacement hardware. It also does not make maintenance free. Software can be updated remotely, but physical repairs require crewed or robotic access, a resupply architecture, or disposal and replacement. Axiom explicitly lists software updates, resupply, and replacement modules as different maintenance paths.^[7]

Lower altitude can shorten passive deorbit time but increases drag and propulsion demand. Higher altitude can lengthen life and reduce drag, but failed spacecraft persist longer. Hardware may also become economically obsolete before its solar arrays or structure wear out. Designs therefore need a full cycle from production through launch, useful compute-years, replacement, and end-of-life disposal.^[14][17][20]

Regulation, debris, and public risk#

The largest proposal in the public regulatory record is SpaceX's Orbital Data Center System. The FCC's February 4, 2026 public notice says SpaceX applied for up to one million satellites at 500 to 2,000 km, using 30-degree and sun-synchronous inclinations, optical inter-satellite links, and Ka-band Earth links that could route through Starlink.^[26] The application predates the Starmind name and does not mention Starmind or AI1. Musk confirmed on June 23, 2026 that Starmind would be the constellation's official name,^[34] and SpaceX said on August 4 that each Starmind satellite would carry NVIDIA Rubin GPUs and Vera CPUs.^[35] "Accepted for filing" means the agency found the application complete enough for public processing and comment. It does not grant authority to deploy the system.

The FCC proceeding covers an orbital communications system, specified frequency bands, telemetry, tracking and command, deployment milestones, and requested waivers. It is not blanket approval of every launch, satellite, or reentry. Optical crosslinks reduce dependence on radio-frequency links within the constellation, but SpaceX still requested Ka-band Earth links.^[26]

Four other companies have asked the FCC to license orbital-data-center constellations. The FCC accepted Starcloud's application for up to 88,000 satellites for filing on March 13, 2026.^[37] Blue Origin filed its 51,600-satellite Project Sunrise request in March,^[38] and the FCC accepted it for filing on April 17.^[57] Cowboy Space's Stampede system, 20,000 satellites in dawn-dusk sun-synchronous orbits at 700 to 1,000 km, was accepted on June 12,^[59][64] and Orbital Compute's request for up to 100,000 satellites in sun-synchronous orbits at 500 to 850 km was accepted on July 31.^[65] On September 10, Orbital picked Reflex Aerospace to design the bus for Orbital-1, a satellite planned with peak power approaching 250 kW.^[66] Cowboy Space, which was founded as Aetherflux to work on space-based power and renamed itself on May 11, 2026 as it expanded into orbital data centers and launch,^[79] launched Reason-1 on SpaceX's Transporter-18 rideshare on October 1, 2026. The satellite is designed to test beaming power from orbit to Earth, not computing.^[80]

On August 28, 2026, the FCC's Space Bureau told SpaceX that its application needed more information before further processing, including orbital debris details missing from the filing.^[55] Letters dated the same day appear in the FCC records of the other four applications.^{[56][57][58][59]} SpaceX replied on September 16 that each Starmind satellite would have a mass under 4,000 kg, less than 1,500 m^2 of solar arrays, and less than 200 m^2 of radiators; that the satellites would make targeted reentries toward the ocean; and that no part would reach the ground with more than 15 joules of kinetic energy.^[67] Starcloud's reply, dated the same day, gave each satellite's gross footprint as roughly 6 by 220 m, most of it taken up by two deployable silicon solar arrays measuring about 5 by 107 m apiece, and the debris-assessment software runs filed with it assume a 2,000 kg spacecraft, which they name Starcloud-3.^[87][88] Cowboy Space answered on September 18, after the bureau had extended its deadline. It said its satellite design was not final, said each satellite would carry a pair of equal solar arrays, each roughly 57 m long and 30 m wide, with a maximum cross-section of about 3,870 m^2, and said the satellites would make targeted reentries.^[78] As of November 1, 2026, none of the five applications had been granted.^{[36][56][57][58][59]} The American Astronomical Society had petitioned on March 6 to deny SpaceX's application,^[40] and environmental groups asked the FCC on July 8 to pause licensing orbital data centers until it had completed a programmatic environmental review.^[41] On September 30, the FCC adopted rules determining that the launch, deployment, and operation of space stations are not "major Federal action," the prerequisite that triggers an agency's obligations under the National Environmental Policy Act.^[60]

Post-mission disposal is a design input, not an afterthought. The FCC rule for US-licensed spacecraft ending missions in or passing through the region below 2,000 km and using uncontrolled reentry requires disposal as soon as practicable and no later than five years after mission end.^[27] NASA's active debris standard limits the calculated worldwide human-casualty risk from a single uncontrolled reentry to 1 in 10,000.^[28] Large structures may need design-for-demise, controlled reentry, or recovery. Even when each reentry meets an event-level criterion, a high replacement cadence creates an aggregate-risk and atmospheric-mass question.^[29]

