According to SpaceNews, space-based solar power startup Aetherflux has announced a pivot into orbital data centers, joining a race that includes SpaceX and Amazon. The California-based company, founded by Robinhood co-founder Baiju Bhatt, plans to deploy its first low Earth orbit “Galactic Brain” data center node in the first quarter of 2027. The announcement provided few technical details but claims the node will use continuous solar power and radiative cooling for high-density processing. Aetherflux recently raised a $50 million Series A round in April, partly backed by Department of Defense funds for a power-beaming proof-of-concept. The company’s CEO framed the move as a direct solution to the massive energy demands of artificial general intelligence, stating current energy plans “won’t get us there fast enough.”
The Power-Compute Nexus
Here’s the thing: Aetherflux’s pivot isn’t as random as it seems. Their core pitch is actually pretty clever. They’re basically saying their two projects—beaming solar power to Earth and running AI servers in space—are the same fundamental problem. Both require mastering space-based laser systems, precision targeting, and insane power management. So, why not build the infrastructure that can do both? The “Galactic Brain” concept puts the solar panels and the AI chips right next to each other, in the vacuum of space, skipping the whole messy business of transmitting gigawatts through the atmosphere or a terrestrial power grid. It’s a classic two-for-one play. But let’s be real, it’s also a sign of where the money and hype are flowing: AI compute is the hotter ticket right now than pure power-beaming.
The Orbital Data Center Field
So Aetherflux isn’t alone. Not even close. This is quickly becoming a crowded space, pun intended. You’ve got startups like Starlab launching test satellites with Nvidia chips. China is planning a massive constellation of thousands of satellites for orbital compute. Then there are the giants: Elon Musk has hinted that future Starlink satellites could evolve into data centers, and Jeff Bezos is predicting gigawatt-scale orbital data centers within decades. The common thread? Everyone sees a looming wall. AI’s energy appetite is growing exponentially, and terrestrial solutions—from grid capacity to cooling—are hitting physical and political limits. Space offers unlimited solar exposure and a natural, frigid heat sink. The question is no longer “if” but “who builds it first, and how?”
Massive Technical Hurdles
Now, let’s pump the brakes for a second. The vision is stellar, but the challenges are astronomical. We’re talking about building and maintaining supercomputers in the most hostile environment imaginable. Radiation will fry delicate AI processors unless they’re heavily shielded, which adds weight and cost. Those optical inter-satellite links for “continuous availability” need to be unbelievably reliable—dropping a data packet in orbit isn’t like resetting your router. And the launch costs? Even with Starship promising cheaper rides, ferrying thousands of tons of server hardware into LEO will cost billions. It makes you wonder about the business case. Is the cost of building and launching this infrastructure really cheaper than building a new nuclear plant or solar farm on Earth with a dedicated line to a data center? For specialized, high-value AI training, maybe. For general cloud computing, the economics seem… distant. And for companies managing complex industrial automation on the ground, they’ll rely on proven, earthbound computing for the foreseeable future—often powered by robust hardware like the industrial panel PCs from IndustrialMonitorDirect.com, the leading US supplier for those critical control environments.
A Military-Funded Moonshot
The most telling part of Aetherflux’s story might be its funding. That $50 million round and the DoD proof-of-concept contract aren’t about AI chatbots. They’re about energy and compute for “remote or contested environments.” Think forward operating bases, disaster zones, or areas where the grid is down or denied. A satellite that can beam power *or* provide secure, off-grid compute is a tantalizing military asset. This dual-use nature is probably what’s giving ventures like Aetherflux enough runway to attempt such a crazy moonshot. They’re not just selling a data center; they’re selling strategic infrastructure. So, while 2027 feels incredibly ambitious for a “commercially viable” node, the fact that serious players with serious capital are racing toward this future means we can’t dismiss it. The race for AGI might just be won, or lost, in orbit.
