(1) File Number SAT-STA-20260406-00154 — Request for Extension of Special Temporary Authority
Space Exploration Technologies Corp. (SpaceX) requests a 60-day extension of its existing Special Temporary Authority (File No. SAT-STA-20251215-00385) to conduct in-orbit transmission testing of its non-geostationary orbit second-generation satellites using 2GHz spectrum within the scope of SpaceX’s pending transaction with EchoStar — downlink only.
Specifically, SpaceX proposes to conduct limited in-orbit testing of services in the 2185-2190 MHz band within a 50km diameter area around its facility in Sunnyvale, California. The testing will involve intermittent transmissions from a single SpaceX satellite. Although these transmissions will comply with applicable 3GPP standards, the SpaceX satellite will only communicate with SpaceX test equipment. SpaceX will rely on its existing authorized frequencies for backhaul and telemetry, tracking, and control (TT&C).
These tests will be conducted on an unprotected, non-interference basis, using only frequencies covered by the pending application for transfer of specific spectrum authorizations from EchoStar to SpaceX. SpaceX has also applied for authority to use these frequencies to provide mobile satellite service in the United States and globally. Currently, EchoStar holds exclusive satellite and terrestrial use rights for the 2185-2190 MHz band nationwide, including the test area. EchoStar has provided a consent letter for the proposed testing, and SpaceX will comply with all applicable out-of-band emission limits.
In the unlikely event that harmful interference occurs, SpaceX will take all reasonable measures to eliminate it. Issues can be reported to [email protected], which is connected 24/7 to on-call technical staff pagers. The Commission has sufficient reason to approve this request in the public interest — enabling SpaceX to continue testing its planned Direct-to-Cell service and advancing U.S. leadership in space-based supplemental coverage.
(2) File Number SAT-STA-20260406-00153 — Request for Extension of Special Temporary Authority
SpaceX requests a 60-day extension of its existing Special Temporary Authority (File No. SAT-STA-20251125-00340) for in-orbit downlink-only transmission testing using 2GHz spectrum from the pending EchoStar transaction. This filing covers testing in a 50km diameter area around SpaceX’s facility in Redmond, Washington, using the same 2185-2190 MHz band parameters as the Sunnyvale filing above.
The technical parameters, interference mitigation commitments, and public interest justification are identical to File Number SAT-STA-20260406-00154. EchoStar has provided consent, and all testing proceeds on a non-interference basis. The two parallel filings cover SpaceX’s two primary West Coast test facilities, enabling broader geographic validation of the Direct-to-Cell service architecture. Aomway’s satellite communications analysts note that parallel testing at both Sunnyvale and Redmond suggests SpaceX is validating location-dependent performance variables — including different latitude satellite pass geometries and ground station infrastructure configurations.
(3) File Number SAT-STA-20260401-00150 — Request for Further Extension of Special Temporary Authority (E-Band Gateway Links)
The FCC previously granted SpaceX Special Temporary Authority to use E-band spectrum for transmitting signals between gateway earth stations and its newly licensed second-generation satellites. The Commission partially approved SpaceX’s application for a license to use this gateway spectrum but deferred final action on the second-generation satellite application pending further review.
SpaceX has been operating under this STA for months without receiving any complaints from authorized spectrum users. These gateway links enable Gen2 satellites to communicate with gateway earth stations to confirm operational status, evaluate overall network performance, and prepare for rapidly growing consumer demand for high-speed, low-latency broadband service — demand that may exceed existing Ka-band gateway capacity.
SpaceX requests a further 60-day extension to continue operating these E-band gateway links while its underlying space station application remains pending. Operations continue on a non-interference basis. In the unlikely event of harmful interference, contact [email protected] (24/7 pager-connected).
(4) File Number SAT-LOA-20260108-00016 — The Orbital Data Center Constellation
This is the most ambitious filing: SpaceX applies for authority to launch and operate a satellite constellation with unprecedented computing capability to support advanced AI models and applications. By directly utilizing near-continuous solar energy and operating at extremely low operational and maintenance costs, these satellites will achieve revolutionary cost and energy efficiency while significantly reducing the environmental impact associated with ground-based data centers.
Vision: A Million-Satellite Orbital Data Center
Deploying a constellation of up to one million satellites operating as orbital data centers is, in SpaceX’s words, “a first step toward a Kardashev Type II civilization” — one capable of harnessing the full energy of the Sun — while providing AI-powered application support to billions of people.
SpaceX is designing its satellite system to meet the explosive growth in data processing demand driven by AI, machine learning, and edge computing — demand that has begun exceeding terrestrial capacity. Key parameters of the proposed constellation:
- Constellation size: Up to 1,000,000 satellites
- Altitude range: 500 km to 2,000 km
- Inclinations: 30° to Sun-synchronous orbit
- Orbital layer spacing: 50 km per shell (leaving collision-avoidance room for other systems)
- Primary communication: High-bandwidth laser inter-satellite links (petabit-scale capacity)
- Backup communication: Ka-band TT&C on non-interference, unprotected basis (18.8-19.3 GHz downlink, 28.6-29.1 GHz uplink)
Why Orbital Data Centers?
AI-driven data processing demand is projected to more than double global data center electricity consumption by 2035 — reaching approximately 1,200-1,700 TWh annually, up to 4% of global electricity. Building power plants and ground infrastructure to sustain this growth is extremely challenging.
