Nvidia invests in "space mining Bitcoin," and the startup Starcloud plans to send ASIC miners into orbit within the year

Supported by Nvidia, the space data center startup Starcloud announced that this year it will install Bitcoin ASIC miners on its second spacecraft, aiming to become the first company to mine cryptocurrencies in space. CEO Philip Johnston believes that Earth-based mining will eventually be replaced—however, there are still several unclear aspects, from satellite launch costs to orbital latency.

(Background: This man wants to send Bitcoin miners into space: unlimited sunlight + zero cooling costs make it a mining paradise)

(Additional context: The strategic game of Bitcoin mining companies: continue holding coins or switch to AI?)

Table of Contents

Toggle

• Cost gap between ASIC and GPU
• Proven results, but still small scale
• “Mining on Earth makes no sense”—is this judgment valid?
• Market background

Last Saturday, Starcloud CEO Philip Johnston stated on X that the company plans to connect Bitcoin ASIC miners on its second spacecraft, scheduled for launch later this year. He claims this will be the “first Bitcoin mined in space.”

Previously, he discussed economic reasons in an interview with HyperChange.

Cost Gap Between ASIC and GPU

Johnston pointed out that GPUs cost about 30 times more per watt than ASICs. A 1-kilowatt B200 chip costs around $30,000, while an equivalent ASIC costs about $1,000.

This figure itself is not problematic. It only reflects hardware procurement costs and does not include launch, deployment, maintenance, and communication expenses. Sending one kilogram of payload to low Earth orbit still costs thousands of dollars, and in space, additional engineering is needed for heat dissipation, power management, and radiation resistance.

Comparing ground hardware prices without accounting for the incremental costs of space deployment gives an incomplete picture.

Proven Results, but Still Small Scale

Starcloud is not just talk. Founded in early 2024, the company successfully launched a satellite equipped with NVIDIA H100 GPUs into orbit in November 2025. This is the first time such high-level GPUs have operated in space. Reports indicate the satellite has completed AI model training and chatbot testing.

The company has applied to the U.S. Federal Communications Commission (FCC) to deploy approximately 88,000 satellites powered mainly by solar energy.

Running an H100 on one satellite is an engineering milestone. But moving from one satellite to 88,000 involves more than time—it requires funding, regulatory approval, and supply chain capabilities.

“Mining on Earth Makes No Sense”—Is This Judgment Valid?

Johnston’s core argument is that Bitcoin mining consumes about 20 GW of electricity, competing with residential and industrial energy needs, so “mining on Earth is pointless; everything will eventually be done in space.”

Space does have energy advantages. Certain orbits provide near-continuous solar power, the vacuum environment offers natural cooling, and there are no land costs.

However, this logic overlooks several assumptions. First, the durability of ASIC miners under cosmic radiation and extreme temperature variations has not been proven over the long term. Second, Bitcoin mining requires low-latency connections to mining pools; in low Earth orbit, communication latency is about 20-40 milliseconds—acceptable but network stability is another concern. Third, there is no public data showing that the total cost per MH/s in orbit (including launch amortization) is lower than on the ground.

Previously, researchers Jose E. Puente and Carlos Puente argued that Bitcoin could be transmitted to Mars via NASA or Starlink optical links in as little as three minutes. While theoretically feasible, mining on Mars remains impractical due to planetary latency. Low Earth orbit conditions are much better, but moving from theory to commercial reality involves many more steps.

Market Background

As of press time, Bitcoin’s price has fallen nearly 48% from its peak of $126,080 on October 6. Mining difficulty has decreased from a record 155.9 trillion in November to 145 trillion, a drop of about 7%.

On the ground, mining companies face dual pressures: high electricity costs and the shift toward AI computing power. Space mining, as a concept, has its appeal. But until the second satellite completes actual mining tests, it remains just a plan.