Space is becoming the next frontier for artificial intelligence infrastructure, with solar panels in orbit generating five times more power than identical panels on Earth. Elon Musk predicted at the World Economic Forum in January 2026 that "the lowest cost place to put AI will be space and that will be true within two to three years, three at the latest." Just one week later, SpaceX announced its acquisition of xAI, signaling a major commitment to building vertically integrated space-based AI systems. Why Would Anyone Put Data Centers in Space? The math behind space-based AI is compelling. Traditional data centers consume enormous amounts of electricity and require intensive cooling systems. Musk explained the physics during his Davos interview: "When you have solar in space you get five times more effectiveness, maybe even more than that, than solar on the ground. It's always sunny, so you don't have a day-night cycle or seasonality or weather and you get about 30% more power in space because you don't have atmospheric attenuation of the power". Space also offers natural cooling advantages. "It's very cold in space. When you're in the shadow, it's very cold in space, 3 degrees Kelvin," Musk noted. This means operators can position solar panels facing the sun and radiators pointing away from it, creating an extremely efficient cooling system without any additional energy cost. The combination of superior power generation and passive cooling creates what Musk calls "a no brainer for building solar powered AI data centers in space." SpaceX's statement on the xAI acquisition emphasized that "current advances in AI are dependent on large terrestrial data centers, which require immense amounts of power and cooling. Global electricity demand for AI simply cannot be met with terrestrial solutions, even in the near term, without imposing hardship on communities and the environment". What's the Timeline for Making This Actually Happen? The critical bottleneck is launch cost. SpaceX's Starship rocket, currently in advanced testing, could be the game-changer. On March 18, 2026, SpaceX completed the first static fire test of Starship Version 3's Booster 19, successfully firing 10 of the booster's Raptor engines. Engineers plan to conduct a full 33-engine static fire test in the coming weeks, with an inaugural launch targeted for April 2026. If Starship achieves full reusability as planned in 2026, the cost implications would be dramatic. Musk stated that "the cost of access to space would drop by a factor of 100 when you achieve full reusability." This would bring launch costs below $100 per pound, making it cheaper than aircraft freight. The scale required is staggering. SpaceX's xAI statement outlined the vision: "launching a million tons per year of satellites generating 100 kW of compute power per ton would add 100 gigawatts of AI capacity annually, with no ongoing operational or maintenance needs. Ultimately, there is a path to launching 1 TW/year from Earth". For context, in 2025, the most prolific year in orbital launch history, only about 3,000 tons of payload reached orbit. Who Else Is Building High-Powered Satellites for Space Computing? SpaceX isn't alone in this vision. K2 Space, founded by former SpaceX engineers Karan and Neel Kunjur in 2022, is launching its first high-powered satellite called Gravitas as soon as the end of March 2026 aboard a SpaceX Falcon 9 rocket. The satellite weighs two metric tons and features a 40-meter wingspan when its solar panels unfold. Gravitas is designed to produce 20 kilowatts of electricity for payloads including powerful sensors, transceivers, and computers. For comparison, most spacecraft generate just a handful of kilowatts, while ViaSat-3 generates more than 25 kilowatts and Starlink V3 satellites generate 20 kilowatts. "The future is higher power," K2 CEO Karan Kunjur explained. K2 has raised $450 million and was valued at $3 billion by investors in December 2025. The Gravitas mission will carry 12 undisclosed payload modules from several customers, including the Department of Defense, plus a 20-kilowatt electric thruster that the company expects will be the most powerful ever flown in space. Kunjur is planning aggressively for scale. K2 intends to launch 11 satellites over the next two years in demonstration and commercial missions, with expectations to produce satellites for customers by 2028. The company has designs ready for a 100-kilowatt satellite already taped out on its factory floor. What Manufacturing Challenges Need to Solve? Building space-based AI infrastructure requires massive solar panel manufacturing capacity. Musk revealed that "the SpaceX and Tesla team, both separately, are working to build to 100 GW a year of solar power in the US, of manufactured solar power. That'll probably take us three years or something". Tesla's track record in solar manufacturing has been mixed. Previous plans to produce 1 gigawatt of panels at its Buffalo, New York facility acquired in 2016 did not materialize. However, in January 2026, Tesla's Energy team revealed plans to ramp up the Buffalo Gigafactory to 300 megawatts per year in an initial phase. Space-based solar panels will require different technology than terrestrial panels. The specific cell and panel technology SpaceX plans to use remains undisclosed, but scale will be critical. As production scales, economies of scale kick in and costs decline, just as they have for solar and battery storage on Earth. How to Prepare for the Space AI Economy - Monitor Launch Cost Reductions: Track Starship's progress toward full reusability and the resulting impact on launch pricing. When costs drop below $100 per pound, the economics of space-based infrastructure shift dramatically in favor of orbital deployment. - Follow High-Power Satellite Development: Companies like K2 Space are proving that 20 to 100-kilowatt satellites are feasible. Watch for successful Gravitas mission results and subsequent K2 launches to validate the business case for orbital compute. - Assess Solar Manufacturing Capacity: Tesla and SpaceX's plans to build 100 gigawatts per year of solar manufacturing capacity will be essential. Delays or successes in ramping Buffalo and other facilities will signal whether space-based AI timelines are realistic. Another venture focused on this opportunity is Starcloud Inc., based in Redmond, Washington. Their September 2024 white paper described how "orbital data centers can be scaled almost indefinitely without the physical or permitting constraints faced on Earth, using modularity to deploy them rapidly." The analysis showed that a 5-gigawatt AI data center in space would require a solar array measuring just 4 kilometers by 4 kilometers, far smaller than an equivalent terrestrial facility would need. The convergence of reusable rockets, high-powered satellites, and space-based solar technology is creating a genuine inflection point. If Starship achieves its reusability targets and launch costs plummet as promised, the economic case for moving AI infrastructure to orbit becomes compelling. The next 12 to 24 months will be critical in determining whether Musk's prediction that space becomes the lowest-cost location for AI computing actually materializes.
