Author: Nancy, PANews
The next battlefield for AI computing power is extending into space, gradually becoming a new direction for commercial narratives.
After successfully launching the first space AI server, a space computing startup recently announced its plan to send Bitcoin mining into outer space.
Planning to Mine Bitcoin in Space This Year, Symbolic Significance May Outweigh Practical Value
Having moved past the first half of the competition focused on chips and models, the battle for AI computing power is quietly shifting toward the争夺 for energy. Electricity, as a core resource in this competition, is rapidly becoming a scarce resource in the global race for computing power. This shift is not only changing the industry landscape but also directly reshaping the cost structure of the Bitcoin mining industry.
In particular, once steady Bitcoin miners are now turning to the AI computing power track. Behind this transformation lies the survival pressure brought by Bitcoin halving, the profit compression caused by intensified competition and rising energy costs, and the huge opportunity presented by the AI narrative.
As Bitcoin mining efficiency is squeezed by the global energy争夺, Starcloud has proposed a bold plan to move Bitcoin mining into space.
In a recent interview with HyperChange, Starcloud CEO Philip Johnston revealed that the company is currently focused on its existing space computing power operations, but also has plans for Bitcoin mining. Starcloud will equip its Starcloud-2 satellite, scheduled for launch later in 2026, with some ASIC hardware specifically designed for Bitcoin mining. If successful, Starcloud would become the world's first spacecraft to mine Bitcoin in space.
Johnston believes that space offers multiple natural advantages over Earth. Firstly, space has an unlimited and continuous supply of solar energy, which is more stable and cheaper than renewable energy on Earth. At the same time, the environment in space is superior; although temperature differences and radiation are extreme, they can significantly reduce the heat dissipation energy consumption of hardware, lowering cooling costs and equipment maintenance burdens. Most importantly, Bitcoin mining in space can avoid the increasingly tense energy bottlenecks, grid limitations, and regulatory pressures on Earth. Currently, about 20GW of electricity is used for Bitcoin mining on Earth, and such large-scale power consumption is no longer feasible on the ground. In space, utilizing solar energy for cheap power offers a brand-new solution for Bitcoin mining.
Johnston further added that the cost of Bitcoin mining equipment typically ranges from $600 to several thousand dollars, far less than NVIDIA's enterprise-grade GPUs (often over $30,000). This makes the economics of space-based Bitcoin mining highly attractive.
Starcloud views Bitcoin mining as a "future business," leveraging solar energy in space for cheap power, and直言 this is also one of the reasons why it and other companies (including SpaceX) are building data centers in space. Space mining can not only significantly reduce costs but also provide a全新的 resource acquisition model for the global computing power market.
The concept of space mining is not new. Last year, Intercosmic Energy also stated it was researching Bitcoin mining in space.
However, mining Bitcoin in space still faces many challenges. Johnston also admitted that the economics of space-based Bitcoin mining remain unstable. Currently, Bitcoin ASIC devices can run on any cheap energy source, and with the continuous introduction of new equipment, the profitability of mining hardware can decline rapidly.
Furthermore, although launch costs are decreasing year by year, sending hardware into space remains an expensive endeavor. Compared to ground-based mining farms, the startup and maintenance costs of space mining are still high, including expenses for launch, spacecraft integration, satellite communications, equipment upgrades, etc.
More棘手的是, the space environment places extremely苛刻 demands on hardware. Bitcoin mining ASIC devices need to operate stably under极限 conditions such as high radiation and extreme temperature changes, posing severe tests for设备 performance and lifespan.设备 maintenance and upgrades will also become a major challenge, as the cost and difficulty of repairing and replacing hardware will increase significantly if a故障 occurs.
Previously, many crypto organizations have explored bringing blockchain business into space. For example, the veteran Bitcoin community company Blockstream has, since 2017, leased multiple geostationary orbit satellites to broadcast Bitcoin blockchain data globally for free. Even if there is a large-scale internet outage on Earth (such as natural disasters or人为封锁), as long as you have a small satellite dish (receiver), you can synchronize the Bitcoin ledger and complete transactions. SpaceChain installed the first commercial Ethereum node on the International Space Station (ISS) back in 2019. Earlier this year, a new project focused on space commerce, Spacecoin, also attracted market attention, aiming to use a satellite network for cryptocurrency payment settlements.
Therefore, the投入 for space mining in the short term may far exceed the收益. Currently, it holds more symbolic significance, or perhaps serves as a narrative手段 for this startup to attract market attention.
The First Time in Human History, Sending an NVIDIA AI Server into Space
Founded in 2024, Starcloud, formerly known as Lumen Orbit, has already emerged prominently in the global tech circle as one of the first companies to propose building data centers in space.
As a member of the NVIDIA Accelerator Program, and also part of the Y Combinator and Google Cloud incubator programs, Starcloud is not simply moving data centers into space. Its goal is to utilize the unique resources of the space environment to build infrastructure capable of supporting AI computing and large-scale computation.
So far, Starcloud has raised at least $21 million in funding, with backing from知名 investment firms like NFX, Y Combinator, FUSE, Soma Capital, a16z, and Sequoia Capital.
Starcloud has already secured a place in the space AI computing power race. Last November, Starcloud completed the first-ever large model training in space orbit. It launched its Starcloud-1 satellite via a SpaceX Falcon 9 rocket, sending an NVIDIA H100 GPU into Earth's orbit, and successfully ran Google's open-source AI large language model Gemma, sending the first message from space to Earth: "Hello, people of Earth!"
At the time, Philip Johnston stated that space AI is not a gimmick; the company's goal is to achieve orbital data center energy costs that are 10 times lower than those of ground-based data centers.
After achieving initial success, Starcloud's ambitions did not stop. Recently, the company applied to the FCC for permission to deploy a massive constellation of 88,000 satellites to build a distributed, space-based AI training and cloud computing platform. But turning this vision into reality is fraught with huge challenges. From funding scale, regulatory approval, launch capacity, orbital resource allocation, to operational sustainability, this is not just a commercial race but also a systematic engineering challenge, with every step full of uncertainty and complexity.
It's not just Starcloud. As the AI industry's demand for computing resources increases, more and more tech companies are seeking new sources of computing power, and space is gradually becoming the focus of this competition. For example, Google officially launched its Sun catcher project late last year, announcing plans to send its self-developed TPU AI chips into space, aiming to build a prototype solar-powered space data center; not long ago, Musk's SpaceX applied to deploy 1 million satellites in Earth orbit to build orbital data centers; recently, data storage and disaster recovery company Lonestar and semiconductor and storage company Phiso launched a set of data center infrastructure via a SpaceX rocket destined for the moon.
As the concept of space data centers moves from science fiction to reality, a new infrastructure arms race is unfolding. According to Musk's predictions, in five years, annual新增 AI computing power in space will reach hundreds of gigawatts; the annual amount of AI computing power sent into space for operation will exceed the historical cumulative total of all AI on Earth.
By then, the main battlefield of AI computing power competition will truly have moved to space. In the coming years, we will see more commercial exploration and technological innovation. Space mining might just be one of the stepping stones in this wave.
