On Tuesday, tech billionaire Elon Musk endorsed ARK Invest's chief futurist Brett Winton's prediction that putting data centers in orbit could become cheaper and more efficient as AI demand skyrockets.
AI's Growing Demands Push Innovation Beyond Earth
As artificial intelligence workloads surge, traditional land-based data centers face soaring costs for power and cooling.
Winton explained on X that on Earth, the 100th gigawatt deployed will almost certainly be more costly, complex and time-intensive than the first, but in space, the opposite is true.
"The 100th orbital GW could be 1/3rd as costly as the 1st," Winton wrote.
On Earth the datacenter buildout is subject to backwards cost scaling.
The 100th GW deployed will almost certainly be more costly, complex, time intensive and subject to negotiation than the 1st.
In space, the opposite.
The 100th orbital GW could be 1/3rd as costly as the 1st.
Elon Musk Bets On Space To Power AI With Solar Satellites
Musk has been betting on orbit as the future of AI, making it central to his $1.25 trillion plan to mergeSpaceX with his struggling startup xAI.
He plans to take the merged company public this year and predicts that fleets of solar-powered satellites, naturally cooled by space, could soon become the most cost-effective way to power AI within the next three years.
Space Data Centers Could Slash AI Costs With Free Solar Power, Cooling
In space, solar energy is available 24 hours a day and is roughly 30% more intense without atmospheric interference, Baker said.
"Cooling is free…you just point a radiator into the dark side of orbit," where temperatures could approach absolute zero, he added.
Speed And Efficiency Advantages in Orbit
Baker also highlighted latency benefits. Unlike fiber optic cables on Earth, lasers traveling through a vacuum can transmit data faster, offering potential speed advantages for AI networks.
Join thousands of traders who make more informed decisions with our premium features.
Real-time quotes, advanced visualizations, backtesting, and much more.