The Invisible War for Orbital Spectrum: How Nations Are Carving Up Space's Most Valuable Resource

Key Takeaway

Radio spectrum - the range of electromagnetic frequencies that carry data between satellites and the ground - is finite, invisible, and arguably the most valuable resource in space. Nations and corporations are engaged in an escalating contest to claim it through a filing system designed in the Cold War era and administered by a United Nations agency in Geneva. The stakes are measured in trillions of dollars: the US C-band spectrum auction alone raised $81 billion. SpaceX paid $17 billion for a set of satellite spectrum licenses. China has filed plans for over 249,000 satellites. And the system that governs who gets what is buckling under pressure it was never built to handle.

How Spectrum Gets Claimed

The International Telecommunication Union (ITU), headquartered in Geneva, coordinates orbital spectrum allocation for 193 member states. The process works roughly like this:

1. A country files an Advanced Publication Information (API) with the ITU, describing the satellite system’s frequencies and orbital parameters. Cost: 570 Swiss Francs. This puts other operators on notice.

2. A Coordination Request follows, where the filing country must demonstrate that its planned satellite system will not cause harmful interference to existing systems. Cost: 5,710 to 67,000 CHF, depending on complexity.

3. A Notification filing formalizes the frequency assignment. Cost: 15,910 to 116,000 CHF.

4. The satellite must be “brought into use” within 7 years. Since WRC-2019, non-geostationary constellations face milestone requirements: 10% deployed within 2 years of the deadline, 50% within 5 years, 100% within 7 years - or the license narrows to only the satellites actually in orbit.

The ITU receives 400 to 500 requests for new satellite network systems annually. Only about one-tenth ever reach actual launch.

The fundamental principle is filing priority: the first system to file and deploy has priority over later entrants, who must demonstrate their signals will not interfere. This is not ownership - no one owns spectrum under international law - but it functions as a de facto property right. First come, first served.

The Paper Satellite Problem

There is a loophole in this system, and it has been exploited aggressively.

Filing an API costs 570 Swiss Francs. For that price, you reserve spectrum for a satellite system you may never build. These phantom reservations are known as “paper satellites” - orbital slot claims that exist only on paper, blocking access for others without any hardware in space.

The Tonga precedent (1990): The Kingdom of Tonga - population 100,000, GDP roughly $500 million - filed for 16 geostationary orbital positions with the ITU. It then leased slots to Unicorn, a Colorado-based company, and auctioned remaining positions for $2 million per year. Tonga never launched a satellite. It was selling real estate in space that it had claimed for the price of a filing fee.

Rwanda’s 327,320 satellites (2021): In September 2021, Rwanda filed with the ITU for a constellation called Cinnamon-937, comprising 327,320 satellites across 27 orbital shells. The filing was linked to Marvel Space Communications, a company connected to Greg Wyler - the same entrepreneur who founded OneWeb. Astrophysicist Jonathan McDowell called it likely a strategic play to “establish that not all of LEO gets owned by the Americans.”

China’s 193,000-satellite filings (December 2025): The Institute of Radio Spectrum Utilization and Technological Innovation filed two mega-constellations with the ITU: CTC-1 (96,714 satellites) and CTC-2 (96,714 satellites) - totaling 193,428 satellites in 3,660 orbital planes each. These filings came on top of China’s existing Guowang (~13,000), Qianfan (15,000+), and Honghu-3 (~10,000) programs, bringing China’s total filings to approximately 249,000 satellites.

In a 2009 enforcement effort, the ITU investigated satellite networks on its Master International Frequency Register and removed 145 of them - representing 45% of the networks examined. A University of British Columbia study found filings representing approximately 1 million new satellites across 300 planned constellations globally.

The filing system was designed for an era when building a satellite cost hundreds of millions of dollars and the idea of filing for spectrum you would not use was absurd. In an era where filings cost hundreds of Swiss Francs and the strategic value of spectrum reservation is measured in billions, the system is being gamed at industrial scale.

The $17 Billion Frequency

If anyone doubts that spectrum has real monetary value, SpaceX provided a definitive answer in September 2025.

SpaceX paid $17 billion to acquire EchoStar’s AWS-4 and H-block spectrum licenses - 50 MHz of bandwidth in frequencies suitable for direct-to-cell satellite communications. The deal included up to $8.5 billion in cash, $8.5 billion in SpaceX stock, and approximately $2 billion in interest payments.

Fifty megahertz. Seventeen billion dollars. That is $340 million per megahertz.

For context, the US C-band spectrum auction in January 2021 - the largest spectrum auction in history - raised $81.11 billion for 280 MHz of bandwidth between 3.7 and 3.98 GHz. Verizon spent $45.45 billion. AT&T paid $23.41 billion. Since 1994, FCC spectrum auctions have raised over $233 billion for the US Treasury across 100 auctions.

