When Bitcoin miners take off into space

null

Multiple media outlets report that Elon Musk’s space exploration technology company, SpaceX, will soon file its IPO prospectus with the U.S. Securities and Exchange Commission (SEC). The target valuation is $1.75 trillion, and the company is expected to raise more than $75 billion. If it goes through, it would be the largest IPO in human history, far surpassing the $29.4 billion record set by Saudi Aramco in 2019. It would also be the most closely watched IPO of the year.

What’s intriguing is that in February 2026, SpaceX suddenly acquired Musk’s other AI company, xAI, and wrote the “orbital data center” into its core strategy: using the vacuum environment of space for heat dissipation and powering continuously with solar energy to put AI compute power into low Earth orbit. Musk believes that, in the long run, AI in space is the only approach that can achieve scalable development.

At the same time, Nvidia is also actively building in this direction. It invested in the orbital data center startup Starcloud, which in November 2025 successfully sent an Nvidia H100 GPU into orbit and completed what is believed to be the first-ever in-space AI model training and inference on a human history record.

As SpaceX brings AI compute power up to the sky, many people have also started wondering: since it relies on compute chips like Bitcoin mining and can also use solar energy, could Bitcoin mining be moved to space as well? But this question is, in reality, far more complex than most people imagine.

A satellite, a solar panel, a mining rig

Mining is a competitive kind of mathematical computation. Millions of mining rigs around the world run simultaneously, competing to be the one that solves a specific hash value the fastest. The winner receives the Bitcoin block reward for the current block. This process is called Proof of Work. Its cost is a large amount of electricity. The global Bitcoin network’s continuous power consumption is about 20 gigawatts—equivalent to the total industrial electricity consumption of a medium-sized country. Most of miners’ profit margins are determined entirely by electricity prices; once electricity prices rise, profit margins get squeezed.

And in space, an endless supply of sunshine perfectly corresponds to the most core cost variable in Bitcoin mining: power.

In Earth orbit, the solar radiation intensity is about 1,380 watts per square meter—about six times the ground average level—and it is not affected by clouds, day-night cycles, or seasons. In certain sun-synchronous orbits, satellites can receive sunlight and generate power almost around the clock. Mounting a mining rig on the back of a solar panel, sending it into orbit so it can mine forever—this is the underlying logic of space mining.

Bitcoin core developer Peter Todd published a technical analysis in December 2024, taking the idea from concept to engineering blueprint. He proposed the concept of a “flat mining rig”: install ASIC chips directly on the back of a solar panel, with the front facing the sun to generate electricity, while the chips on the back consume power for mining. The overall structure radiates waste heat in both directions.

Space heat dissipation is a counterintuitive challenge. On Earth, chip heat can be removed via air convection; but in the vacuum of space, there is no air—heat can only be emitted through radiation. Todd’s calculations indicate that without adding any extra heat-dissipation equipment, the thermal equilibrium temperature of this structure in orbit is about 59°C, well within the normal operating range of the chips. If you think the temperature is too high, you can simply tilt the entire panel slightly relative to the sun to reduce the illuminated area, which would further improve the heat-dissipation issue.

In terms of communications, it’s also surprisingly simple. Communication between miners and mining pools is essentially receiving new block headers and submitting computation results. The amount of data generated each day is about 10MB—less than the traffic consumed by streaming a single song. Communication latency in low Earth orbit (500 to 1,000 kilometers above Earth) ranges from 4 to 30 milliseconds. This leads to a probability of wasted blocks—i.e., blocks whose submitted computation results are already outdated—of less than 0.01%, which is on the same order of magnitude as the vast majority of on-ground miners, with no material difference. In fact, Blockstream began broadcasting the complete Bitcoin blockchain to the world using Earth-synchronous satellites as early as 2017, proving that combining satellites with a blockchain has never been an unsolved problem.

If it’s physically feasible and the engineering framework is feasible, why hasn’t it been widely adopted? The reason is that the cost of rocket launches is too high.

Can’t make the economics work

Using Space X’s Falcon 9 to send cargo to low Earth orbit costs about $2,720 per kilogram at current rates.

Peter Todd estimates that a complete 20-kilowatt space mining system—including solar panels, thermal radiators, an ASIC chip array, structural support components, and a communications module—has a total weight of roughly 1,600 to 2,200 kilograms. At current prices, the cost to launch it just once is as high as $4.3 million to $6 million.

How much compute power can this system deliver each day, and how many coins can it mine? Researcher Nick Moran provided the answer: daily returns of about $92.7, or about $34,000 annually. The payback period is over 100 years.

Starcloud CEO Philip Johnston calculated that launch costs must drop to below $200 per kilogram for space mining to have basic commercial logic. That means costs still need to fall by 13 times.

