Elon Musk has once again sparked debate in the tech world with a bold claim about the future of artificial intelligence. Speaking on the Dwarkesh podcast, the Tesla and SpaceX CEO said that within the next 36 months, space may become the cheapest place to run AI systems, not Earth.
According to Elon Musk, the biggest challenge facing AI today is energy. As AI models grow larger and more powerful, they need huge amounts of electricity. Running these systems on Earth is becoming increasingly expensive due to rising power costs, land limitations, and environmental concerns.
Why Musk Thinks Space Makes More Sense?
Elon Musk believes space offers a major advantage which is uninterrupted solar energy. Solar panels in space are not affected by clouds, pollution, or nightfall. This means they can produce energy continuously, without needing batteries. Elon Musk claims that space-based solar panels can generate around five times more energy than those installed on Earth.
Because of this constant energy supply, Elon Musk argues that data centres in orbit could eventually be cheaper to operate than those on the ground. In his view, once the infrastructure is set up, space could solve many of the energy problems AI companies are currently facing.
Combining AI and Space Technology
This idea gains more sense following SpaceX’s recent acquisition of Elon Musk’s AI company, xAI. The move suggests Musk is serious about combining artificial intelligence with space technology. Analysts say this could help him explore new ways of running large AI systems without relying heavily on Earth-based resources.
Elon Musk also addressed concerns about hardware reliability in space. He said that once GPUs and other equipment pass initial testing, they should continue working for long periods without major issues though experts remain divided on this point.
Industry Leaders Remain Skeptical
Not everyone agrees with Elon Musk’s vision. Several tech leaders have questioned whether space-based data centres are realistic in the near future. Critics point out that launching equipment into space is still extremely expensive, and maintaining or repairing systems in orbit is far more difficult than on Earth.
Some cloud computing executives have dismissed the idea, saying that current technology and costs make space-based AI infrastructure impractical for now.
Conclusion
Elon Musk’s prediction may sound futuristic, but it highlights a growing problem Earth’s limits when it comes to supporting AI’s rapid growth. Whether space truly becomes the cheapest home for AI in the next three years remains uncertain. Still, the idea has already started an important conversation about how and where the future of artificial intelligence will be built.