On terrestrial networks, spectrum is the primary bottleneck. Every wireless carrier, every WiFi router, every Bluetooth device operates within allocated frequency bands. When demand exceeds available spectrum, the result is congestion, throttling, and dropped connections. In space, the same constraint applies - multiplied by the fact that thousands of satellites sharing the same frequencies must avoid interfering with each other and with ground-based systems.

SpaceX vs. Amazon: The Filing Wars

The competition between SpaceX and Amazon has extended from the launch pad to the regulatory arena, with the FCC as the battlefield.

SpaceX’s Gen2 constellation: SpaceX filed for up to 29,988 second-generation satellites across multiple orbital shells and frequency bands (Ku, Ka, and E-band). In December 2022, the FCC granted a partial authorization for 7,500 satellites. In January 2026, the FCC approved an additional 7,500, bringing the total to 15,000 approved Gen2 satellites with multi-band frequency support and direct-to-cell capability. The remaining ~15,000 are still pending.

Amazon Leo (formerly Kuiper): Amazon holds authorization for 3,236 satellites with FCC milestone deadlines: 50% deployed by July 2026, 100% by July 2029. In January 2026, Amazon filed for an extension of the July deadline.

The corporate warfare has been unusually public:

• SpaceX filed a formal FCC complaint accusing Amazon of violating orbital debris mitigation rules, claiming Amazon launched satellites into orbits above 450 km on eight occasions without submitting an amended debris plan
• SpaceX claimed its satellites had to perform 30 collision-avoidance maneuvers within hours of an Ariane launch deploying Amazon satellites
• Amazon fired back that SpaceX itself had launched Amazon satellites to 460 km altitude “without raising objections at the time”
• Amazon called on the FCC to reject SpaceX’s revised Gen2 plan, arguing it would create unacceptable interference with Kuiper’s operations

This is not a polite regulatory discussion. It is a multi-billion-dollar battle for orbital real estate, fought through FCC filings and public accusations. The spectrum that each company secures - or blocks the other from accessing - will determine the competitive landscape for decades.

China’s 249,000-Satellite Offensive

China’s satellite filings represent something qualitatively different from a commercial spectrum grab. They are a strategic national program with military, economic, and geopolitical objectives.

The constellations:

Program	Operator	Satellites Filed	In Orbit (Apr 2026)
Guowang (GW-A59 + GW-2)	China SatNet (state-owned)	~13,000	~100
Qianfan (Thousand Sails)	Shanghai Spacecom	15,000+	~90
Honghu-3	Landspace/Hongqing Tech	~10,000	0
CTC-1	RSDTII	96,714	0
CTC-2	RSDTII	96,714	0
China Mobile	China Mobile	~2,664	0
Total		~249,000+	~190

China SatNet - the company behind Guowang - was established by the Chinese government in April 2021 specifically to build a national satellite internet constellation. The timing was not coincidental. It came after Starlink’s high-profile deployment in beta testing demonstrated the strategic potential of LEO broadband.

The military driver: PLA researchers from the National University of Defense Technology published papers describing Starlink as “a security threat in nuclear, space, and cyber domains.” A study led by researcher Ren Yuanzhen from the Beijing Institute of Tracking and Telecommunications (under the PLA Strategic Support Force) proposed “a combination of soft and hard kill methods” to disable Starlink satellites. A simulation found that approximately 1,400 Starlink satellites could be approached by just 99 Chinese satellites in 12 hours.

RAND analysts have confirmed that China’s military assessment of Starlink as a weapon system directly shapes its own LEO constellation strategy. The constellations are dual-use by design - civilian connectivity and military communications infrastructure built into the same network.

The December 2025 filings for 193,000 additional satellites (CTC-1 and CTC-2) may or may not represent systems that will actually be built. But as ITU filings, they serve a strategic purpose regardless: they claim spectrum, create coordination burdens for other operators, and establish a negotiating position for future WRC conferences.

Starlink in Ukraine: The Proof of Concept

The event that transformed orbital spectrum from a technical regulatory issue into a matter of national security was Starlink’s deployment in Ukraine.

Within days of Russia’s February 2022 invasion, SpaceX activated Starlink service in Ukraine. At its peak, as many as 200,000 Starlink terminals operated in the country - making Ukraine the largest Starlink deployment in Europe. Poland alone contributed 19,500 of the 47,000 terminals initially delivered.