Collision avoidance becomes harder as count, cross-sectional area, and differential orbital geometry increase. ESA's May 2026 environment report identifies 400 to 600 km as an altitude range with high concentrations of active and maneuverable satellites, says current global mitigation performance is insufficient for long-term sustainability, and reports 1,200 intact-object reentries during 2025.^[29] A compute constellation therefore needs maneuverability, tracking data, autonomous conjunction response, passivation, reliable disposal, and enough propulsion reserve to execute them. A failed, non-maneuvering server satellite is still space debris.

Environmental and astronomy impacts#

Claims that orbital data centers could be carbon neutral remain model-dependent. The 2025 Nature Electronics paper is peer-reviewed and provides a lifecycle accounting framework, but its title describes the objective being explored, not an operating carbon-neutral system.^[1] ASCEND similarly conditioned its favorable case on a much less emissive future launcher.^[12] A lifecycle assessment must state the terrestrial electricity mix, launch vehicle and propellant, reuse, payload mass fraction, lifetime, replacement rate, manufacturing emissions, and avoided ground infrastructure.

Launch and reentry emissions occur high in the atmosphere, where their chemistry and residence times differ from surface emissions. A peer-reviewed 2025 Journal of Geophysical Research study modeled a hypothetical 10 Gg per year injection consisting entirely of aluminum oxide from reentries. It found scenario-dependent accumulation and atmospheric changes, while stressing that the particles produced by actual vaporization remain poorly characterized.^[30] A 2026 peer-reviewed Earth's Future study likewise modeled radiative forcing and ozone effects from megaconstellation launch and reentry pathways and highlighted sensitivity to assumptions.^[31] These studies establish a risk requiring measurement and assessment, not a precise impact for an unbuilt orbital data center.

Astronomy effects depend on spacecraft number, altitude, attitude, size, reflectivity, radio emissions, and operational behavior. A 2026 peer-reviewed Monthly Notices of the Royal Astronomical Society paper modeled one hypothetical 5 GW facility with a roughly 4 km solar array. Under its assumptions, such a structure in sun-synchronous low Earth orbit would be exceptionally bright during twilight and large enough in angular extent to affect observations.^[32] Those results should not be applied directly to a small edge-compute satellite, but they show why operators of large arrays need credible brightness, occultation, radio-frequency, and debris analyses before deployment. More general peer-reviewed constellation studies find that no single scheduling or brightness mitigation protects every optical and near-infrared observation.^[33]

Open technical evidence gaps#

Nine question areas separate what orbital computing has demonstrated from what its proposed scale requires.

The evidence supports orbital edge computing today. It supports experiments with shared hosted compute and increasingly powerful commercial accelerators. It does not yet support treating proposed orbital clouds as proven substitutes for terrestrial data centers. The decisive measurements will come from openly reported power balance, radiator performance, fault rates, useful workload throughput, link availability, service utilization, replacement cadence, and lifecycle impacts from actual flight systems.^{[1][7][14][20]}

