Space Race: The Technological Ambition and Geopolitical Impact of the Million-Satellite Constellation Plan
01/02/2026
In late January 2026, the U.S. Federal Communications Commission received an application document significant enough to reshape the scale of human space activities. SpaceX formally requested approval to launch up to one million satellites, constructing a solar-powered computing network called the Orbital Data Center in orbits ranging from 500 to 2,000 kilometers above Earth. This initiative aims to provide unprecedented computing power support for artificial intelligence applications serving billions of users worldwide. The eight-page document, submitted on Friday, January 30, not only outlines Elon Musk's grand vision of relocating data centers away from Earth but also directly addresses core issues such as the energy bottlenecks in the global AI race, competition for near-Earth orbital resources, and the legal boundaries of space commercialization.
Technical Vision: The Quantum Leap from Starlink to "Orbital Cloud"
SpaceX's orbital data center system, as described in the document, is essentially a computational power upgrade of its existing Starlink constellation. Currently, there are approximately 9,600 Starlink satellites in orbit, primarily providing broadband internet access. The new satellites in the plan will be equipped with high-performance computing units, interconnected via laser links to form a distributed space supercomputer. These satellites will be deployed within multiple narrow orbital layers at altitudes ranging from 500 km to 2000 km and inclinations from 30 degrees to sun-synchronous orbit, with each orbital layer having a thickness not exceeding 50 km.
From a technical perspective, this plan relies on several key assumptions. The first is energy acquisition efficiency. Orbital data centers will utilize nearly constant solar radiation for power supply, avoiding the reliance on the power grid by ground-based data centers and the current situation where up to 40% of energy is consumed by cooling systems. The near-absolute-zero background temperature in space provides a natural radiative cooling environment, eliminating the need for water consumption in thermal management. The second is a revolution in launch costs. SpaceX is betting that its Starship reusable rocket, currently under development, can reduce the cost per kilogram of payload to orbit to below $100. The document clearly states: fully reusable launch vehicles like Starship can deploy millions of tons of mass into orbit annually, enabling on-orbit processing capabilities to reach a scale and speed unmatched by ground-based construction.
The third is network architecture continuity. The orbital data center will connect with the existing Starlink constellation via optical links, utilizing the latter as the user data downlink. This design leverages already deployed infrastructure while avoiding potential atmospheric attenuation and bandwidth limitations associated with direct data transmission from the orbital data center to the ground. However, the document lacks details on key engineering parameters such as satellite mass, computing unit specifications, and thermal control system design. It only mentions that different versions of satellite hardware will be designed and operated to optimize cross-orbit layer operations. Satellite industry analyst Tim Farrar pointed out to PCMag that this application document appears quite rushed and is at a very early stage.
Business Logic: The Capital Narrative of Mergers and the Arms Race
SpaceX chose this moment to submit its application for millions of satellites, closely intertwined with its capital market maneuvers. Multiple sources confirm that SpaceX is preparing for an initial public offering, expected to raise up to 50 billion dollars. At the same time, Bloomberg reported that SpaceX is exploring a potential merger with Musk's artificial intelligence company xAI, with another consideration being integration with Tesla. These moves collectively point to a strategic goal: building a sustainable funding cycle for the capital-intensive project of orbital data centers.
Analysts point out that SpaceX's existing Starlink and Starship projects have difficulty absorbing the massive funds that an IPO could potentially raise. Farrar analyzed: SpaceX cannot allocate the entire $50 billion expected to be raised from the IPO to its existing Starlink and Starship projects, while xAI urgently needs as much funding as possible to keep up with competitors. Combining SpaceX's satellite manufacturing and launch capabilities with xAI's AI model development needs can create a vertically integrated advantage from hardware to services. Musk himself bluntly revealed the business logic in his speech at the World Economic Forum in Davos: Building AI data centers in space is self-evident. The lowest-cost location for AI will be in space, and this will become a reality within two years, or three at most.
The deeper reason is that the global AI computing power competition is facing physical limits. Training large language models at the level of GPT-4 requires tens of thousands of NVIDIA H100 GPUs to run continuously for several months. The power consumption of a single ultra-large-scale data center can reach hundreds of megawatts, equivalent to the electricity consumption of a medium-sized city. Tech giants like Microsoft and Google are scrambling to purchase power resources across the United States, leading to a sharp increase in grid pressure. Orbital data centers can theoretically expand infinitely without being constrained by ground grid capacity or land resources, providing a seemingly perfect solution for the continuous scaling of AI models. In its application documents, SpaceX emphasized in bold font: **Orbital data centers are the most effective way to meet the accelerating demand for AI computing power.**
Regulatory Challenges: Congestion in Low Earth Orbit and the Legitimacy of "Kardashev Civilizations"
SpaceX's million-satellite plan immediately places regulatory bodies under unprecedented pressure. The Federal Communications Commission (FCC) has just approved SpaceX's application to launch an additional 7,500 second-generation Starlink satellites in early January 2026, but rejected the initial request for a total of 22,488 satellites. Currently, the total number of all artificial satellites in Earth's orbit is approximately 15,000. The number requested by SpaceX alone would reach nearly 70 times the existing global total.