The military implications were immediate and profound:

• Ukrainian forces used Starlink for command and control, drone operations, and battlefield communications
• When Russia’s initial attack destroyed much of Ukraine’s telecommunications infrastructure, Starlink provided resilient backup connectivity for hospitals, railways, and government operations
• Russia attempted to jam Starlink signals. SpaceX deployed software updates that neutralized the jamming within hours - demonstrating an asymmetric advantage where a software-defined satellite network could adapt faster than an adversary could attack

The Pentagon eventually signed a $537 million contract for Starlink terminals in Ukraine through 2027, giving the military direct control over where the service operated.

Russia’s response extended beyond jamming to broader electronic warfare. Between August 2023 and April 2024, approximately 46,000 GPS interference incidents were reported over the Baltic Sea, attributed to Russian equipment at Kingissepp, 20 km from Estonia’s border. 310,000 flights were affected by GPS interference globally in 2024 - a 400% increase over three years. Satellites now face over 10,000 interference incidents per year.

Iran deployed its own counter-Starlink measures in January 2026: a three-mechanism “kill chain” combining GPS blinding, Ku-band saturation, and physical hardware targeting. The GPS spoofing caused Starlink terminals to report false locations, resulting in 30-80% packet loss.

Orbital spectrum is no longer just a commercial resource. It is contested military terrain.

When the UK Bought a Satellite Company for Its Spectrum

OneWeb’s bankruptcy in March 2020 became a case study in the strategic value of orbital spectrum rights.

OneWeb had raised $3.4 billion and deployed 74 satellites before COVID-19 forced SoftBank to cut off funding. The company filed for Chapter 11. On paper, it was a failed satellite venture - another entry in the industry’s long history of bankruptcies.

But OneWeb had one asset that made it worth rescuing: global Ku- and Ka-band spectrum priority rights. In August 2019 - just months before the bankruptcy - OneWeb had met the ITU’s minimum requirements by having 6 satellites broadcasting at the required frequencies for 90 consecutive days. This secured its filing priority.

Under the new ITU rules that took effect in January 2020, if OneWeb had been liquidated instead of rescued, those priority spectrum rights would have eventually lapsed - and been available for other operators (read: SpaceX or Amazon) to claim.

The UK government paid $500 million for a 42% stake - not because it wanted to be in the satellite business, but because it wanted to control spectrum rights that post-Brexit Britain could leverage for sovereign communications capability. The government secured a “golden share” giving it final say over any future sale of the company and over future access to OneWeb technology by other countries on national security grounds.

The spectrum was the asset. The satellites were secondary.

OneWeb later filed for an expanded 48,000-satellite constellation during its bankruptcy proceedings - a move widely interpreted as a defensive spectrum claim rather than a realistic deployment plan. When Eutelsat completed its $3.4 billion merger with OneWeb in September 2023, the combined entity held one of the most comprehensive sets of NGSO spectrum rights in the industry.

The Developing World’s Access Problem

The ITU’s filing-priority system was designed to be equitable. In practice, it favors nations with the technical capacity and financial resources to file quickly, build satellites, and meet deployment milestones.

Consider the numbers: filing an API costs 570 CHF, but building and launching a satellite constellation costs billions. The system is nominally open to all 193 member states. In reality, only a handful of countries can afford to participate meaningfully. The United States, China, the UK (through OneWeb/Eutelsat), Japan, and a few European nations dominate satellite spectrum allocations.

For developing nations - particularly in Africa and South America, where satellite internet could have the greatest impact on connectivity - the system creates a paradox. LEO satellite internet offers a path to connecting populations that terrestrial infrastructure will never economically reach. But the spectrum required for that connectivity is being claimed by wealthy nations and corporations at a pace that outstrips developing-world policy development.

The African Telecommunications Union is working to harmonize spectrum-allocation policies across the continent, and multiple African nations participated actively at WRC-23 for the first time. Resolution 219 from the ITU’s 2022 Plenipotentiary Conference directly encouraged member states to study equitable access to orbit and spectrum resources. WRC-23 formally agreed to study equitable procedures for Q and V band spectrum access.

But “studying equity” while Starlink deploys 10,000 satellites and China files for 249,000 is a race the developing world is losing by the day. Each satellite deployed, each filing registered, each coordination requirement established creates one more obstacle for late entrants.

2.6 billion people worldwide still lack internet access - many in regions where satellite may be the only viable delivery mechanism. The question is whether the spectrum they need to receive that service will still be available by the time their nations are in a position to use it - or whether it will have been claimed by operators who serve different markets entirely.

The Laser Escape Hatch

There is one technology that sidesteps the entire spectrum problem: laser optical communications.

Starlink satellites are equipped with three space lasers each - optical inter-satellite links (ISLs) operating at approximately 200 Gbps per link. By the end of 2023, over 9,000 lasers connected Starlink satellites in orbit, moving 42 petabytes of data per day through the vacuum of space.

