Table of Contents
If it were up to tech companies, there’d be data centers floating in the space today. But reality tells a different story
Conversations about orbital data centers have started to shift from if to when. But advocates and skeptics agree on one thing: there are some very tall barriers to getting it off the ground. Tim Farrar, a veteran analyst in the satellite industry and this week’s guest on Pulse, says the concept might be useful but it faces a long road ahead.
Farrar is a believer in the vision. “There’s definitely a need for orbital data centers in space to process data in real time,” he said, pointing to use cases like missile defense systems that require rapid response times. “So doing that all in space sounds like a good idea,” he said.
Farrar however expressed doubts about the cost advantage of the project. “Whether this scales up to something that according to Elon Musk is going to be cheaper to do in space than it is on Earth is much more doubtful,” he said.
The concept of space data centers is backed by some of the biggest industry heavyweights, like Elon Musk, Jeff Bezos, Sundar Pichai, Jensen Huang, and Sam Altman. It is built on the premise that data centers in space will have access to abundant solar energy, and can therefore compute endlessly without the energy bottleneck on Earth. “It’s always sunny in space!” as Musk wrote in a blog post, explaining the idea.
But the viability of the concept is debatable. For one thing, it requires enormous capital. Just building the foundation and hauling components into orbit would cost billions of dollars. For reference, flying each ton of payload into space costs close to a million dollars. Suffice to say, the financial aspect of it is harrowing as it is.
The other thing is physics. There is no air in space — hence no medium for cooling. And as AI data centers famously produce excessive amounts of heat, and in space, will spend a good amount of time in direct sunlight, they have to be cooled using thermal radiation. This will require massively complex radiators engineered to redirect the heat away from the servers and keep the systems thermally stable.
But that’s only the tip of the iceberg. Scientists worry that the excess heat from space-borne data centers can radiate back to Earth causing the atmosphere to melt.
Third is launch capacity. This is where SpaceX has a massive advantage with the Falcon 9, a workhorse launch vehicle powered by reusable technology. The upcoming Starship, an improvement on the Falcon 9, will pack 150 metric tons of capacity, allowing it to carry much more payload in one ride. SpaceX describes it as “the world’s most powerful launch vehicle.”
“There is a lot of work needed to be done to get orbital data centers into space in any sort of volume,” Farrar said. “The Starship rocket needs to be working and move at a regular tempo to be able to launch these satellites into the correct orbit,” he said. Starship is currently in testing.
Although many nations are now building their own sovereign launch capabilities, Farrar said that the trick is to have a vertically integrated model that combines satellite production, launch and deployment, which SpaceX has. “Integrating, launch and manufacturing of payload is critically important to gaining scale and critical mass,” he said.
With first-mover advantage in a number of plays within the satellite space, Farrar argued that Starink will continue to dominate the market for the next decade at least. “Starlink is going to be dominant for the next decade or more,” he said, adding that countries like China, Europe and Canada may eventually rise to challenge it with their sovereignty push, but whether those will be truly competitive from an economic perspective remains questionable.