SpaceX’s Starship is widely seen as the key to making that leap. A fully reusable Starship could, in theory, reduce the launch cost per kilogram to below $100 and even lower. That’s also one of the premise assumptions behind SpaceX’s orbital data center vision in this IPO. But when—or whether—that cost curve will actually be realized remains an unresolved variable.

Another challenge is Bitcoin’s automatic adjustment of the network-wide mining difficulty. The Bitcoin protocol counts the total network hash rate every two weeks and automatically adjusts the mining difficulty so that the block production speed stays at about 10 minutes per block. In other words, if a large number of space mining rigs flood the market and the network-wide compute power increases significantly, the mining difficulty will be raised accordingly, and all miners—including those in orbit—will have their profit margins compressed in sync.

There’s always someone busy hunting for treasure in this world

Even so, a group of startups is still working to push this forward.

Starcloud, formerly known as Lumen Orbit, is currently the company closest to practical deployment and also one of the most important observation samples in the entire space mining track. Founded in 2024, it is headquartered in Raymond, Washington. Backers include the angel funds of NFX, Y Combinator, a16z, and Sequoia Capital, as well as Nvidia. Total funding is about $200 million. Its CTO previously worked for 10 years in Airbus Defense and Aerospace, and the chief engineer previously worked at SpaceX on the Starlink project.

In November 2025, Starcloud successfully sent its first satellite carrying an Nvidia H100 GPU into orbit, running Google’s Gemma language model in space and sending to Earth the first piece of information in human history generated in orbit by AI. In March 2026, Starcloud announced that its second satellite would carry both Bitcoin ASIC chips and Nvidia’s latest-generation Blackwell GPU, aiming to become the first organization in human history to mine Bitcoin in space. In addition, the company has applied to the U.S. Federal Communications Commission (FCC) for a constellation plan to deploy up to 88,000 satellites, with a long-term vision to build a total 5-gigawatt compute infrastructure in orbit.

SpaceChain is the OG player in this track, co-founded by former Bitcoin core developer Jeff Garzik and Zheng Zhong. Since 2017, SpaceChain has launched at least seven blockchain payloads to satellites and the International Space Station. In June 2020, Garzik completed the first-ever space Bitcoin transfer on an orbit about 400 kilometers above Earth, amounting to 0.0099 BTC, using a multi-signature wallet node installed on the SpaceChain aboard the space station. SpaceChain’s core focus is not active mining, but orbital security nodes for blockchain transactions: locking private keys in space so that no hacker or government on Earth can physically access them.

Cryptosat was founded by two Stanford PhDs and currently operates three satellites in orbit, mainly providing anti-tampering orbital cryptography services. In 2023, Cryptosat participated in the largest trusted setup ceremony (KZG Ceremony) in Ethereum’s history. By generating some random parameter values through orbital nodes, it ensured at the governance level that none of these parameters could be controlled by any single ground-based institution. What it explored is another possibility for space-based blockchains: not mining, but making the entire crypto economic system harder to attack.

From orbit to the market: what does this mean for mining?

For currently operating Bitcoin mining companies, space mining may not pose a practical competitive threat in the short term. However, many startups are still continuously trying, which shows that the cost-reduction potential represented by it—and the associated attraction and imagination space for the industry—remains very large. This also reflects that the entire industry is facing structural cost pressure.

After the 2024 halving, the network-wide hash rate and difficulty have continued to hit all-time highs. Energy costs account for 70% to 90% of total operating costs. In such a context, whoever can reliably secure clean power at the lowest cost has the deepest moat. Hydropower, wind power, and associated natural gas resources in the United States, the Middle East, and Africa are becoming the core drivers behind the next round of mining M&A and site selection.

The logic of space mining is the ultimate extrapolation of the above trend: if cheap electricity on the ground inevitably gets squeezed due to competition for demand, then go where energy is most abundant—namely, the cosmos.

Of course, even if Starcloud-2 in 2026 mines if it can mine the first Bitcoin, for the total network compute power of over 900 exahashes per second (EH/s) worldwide, it would be roughly like a grain of sand falling into the sea. But the symbolic value itself has real staying power. Just like the 0.0099 BTC space transfer in 2020, its value isn’t in the amount—it’s in proving that this is something that can be made to work.

From SpaceX’s IPO narrative to Nvidia’s orbital compute power layout, and then to Starcloud’s ASIC satellite plan, a shape is emerging: space is becoming a competitive arena for the next generation of compute infrastructure. AI compute power is launching first, and Bitcoin compute power follows closely behind.

Someday, the global digital network described in Satoshi Nakamoto’s white paper—connecting every corner of Earth—could also break beyond Earth and float in the universe, searching for new opportunities.