References#

1. The development of carbon-neutral data centres in space - Nature Electronics, peer-reviewed perspective, October 27, 2025.
2. Artificial Intelligence for Earth observation - European Space Agency.
3. Cutting-Edge Computing Goes Spaceborne - NASA Spinoff, February 11, 2025.
4. HPE Spaceborne Computer-2 returns to the International Space Station - Hewlett Packard Enterprise, January 30, 2024.
5. Researchers successfully train a machine learning model in outer space for the first time - University of Oxford, July 28, 2023.
6. Fast model inference and training on-board of satellites - arXiv preprint, July 17, 2023.
7. Orbital Data Centers - Axiom Space, accessed October 16, 2026.
8. Starcloud-1 - Starcloud, accessed August 12, 2026.
9. Bandwagon-4 - SpaceX, November 2, 2025.
10. Space based data centres - European Space Agency Nebula Public Library, 2022 to 2024.
11. Advanced Space Cloud for European Net zero emissions and Data sovereignty - European Commission CORDIS, updated January 14, 2026.
12. Thales Alenia Space reveals results of ASCEND feasibility study on space data centers - Thales Alenia Space, June 27, 2024.
13. Project Suncatcher explores powering AI in space - Google, November 4, 2025.
14. Towards a future space-based, highly scalable AI infrastructure system design - arXiv preprint, revised June 17, 2026.
15. NVIDIA Launches Space Computing, Rocketing AI Into Orbit - NVIDIA, March 16, 2026.
16. Solar Irradiance Science - NASA Goddard Space Flight Center.
17. Aeolus satellite - European Space Agency.
18. NIAC Phase I Final Report: Molecular Composition Analysis of Distant Targets - NASA Technical Reports Server, 2016.
19. State-of-the-Art of Small Spacecraft Technology: Thermal Control - NASA Small Spacecraft Systems Virtual Institute, May 7, 2026.
20. Orbital Data Centers: Spacecraft Constraints and Economic Viability - arXiv preprint, April 29, 2026.
21. Radiation-Hardness Components at scaled technology nodes - European Space Agency.
22. Delay-Tolerant Networking for Low Earth Orbit - NASA Technical Reports Server, 2009.
23. NASA's Record-Breaking Laser Demo Completes Mission - NASA, September 2024.
24. State-of-the-Art of Small Spacecraft Technology: Communications - NASA Small Spacecraft Systems Virtual Institute, 2026.
25. Smallsat Rideshare Program - SpaceX, accessed August 12, 2026.
26. Space Bureau Accepts for Filing SpaceX's Application for Orbital Data Centers - Federal Communications Commission, February 4, 2026.
27. Mitigation of Orbital Debris in the New Space Age, FCC 22-74 - Federal Communications Commission, September 29, 2022.
28. NASA-STD-8719.14C: Process for Limiting Orbital Debris - NASA, November 5, 2021.
29. ESA's Annual Space Environment Report, Issue 10.0 - European Space Agency Space Debris Office, May 1, 2026.
30. Investigating the Potential Atmospheric Accumulation and Radiative Impact of the Coming Increase in Satellite Reentry Frequency - Journal of Geophysical Research: Atmospheres, peer-reviewed article, 2025.
31. Radiative Forcing and Ozone Depletion of a Decade of Satellite Megaconstellation Missions - Earth's Future, peer-reviewed article, 2026.
32. The impact on astronomy of data centres orbiting Earth - Monthly Notices of the Royal Astronomical Society, peer-reviewed article, May 1, 2026.
33. Analytical simulations of the effect of satellite constellations on optical and near-infrared observations - Astronomy & Astrophysics, peer-reviewed article, 2022.
34. Confirmation of the Starmind name for SpaceX's AI-satellite constellation - Elon Musk on X, June 23, 2026.
35. SpaceX announcement of the Starmind AI1 partnership with NVIDIA - SpaceX on X, August 4, 2026.
36. Application summary for SAT-LOA-20260108-00016 - Federal Communications Commission ICFS, accessed November 1, 2026.
37. Applications accepted for filing, Report SAT-01982 - Federal Communications Commission, March 13, 2026.
38. 51,600 more satellites? Blue Origin adds another twist to the data center space race with Project Sunrise - GeekWire, March 21, 2026.
39. Starcloud-2 - Starcloud, accessed November 1, 2026.
40. Petition to deny SpaceX's orbital-data-center application - American Astronomical Society, March 6, 2026.
41. Coalition calls for rigorous federal review of space-based data center proposals - Earthjustice, July 8, 2026.
42. Why orbital data centers are harder than Silicon Valley thinks - IEEE Spectrum, June 11, 2026.
43. Forecast that orbital compute could become the only way to scale AI around 2029 - Elon Musk on X, August 14, 2026.
44. SpaceX EU prospectus - Space Exploration Technologies Corp., June 5, 2026.
45. Science & Tech Spotlight: Data Centers in Space - U.S. Government Accountability Office, GAO-26-109012, April 28, 2026.
46. The Cost and Network Limits of Space-Based AI Compute - Kees van Berkel, arXiv preprint, July 15, 2026.
47. Planet Launches Suncatcher, Tanager-2, and 18 SuperDove Satellites - Planet Labs PBC, October 1, 2026.
48. Google's Project Suncatcher prototype satellite is in orbit - Google, October 1, 2026.
