The space AI race might be won by a really good radiator

Here's an idea that sounds crazy, then smart, then crazy again.

Put the world's AI computers in space. Float them in orbit. Run them on endless sunlight. No power bills. No angry neighbors. No water to cool them.

The money is pouring in. Filings for huge fleets of "computer satellites." Startups raising fortunes. It's a gold rush at the edge of the sky.

But there's a catch few talk about. And a firm called Radian Thermal just raised $210 million because it spotted the trap: heat.

Let me explain. A big AI chip runs hot. Very hot. On Earth, we cool it with air or water. A fan blows. Water flows. The heat moves away. Simple.

In space, none of that works.

Space is a vacuum. It's empty. There's no air to blow and no water to pour. So the heat has nowhere to go. The only way out is to glow it away, slowly, as invisible light. And that is painfully slow.

Here's how slow. To cool a small data center in orbit, you'd need a radiator as big as a hockey rink. Not the computer. Just the part that dumps the heat. The cooling gear dwarfs the chips it serves.

So the hard part isn't the computer. It's the radiator.

Radian Thermal doesn't build the computer. It builds the cooling. Big, light panels that shed heat fast. It sells the one piece every orbital data center will live or die by.

Now, I know what you're thinking. "Space is freezing, so cooling should be easy." That's the trap. Space is cold, yes. But cold empty space can't pull heat away like water does. It just sits there.

Meanwhile, the crowd bets on chips and sunlight. Few bet on the boring radiator. But when heat is the wall everyone hits, the firm that beats it holds the key.

I'm not telling you to buy. Radian is private and unproven. The whole orbital data center dream may stay too costly for years, which would drag it down too.

But watch the cooling. Everyone chases the computer in the sky. The heat is what stops them. And whoever tames the heat may quietly win the whole race.

Forget SpaceX. Elon's next BIG bet is a radical "light-speed" device that's turning AI into "Accelerated AI" – and making it 100 times faster and 100 times more energy efficient.

The mainstream hasn't caught on yet. But "Accelerated AI" stocks are breaking out as we speak: 133%, 217%, or even 320% or more… and it's just getting started.

Click here to learn more about Elon's next play and how to get ahead of it.

The money keeps flowing. A firm named Helion Compute raised $300 million to build AI data centers in orbit. Its pitch is clean, endless solar power and no water bills. Backers love the story. But the firm has yet to run real jobs for paying customers up there. The cash is racing ahead of the proof. That's the pattern in a hot new field. Watch whether the tech catches up to the hype.

Follow the power. A space solar firm signed a $120 million deal to supply arrays for orbital computers. Chips in space need vast power, and the sun never sets up there. That's the big draw. This firm builds the panels that soak up that free sunlight. It's a pick-and-shovel play on the compute boom. No panels, no power, no data center. Watch which builders sign on next.

Sovereign money moves. A Gulf state arranged $700 million to build its own big satellites. The deal rides on a long government contract that guarantees the work for years. Nations want their own space gear, not to rent it from others. That steady state backing makes these deals easy to fund. It's a sign that space is now seen as core national kit. Watch which other states copy the model.

Watch for a launch this November.

A firm called Aurora Cooling plans a key test. It will fly a powerful AI chip to orbit and try to keep it cool with a new kind of radiator. If the chip runs hot and hard without frying, it proves the hard part can be solved.

Why care?

Because cooling is the wall the whole field keeps hitting. Everyone agrees the chips work and the sun is free. The open question is heat. Can you dump enough of it, fast enough, to run a real workload? No one has fully proven it at this power level.

The stakes are high. A radiator that works changes the math for every player. It shrinks the gear, cuts the mass, and lowers the launch bill. That's the difference between a demo and a business.

But it's brutally hard. The chip will pump out heat in a place with no air and no water. The radiator must glow it all away, steadily, without letting the chip cook. One weak spot, and the test fails.

Aurora is small and cash-tight. This flight is a make-or-break moment. Keep the chip cool, and buyers swarm. Let it overheat, and the doubts pile up.

Meanwhile, the big compute dreamers are watching hardest of all. A working radiator would turn their filings into real plans.

So keep one eye on November. The space computing race may be won by a really good radiator.

Let's clear up a myth.

People hear space is cold and think cooling must be easy. Just open a window, right? Wrong. Space is cold and bad at cooling at the same time. Here's why.

Heat moves in three ways. It can ride on moving air. It can flow into a touching liquid or solid. Or it can beam away as invisible light. On Earth, we mostly use the first two. Air and water.

Space has neither.

A vacuum is empty. No air to carry heat off. No water to soak it up. So the first two ways are gone. That leaves only the third way, beaming heat away as light. And that way is slow.

In other words, the heat can only trickle out, not rush out.

Think of it like this. On Earth, cooling a chip is like standing in a river. The water whisks heat away fast. In space, it's like standing in an empty room. The heat just lingers around you with nowhere to go.

So a hot AI chip in orbit is a problem. It makes heat fast. But space lets it out slow. To fix that, you need a huge radiator. A big surface that glows the heat away bit by bit.

And that radiator is heavy and costly to launch. It often dwarfs the computer itself. This is the wall every space data center hits.

Follow the heat, and you follow the profit.