The key distinction: optical transmission between satellites is license-free. There is no ITU filing required, no spectrum allocation, no coordination with other operators. Laser frequencies are so high (hundreds of terahertz) and the beams so narrow that interference between satellite laser links is virtually impossible.

This creates a parallel, unregulated communications layer in space. Data can travel across Starlink’s constellation - from satellite to satellite to satellite - without ever touching the regulated radio spectrum, except for the final uplink and downlink to the ground.

For the ground link, spectrum is still required, and the regulatory battle continues. But the inter-satellite backbone - the long-haul network connecting continents through space - operates outside the spectrum framework entirely.

Laser links also offer advantages beyond regulatory freedom:

• Faster: Light in vacuum travels 47% faster than in fiber optic glass
• Higher bandwidth: 200 Gbps per link, with 1.2 Gbps already demonstrated for ground-to-space (NASA’s ILLUMA-T on the ISS)
• More secure: Laser beams are extremely difficult to intercept or jam compared to radio signals
• No atmospheric attenuation (in the space-to-space segment)

SpaceX is now developing laser connectivity between Starlink and third-party satellites, potentially creating a shared optical backbone in space. In September 2024, SpaceX demonstrated military-standard laser communications between satellites.

The implication: the most valuable part of the satellite network - the long-haul backbone - is migrating to a technology that does not require spectrum. The spectrum wars may ultimately determine who can talk to users on the ground, while laser links determine who controls the highways in the sky.

WRC-27: The Next Battleground

The next World Radiocommunication Conference will take place in Shanghai from October 11 to November 12, 2027. It will be the first WRC hosted in Asia - China won the hosting bid by a narrow margin of 25 out of 48 voting member states. The US opposed China’s bid.

Approximately 80% of WRC-27’s agenda items address space services - the highest proportion in the conference’s history. Key items include:

• V-band spectrum for satellite earth stations in motion (47.2-51.4 GHz)
• Gateway earth station access in 51.4-52.4 GHz
• Lunar surface communications spectrum allocation
• Space-to-space inter-orbit links - regulation of the very laser and radio links between satellites
• 70/80 GHz band sharing between satellite and terrestrial services

The fact that China - which has filed for 249,000 satellites and views Starlink as a military threat - is hosting the conference that will shape space spectrum policy for the next four years adds a geopolitical dimension that previous WRCs lacked.

A US Senate hearing in March 2026 on “US Leadership at the World Radiocommunication Conference 2027” focused specifically on the strategic challenges of a China-hosted WRC, with particular attention to spectrum policies that could advantage Chinese constellations at the expense of American ones.

Key dates ahead:

• April 2026: FCC votes to overhaul the EPFD framework, modernizing spectrum sharing rules
• July 2026: Amazon Leo FCC deadline - 50% of constellation must be deployed
• October 2027: WRC-27 in Shanghai - 80% of agenda items on space services
• 2029: Amazon Leo FCC deadline - 100% of constellation deployed

Who Controls the Spectrum Controls the Future

The invisible war for orbital spectrum is being fought across multiple fronts simultaneously:

At the ITU in Geneva, where filing priority determines who gets to operate and who must prove they will not interfere. A system designed for dozens of geostationary satellites is being asked to manage filings for a million non-geostationary ones.

At the FCC in Washington, where SpaceX and Amazon trade accusations and the commission rewrites interference rules that have governed satellite operations for decades. The April 2026 vote to replace the EPFD framework is the most significant US spectrum regulatory action in a generation.

In Beijing, where state enterprises file for hundreds of thousands of satellites that serve both commercial and military objectives, and PLA researchers publish papers on methods to disable foreign satellite constellations.

In orbit itself, where 10,168 Starlink satellites have established a physical presence that creates coordination burdens on every future entrant. SpaceX’s first-mover advantage is not just commercial - it is electromagnetic. Every satellite broadcasting on Ku-band or Ka-band frequencies is a fact on the ground (or rather, in the sky) that subsequent operators must design around.

And in the laboratories where laser optical links are being developed to bypass the radio spectrum framework entirely - potentially making the entire regulatory battle over radio frequencies a rearguard action in a war that has already moved to a different theater.

Spectrum is invisible. You cannot see it, touch it, or store it. Under international law, no one owns it. But whoever controls the practical ability to use it - through filings, through deployment, through regulatory capture, or through sheer scale of operation - controls the future of connectivity from space.

The first transatlantic telegraph cable used a single frequency. Today, the fight is over millions of frequencies across dozens of bands, from 10 GHz to 86 GHz and beyond - governed by a UN agency in Geneva, contested by the world’s largest corporations, and shaped by military strategies that view the electromagnetic spectrum as the next domain of warfare.

The spectrum is invisible. The war for it is not.