49. Behind Project Suncatcher, our moonshot to put AI in space - Google, September 24, 2026.
50. China launches eight satellites for space-based computing - CGTN, October 14, 2026.
51. Chinese launches boost Guowang constellation, add space computing sats - SpaceNews, October 15, 2026.
52. SpaceXAI Adopts NVIDIA Vera CPU to Accelerate Agentic AI at Massive Scale - NVIDIA, August 24, 2026.
53. Space-optimized Vera Rubin NVL72 system planned for launch to orbit in the fourth quarter of 2027 - Elon Musk on X, August 24, 2026.
54. Starcloud raises $250 million for orbital data centers as launch options dry up - TechCrunch, August 21, 2026.
55. Letter requesting additional information on ICFS File No. SAT-LOA-20260108-00016 - Federal Communications Commission Space Bureau, August 28, 2026.
56. Application summary for SAT-LOA-20260202-00073, Starcloud Orbital Datacenter System - Federal Communications Commission ICFS, accessed November 1, 2026.
57. Application summary for SAT-LOA-20260310-00118, Project Sunrise - Federal Communications Commission ICFS, accessed November 1, 2026.
58. Application summary for SAT-LOA-20260624-00251, Orbital Datacenter System - Federal Communications Commission ICFS, accessed November 1, 2026.
59. Application summary for SAT-LOA-20260323-00135, Stampede Data Center System - Federal Communications Commission ICFS, accessed November 1, 2026.
60. FCC Modernizes its Approach to NEPA Environmental Regulations - Federal Communications Commission, September 30, 2026.
61. China launches first of 2,800 satellites for AI space computing constellation - SpaceNews, May 14, 2025.
62. Amid failure, Wenchang celebrates milestone as China launches AI satellites and prepares for lunar south pole mission - NASASpaceFlight, August 12, 2026.
63. Toward a future space-based, highly scalable AI infrastructure system design - Joule, peer-reviewed article, October 1, 2026.
64. Applications accepted for filing, Report SAT-02007 - Federal Communications Commission, June 12, 2026.
65. Applications accepted for filing, Report SAT-02024 - Federal Communications Commission, July 31, 2026.
66. Orbital Selects Reflex Aerospace as Exclusive Satellite Platform Partner for Planned Space Data Center Constellation - Reflex Aerospace, September 10, 2026.
67. Ex parte response on the Starmind application, ICFS File No. SAT-LOA-20260108-00016 - Space Exploration Technologies Corp., September 16, 2026.
68. Axiom Space, Kepler Ready Orbital Computing for Quantum Era - Axiom Space, October 6, 2026.
69. Axiom Space, Spacebilt Announce Orbital Data Center Node Aboard International Space Station - Axiom Space, September 16, 2025.
70. China launches space computing satellite constellation - Xinhua via China Daily, May 15, 2025.
71. Space-AI computing solutions paving way for faster smart tech development - China Daily, July 29, 2026.
72. 活力中国调研行丨探访之江实验室：科创活水充分涌动 (A visit to Zhejiang Lab) - Xinhua, July 28, 2026.
73. 沐曦股份上市后首份半年报出炉：同比扭亏为盈 GPU产品出货量显著提升 (MetaX releases its first half-year report since listing) - Cailian Press, August 30, 2026.
74. Is It Lunacy to Put a Data Center on the Moon? - IEEE Spectrum, February 25, 2025.
75. Intuitive Machines' IM-2 Moon mission ends with lander on its side - Spaceflight Now, March 7, 2025.
76. Lonestar Expands Launch Plans for Orbital Data Storage Payload - Via Satellite, April 15, 2026.
77. Sidus Space to Launch LizzieSat-4 and LizzieSat-5 Together on SpaceX Bandwagon-6 - Sidus Space, September 28, 2026.
78. Supplement responding to the Space Bureau letter, ICFS File No. SAT-LOA-20260323-00135 - Cowboy Space Corp., September 18, 2026.
79. Aetherflux Rebrands as Cowboy Space, Expanding Plans to In-Orbit Compute and Launch - Via Satellite, May 11, 2026.
80. The First Mission Milestones Onboard SpaceX's Transporter-18 Rideshare Launch - Via Satellite, October 1, 2026.
81. We've received our first response from the Project Suncatcher prototype satellite - Google, October 30, 2026.
82. Google's first Suncatcher orbital data center test launches October 1 - Ars Technica, September 24, 2026.
83. Google's first data center satellite is online and writing poetry - TechCrunch, October 30, 2026.
84. Designing Dense Satellite Clusters for Distributed Space-based Datacenters - Jules Pénot and Hamsa Balakrishnan, arXiv preprint of AAS 26-754, revised September 28, 2026.
85. 'Greetings, earthlings': Nvidia-backed Starcloud trains first AI model in space as orbital data center race heats up - CNBC, December 10, 2025.
86. Nvidia GPUs in Orbit: Inside Starcloud's $2B Factory - Sourcery with Molly O'Shea (YouTube), October 31, 2026.
87. Supplemental response to the Space Bureau letter, ICFS File No. SAT-LOA-20260202-00073 - Starcloud, Inc., September 16, 2026.
88. Debris Assessment Software logs for Starcloud-3, ICFS File No. SAT-LOA-20260202-00073 - Starcloud, Inc., September 16, 2026.
89. Starcloud raises $170 million Series A to build data centers in space - TechCrunch, March 30, 2026.