Space could become cheapest place to run AI within 36 months, says Elon Musk

Elon Musk has made a striking prediction about the future of artificial intelligence, suggesting that outer space could soon become the most cost-effective location to run AI systems, potentially within the next 36 months. The comments, reported by The Times of India, have sparked renewed debate about the growing energy demands of AI and how the industry might meet them.

Musk shared his views during an appearance on the Dwarkesh Podcast, where he discussed the challenges of scaling artificial intelligence infrastructure on Earth. As AI models become larger and more complex, the demand for computing power and the electricity required to run it is rising sharply. Musk argued that Earth-based power generation may struggle to keep pace with this growth.

According to The Times of India, Musk believes space offers a clear advantage due to its access to uninterrupted solar energy. Solar panels in orbit are not affected by night cycles, weather conditions or atmospheric interference, allowing them to generate significantly more power than those on Earth. He estimated that solar energy collected in space could be up to five times more efficient, making space-based data centres increasingly attractive from a cost perspective.

Musk also pointed to limitations in global power production, noting that electricity generation has remained relatively flat in many parts of the world in recent years, even as demand from AI data centres continues to surge. Outside of China, he said, scaling energy production quickly enough to support the next phase of AI development could prove difficult. In contrast, space allows for continuous energy generation without the need for extensive battery storage, which could further reduce operational costs.

The remarks come at a time when Musk’s ventures in artificial intelligence and space exploration are becoming more closely linked. SpaceX recently integrated Musk’s AI startup, xAI, a move that has fueled speculation about long-term plans involving AI infrastructure beyond Earth. While Musk did not outline a detailed roadmap, his comments suggest a future where AI computing could expand beyond traditional, Earth-bound data centres.

However, Musk’s prediction has not gone unchallenged. Several technology leaders and cloud computing executives have expressed skepticism about the feasibility of large-scale space-based data centres in the near term, citing challenges such as launch costs, maintenance, and data transmission latency. The Times of India has previously reported that some industry experts believe terrestrial solutions, including renewable energy and more efficient hardware, may remain more practical in the short to medium term.

Despite the doubts, Musk’s comments have reignited discussion about the long-term sustainability of AI infrastructure. As demand for artificial intelligence continues to accelerate, questions around energy availability, cost, and environmental impact are becoming increasingly urgent — and Musk’s vision suggests that the answers may eventually lie beyond the planet.