The pitch is irresistible: move data centers into orbit, beyond the constraints of Earth's geography and power grids. At Nvidia's GTC conference in March, CEO Jensen Huang declared that "Space computing, the final frontier, has arrived"—a statement that effectively gave the concept a mainstream seal of approval.
The investment is already flowing. Elon Musk's SpaceX has acquired xAI, Musk's artificial intelligence company, and is reportedly planning orbital infrastructure to support it. What was science fiction a few years ago has become, according to IEEE Spectrum, "a serious spending category."
But IEEE Spectrum's engineers and researchers are throwing cold water on the excitement. Their argument, in short: Silicon Valley is dramatically underestimating how hard this actually is. The central problem is heat. On Earth, data centers are voracious heat producers managed by elaborate cooling systems—air handlers, water loops, proximity to cold climates. In orbit, those solutions don't translate. There's no atmosphere to carry heat away through convection, maintenance is extraordinarily expensive, and the hardware environment is punishing in ways that terrestrial engineers rarely have to consider.
The gap between a CEO's confident declaration on a conference stage and a functioning, reliable orbital data center is measured not just in dollars but in unsolved physics problems.
If the industry charges ahead without reckoning seriously with these constraints, it risks building the next great infrastructure bubble—one with a very long and expensive fall back to Earth.