The risk of orbital debris is an issue that regulatory bodies cannot avoid. According to data from the European Space Agency, the number of trackable space debris currently exceeds 36,500, and fragments larger than 1 centimeter may number over 1 million. Even though SpaceX claims that satellites will be deployed in widely separated orbital layers with sufficient space to avoid conflicts with other systems with similar ambitions, a satellite constellation on the scale of millions will inevitably significantly increase the probability of collisions. Each collision could generate thousands of new debris pieces, triggering a chain reaction known as the Kessler syndrome, where debris collisions produce more debris, ultimately rendering certain orbits unusable.
Spectrum allocation and interference with astronomical observations are two other major points of contention. In 2024, astronomers collectively protested that Starlink satellites' radio waves were blinding radio telescopes, severely hindering deep-space observations. A constellation of millions of satellites will further exacerbate this interference, affecting not only ground-based optical and radio astronomy but also potentially impacting the work of space telescopes such as the James Webb. When responding to skepticism on the X platform, Musk wrote: Satellites will actually be so far apart that it would be difficult to see one from another. The vastness of space is beyond comprehension. However, this philosophical response has done little to alleviate the concerns of the scientific community.
SpaceX introduced a compelling grand narrative in its application, stating that this is the first step toward a Kardashev Type II civilization—one capable of harnessing the entire energy output of its star. This reference draws from the civilization ranking theory proposed by Soviet astronomer Nikolai Kardashev in 1964, which uses energy consumption as a measure of a civilization's advancement. A Type II civilization can utilize the total energy output of its star. Such a narrative elevates a commercial project to the level of human civilization progress, creating a certain moral pressure for regulatory approval. However, regulatory agencies must make decisions based on existing legal frameworks, such as the International Telecommunication Union's spectrum allocation rules, the U.S. National Environmental Policy Act assessments, and the Convention on Registration of Objects Launched into Outer Space, rather than philosophical concepts.
Geopolitics: The New Frontier of Space Resource Competition and Hegemony
Once SpaceX's orbital data center plan is implemented, it will reshape the geopolitical landscape of the space economy and the AI industry. Currently, commercial activities in low Earth orbit are predominantly led by American companies, with Starlink already holding a first-mover advantage. If the million-satellite constellation is approved, SpaceX will effectively control most of the premium orbital and spectrum resources in low Earth orbit, creating an insurmountable monopoly barrier.
This prospect has already triggered international responses. Amazon's Project Kuiper is applying to the FCC for an extension of the deadline to deploy over 1,600 satellites, citing a lack of rockets, directly competing with SpaceX. Spacefaring nations such as China, the European Union, and Russia are also planning their own mega-constellations, but most are on the scale of tens of thousands of satellites, which is not on the same level as the million-scale. From a strategic perspective, controlling orbital data centers means controlling the high ground of future AI computing power, which can be used for both commercial applications and potentially transformed into military and intelligence advantages.
Environmental justice issues are also beginning to emerge. Orbital data centers are promoted as a green alternative, but rocket launches themselves produce significant carbon emissions and black carbon particles, the latter of which may have a disproportionate impact on the climate in the stratosphere. Starship uses methane as fuel, which burns cleaner than kerosene, but the cumulative impact of millions of launches has not been fully studied. Additionally, when solar-powered satellites burn up upon re-entry into the atmosphere at the end of their lifespan, they disperse metal particles such as aluminum and lithium into the stratosphere, potentially causing unknown effects on the ozone layer.
Space security experts are concerned about dual-use risks. The powerful computing capabilities of orbital data centers could theoretically be used for military purposes such as real-time processing of remote sensing data, commanding drone swarms, and decrypting encrypted communications. Although SpaceX claims to serve billions of users worldwide, during periods of geopolitical tension, the U.S. government may invoke the International Traffic in Arms Regulations or the Export Administration Regulations to impose restrictions on its services, similar to how it regulated the use of Starlink during the conflict in Ukraine.
The actual timeline for the orbital data center remains unclear. The document does not specify a concrete deployment date but mentions that preliminary operations began in early January 2026. Musk's prediction at Davos suggests that space will become the lowest-cost location for AI computing within two to three years. This timeframe is closely linked to the development progress of Starship, the FCC's approval cycle, and the timing window of capital markets. Considering the FCC's significant reductions in previous Starlink applications, the final number of approved million satellites may be drastically reduced. However, even if only 10% are approved, a scale of 100,000 satellites would be sufficient to alter the ecology of low Earth orbit.
Humanity stands at a historical juncture where space industrialization converges with artificial intelligence. SpaceX's million-satellite plan acts like a prism, refracting a multifaceted spectrum of technological ambition, capital logic, regulatory dilemmas, and geopolitical competition. When data centers leave the ground and ascend into orbit, they take away not only computing power and heat but also age-old questions about resource allocation, environmental responsibility, and power balance. These issues will not vanish upon entering space; they will only become sharper and clearer in the vacuum. The fate of orbital data centers will ultimately depend on whether humanity can establish wiser rules beyond Earth than those we have created on it.
Reference materials
https://www.bbc.co.uk/news/articles/cyv5l24mrjmo
https://gigazine.net/news/20260201-spacex-1-million-solar-powered-satellite-data-centers/
https://www.mathrubhumi.com/technology/news/spacex-orbital-ai-data-centers-bzm99ws6