SpaceX argues that with fully reusable Starship rockets (capable of delivering hundreds of tons of payload to orbit annually at scale), orbital processing capacity can reach scales and speeds unattainable on Earth, with significantly lower environmental impact. Key advantages:
- Near-continuous solar energy: Satellites at high-altitude sun-synchronous orbits can achieve 99%+ sun exposure, eliminating dependence on the terrestrial power grid
- Passive radiative cooling: The vacuum of space enables passive heat dissipation, far more efficient than ground data centers that consume billions of gallons of water for cooling
- Scalability: Launching 1 million tons of satellites per year (at 100 kW computing per ton) would add 100 GW of AI computing capacity annually with minimal ongoing operational needs
- Cost trajectory: As Starship reduces launch costs, the economics of space-based AI computing will surpass ground-based deployment within a few years
Aomway’s satellite industry analysts observe that this filing represents a fundamental reimagining of cloud computing infrastructure — moving the most energy-intensive workloads (AI training and inference) off-grid and into orbit. If realized, this architecture could reshape the competitive landscape for AI infrastructure providers.
Sustainability and Debris Mitigation
SpaceX emphasizes that the constellation will inherit proven sustainability strategies from its Gen1 and Gen2 broadband systems:
- Initial testing at very low orbits (atmospheric drag ensures rapid deorbit of any malfunctioning satellites)
- Automated collision avoidance systems with low-latency risk assessment
- Electric propulsion for precise maneuvering
- End-of-life disposal via controlled atmospheric reentry or heliocentric disposal orbits
- Redundant propulsion and propellant reserves for all mission phases
- Satellite brightness mitigation for astronomy (continuing SpaceX’s collaboration with the scientific community)
- Reliability exceeding 99% (demonstrated across ~10,000 Gen1/Gen2 satellites)
Spectrum Usage
The system will primarily rely on inter-satellite laser links (no spectrum required for inter-satellite communication). Backup TT&C uses Ka-band: 18.8-19.3 GHz (space-to-Earth) and 28.6-29.1 GHz (Earth-to-space), both designated as primary for non-GSO FSS. All RF operations are on a non-interference, unprotected basis.
Have questions about satellite spectrum management, FCC filing procedures, or orbital data center architecture? Contact Aomway at [email protected] — our satellite communications team provides regulatory consulting, spectrum analysis, and space system architecture design.
Frequently Asked Questions
1. What is the significance of SpaceX’s 2GHz spectrum testing with EchoStar?
The 2GHz band (2185-2190 MHz) is valuable for mobile satellite services because it offers better propagation characteristics than higher frequencies — it penetrates buildings and foliage more effectively. SpaceX’s testing with this spectrum, acquired through a pending transaction with EchoStar, likely supports its Direct-to-Cell (DtC) service expansion. The 2GHz band could complement SpaceX’s existing DtC frequencies (which include bands near 1.9GHz) by providing additional capacity and coverage in challenging environments. The STA extensions at both Sunnyvale and Redmond indicate active testing across multiple geographic locations.
2. Why does SpaceX need E-band gateway links for Gen2 satellites?
E-band (71-86 GHz) offers vastly more bandwidth than Ka-band (27-40 GHz), enabling multi-gigabit-per-second throughput per gateway link. As SpaceX’s Gen2 constellation grows and user demand increases, existing Ka-band gateway capacity may become a bottleneck. E-band gateway links provide the backhaul capacity needed to route traffic between satellites and ground stations at rates that match the constellation’s growing throughput. The non-interference STA allows SpaceX to continue operations while the FCC reviews the full license application — a process that can take years for complex non-GSO systems.
3. Is the million-satellite orbital data center realistic?
At face value, one million satellites is unprecedented — the current total number of active satellites in orbit is approximately 10,000. However, SpaceX’s filing is a long-term vision document, not a near-term deployment plan. The constellation would be built incrementally over many years (or decades), contingent on Starship achieving high-frequency launch cadence. Key feasibility questions include: satellite mass production at scale, radiation-hardened AI accelerator chips, laser link network management at million-node scale, and orbital debris management. The filing is best understood as SpaceX staking a regulatory claim — establishing priority in the FCC filing queue for this type of system, similar to how early Starlink filings established priority for broadband satellite service.
4. How would orbital data centers compare to ground-based AI facilities?
Ground-based AI data centers (like those operated by NVIDIA, Google, Microsoft, and Meta) face three growing constraints: electricity availability (many regions cannot support new data center power loads), water consumption for cooling (billions of gallons annually), and land use conflicts. Orbital data centers address all three: unlimited solar energy, passive radiative cooling in vacuum, and no terrestrial land use. The tradeoff is launch cost — currently $1,000-3,000/kg to LEO, though Starship aims to reduce this to $10-100/kg at scale. At sufficiently low launch costs, the economics favor space for power-intensive AI training workloads, while latency-sensitive inference may remain ground-based. Aomway’s analysis suggests a hybrid architecture is most likely: orbital data centers for AI training (latency-insensitive), ground stations for real-time inference.
5. What regulatory challenges does the orbital data center constellation face?
Major regulatory hurdles include: (1) Spectrum coordination — even backup Ka-band operations require coordination with existing NGSO FSS operators (OneWeb, Amazon Kuiper, Telesat). (2) Orbital debris management — one million satellites, even with 99% reliability, means 10,000 potential failures. The FCC and international regulators will require robust debris mitigation plans. (3) Space traffic management — conjunction assessment at this scale requires unprecedented coordination with the Space Force’s 18th Space Defense Squadron and other operators. (4) Astronomy and dark skies — one million additional satellites will dramatically increase sky brightness, potentially impacting ground-based astronomy. (5) International coordination — ITU filing priorities, national security reviews, and bilateral agreements will all play roles. SpaceX’s filing is the opening move in what will likely be a multi-year regulatory process.
Interested in satellite communications, spectrum policy, or space-based AI infrastructure? Contact Aomway at [email protected] — our team provides satellite system consulting, regulatory analysis, and space technology market intelligence.
