We assumed the final frontier would be the ocean floor or the Martian desert. Instead, it is the silent, cold void above our heads—and the crypto industry is once again looking up with starry eyes. This week, news broke that SpaceX and Blue Origin have applied to build satellite constellations designed as orbital AI data centers. The author of a Crypto Briefing piece posits that this could revolutionize crypto mining by providing limitless solar energy and low-latency computation. But as a governance architect who has watched the industry chase one infrastructure mirage after another, I see a different picture: a kingdom of ghosts in the machine.
The announcement itself is thin. Both companies filed for FCC permits to launch thousands of satellites, each equipped with AI compute nodes. No technical specs. No timeline. No budget. The narrative is seductive: take mining rigs off Earth, escape energy costs, bypass regulation. Yet the gap between a filing and a functional orbital data center is measured in decades, not quarters. In 2017, I spent six months reading Tezos and Cardano whitepapers, convinced that self-amending governance would fix everything. I learned that a beautiful narrative without a working prototype is just a story. This is no different.
The technical barriers are brutal. Start with latency. A satellite in low Earth orbit has a round-trip time of roughly 10 milliseconds to the ground. That is fine for streaming video, but terrible for mining pools that rely on microsecond synchronization of block headers. Even with laser inter-satellite links, the average latency for a transaction to traverse a satellite network and return to Earth is an order of magnitude higher than a terrestrial fiber line. Mining is a race—every millisecond of delay reduces your chance of finding a block. I once modeled a DePIN network using orbital nodes for a client. The simulation showed that for Bitcoin mining, a 10ms latency disadvantage would make the miner unprofitable unless it had 50% lower energy costs. But energy in space is not free.
A solar panel on Earth generates roughly 200 watts per square meter at peak. In space, you get about 1350 watts per square meter—but that is before accounting for eclipse periods, panel degradation from radiation, and the mass of the panels themselves. To power a single Antminer S19 (3,250 watts), you would need roughly 2.5 square meters of solar panels. That sounds manageable, until you add the mass: a standard space-grade solar panel weighs about 1 kg per square meter. So 2.5 kg for the panels, plus the miner itself (13 kg), plus the satellite bus, thermal management, shielding, and radiation-hardened electronics. Total mass per miner: easily 50-100 kg. Launch costs to LEO are currently around $5,000 per kg. That is $250,000 to $500,000 just to get one mining rig into orbit. The rig itself costs $3,000. The economics are absurd.
Thermal management is another nightmare. In a vacuum, you cannot use fans or liquid cooling that relies on convection. You must radiate heat away, which requires heavy radiators. A single ASIC miner generates 3.25 kW of heat—enough to cook the electronics if not dissipated. Spacecraft engineers spend millions on thermal systems for delicate instruments. A mining rig is not delicate; it is a toaster with silicon. The cooling mass alone would double or triple the satellite weight. And once the rig is in orbit, it cannot be repaired. A single failure—a short, a power surge, a micrometeorite hit—and the entire asset is lost. Terrestrial mining farms have uptimes of 99% because technicians are a call away. In space, failures are permanent.
During the 2020 DeFi Summer, I audited Curve's governance mechanics and discovered that voting power concentrated among whales despite the elegant quadratic formula. The system looked perfect on paper but failed in practice. The same holds for orbital mining. The regulatory hurdles alone make it a non-starter. The FCC requires detailed spectrum coordination to avoid interfering with existing satellites and ground stations. The International Telecommunication Union (ITU) allocates orbital slots. SpaceX and Blue Origin are powerful lobbyists, but getting approval for thousands of satellites carrying high-power AI compute nodes will take years, and each node could interfere with radio astronomy or communications. The cost of compliance is enormous.
But the real issue is that the narrative misunderstands what crypto mining needs. Mining is not about raw computing power; it is about cheap, reliable, and low-latency access to the network. The most successful mining farms are located in hydro-rich regions like Sichuan or Texas, where electricity costs are below $0.03/kWh. Putting a miner in orbit increases energy costs by orders of magnitude when you factor in launch, maintenance, and the inherent inefficiencies of space power systems. The author claims it will “impact crypto mining” by providing cheap energy—but that is backwards. Energy is not free in space; it is the most expensive energy humanity has ever produced.
Let me offer a contrarian angle: I might be wrong about the timing. The gravitational pull of this narrative is strong, and it could attract significant venture capital. If SpaceX or Blue Origin actually launch a prototype and demonstrate that orbital AI inference can work for specific use cases—like high-frequency trading or satellite image processing—the hype could spill into crypto. Mining is a stretch, but the idea of “space-based DePIN” could inspire a new wave of tokenized satellite networks. I have seen this pattern before. After my 2022 burnout, I wrote a private journal titled “The Ethics of Ruin,” analyzing how the industry repeatedly resurrects failed concepts with new branding. The space narrative is the ultimate resurrection: it takes the DePIN thesis and launches it beyond Earth’s atmosphere. But resurrection does not change the underlying corpse.
Moreover, the governance model is entirely centralized. SpaceX and Blue Origin are private companies controlled by billionaires. There is no DAO, no token vote, no community oversight. The satellites will belong to them. If they decide to redirect compute to military contracts or shut down the network, miners have no recourse. This is the opposite of the decentralized ethos that crypto claims to champion. Silence is the only consensus that never forks. When the operator is a single entity, the network is not trustless—it is trust in Elon Musk or Jeff Bezos.
What would need to happen for orbital mining to become viable? First, launch costs must drop to below $500 per kg—a 10x reduction from today. That might happen with fully reusable rockets and mass production, but it is not imminent. Second, solar panel efficiency must double and radiation-hardened electronics must become commodity items. Third, on-orbit servicing must be automated so that failed miners can be swapped out by drones. All of these are possible in a 30-year horizon, but crypto mining exists in a 3-year cycle. To govern the future, we must debug the present. The present is that the most efficient mining is happening in Texas and Kazakhstan, not in orbit.
My takeaway is melancholic but honest: this news is a ghost story. It tells us about our desires—for limitless energy, for escape from regulation, for a clean slate—but not about reality. The crypto industry has a habit of falling in love with infrastructure that does not exist. We built a kingdom of ghosts in the machine—layer-2 solutions that never scale, stablecoins that depeg, DAOs that centralize. The orbital AI data center is the newest phantom. Instead of chasing it, we should invest in improving the physical infrastructure we already have: better ASICs, more efficient cooling, and renewable energy partnerships on Earth. The code is law, but the humans are the bug. We must not let the fantasy of orbital mining distract us from building sustainable ground-level networks.
In the void, we found our own gravity—and it pulled us back down to Earth. That is where the real work lies.