People who follow Musk’s musings on social media can attest to the fact that the man is obsessed with the potential of solar energy technology, sometimes with a twist.
The sun generates an immense amount of energy that could be turned into electricity. However, the various layers of atmosphere filter much of that out. In one way, that is a very good thing. The atmospheric filtering makes life on Earth as we know it possible, yet it also limits solar energy collection.
Musk will get around that filter for one specific use by launching solar arrays into space. He is betting big that these solar arrays will solve his, and indeed many of the world’s, processing problems.
Planned Data Center Dominance
Musk wrote last November that SpaceX’s Starship “should be able to deliver around 300 GW per year of solar-powered AI satellites to orbit” or “maybe 500 GW” if the rocket firm really pushes it.
He argued it was the “per year” bit that we should watch out for. Right now the total annual American electricity consumption is a little less than 500 GW. Even at a more leisurely pace of 300 GW per year, “AI in space would exceed the entire US economy just in intelligence processing every 2 years”.
American regulators have been put on notice for what is coming. In January, SpaceX filed a plan with the US Federal Communications Commission for a data center satellite constellation that could eventually stretch to one million satellites.
Musk Ramps Up Solar Production
By the numbers, securing a large enough number of solar arrays to launch is eminently doable. In fact, solar manufacturing for “terrestrial applications” is already over 1,500 GW every year, Musk pointed out, or three times as much as he would want to launch into space at top velocity.
SpaceX has contracted with existing manufacturers to equip many of the things it launches into orbit, including the Starlink satellites, with solar capacity. But that is changing. Current job postings for “Solar Cell Factory” positions “point to an internal SpaceX division that is racing to stand up its own space-solar manufacturing line”, reported Per Aspara.
They are technically separate companies, but Musk also runs Tesla. It started as a car company. Musk has ambitions for it that go well beyond that, including robotics and its own solar manufacturing. Musk has set a goal for the company of boosting America’s domestic solar capacity.
One use that is currently not in Musk’s plans but that would be a very tempting experiment, as more solar capacity migrates to space, would be to beam some of that energy back to Earth.
Red-Hot Knives Stabbing Musk’s Brain
Whether or not Musk’s vision is “realistic” is almost beside the point, according to Scott Alexander, popular writer of the Astral Codex Ten newsletter. The better question might be: is it even remotely possible? If so, there is a good chance Musk will make something like his vision come true.
Many people who have worked closely with the man have noted that he is both unusually focused and unusually stubborn. When he decides on a course, he “comes up with an absurdly optimistic timeline for how quickly it can happen if everything goes as well as the laws of physics allow”, Alexander writes. Then when the project falls short of those expectations, as it almost always does, “it’s like red-hot knives stabbing his brain”.
In consequence, Musk “gets obsessed, screams at everyone involved, puts in twenty-hour days for months on end trying to get the project ‘back on track’” with his unrealistic timeline. He is famous for coming up with “absurd shortcuts nobody else would ever consider, trying to win back a few days or weeks”. Musk uses verbal abuse and firings to deal with internal resistance and lawsuits and other tactics to remove external roadblocks.
It is not always pretty but the effect of all this activity is to move the needle in a way that few bosses have ever managed. One anonymous employee described Musk’s method as “demanding a car go from LA to [New York City] on a single charge, which is impossible, but he puts in such a strong effort that the car makes it to New Mexico”.
The Space Race Is On
Moreover, this is a genuine space race, with other players jockeying for position as well. Richard Lu, CEO of PowerBank Corporation, pointed out that one AI computation satellite has already been launched into space with little fanfare.
The company behind that satellite was Smartlink AI, and it was launched in December into a low Earth orbit.
“This isn’t theoretical – the Orbit AI satellite is already operational, producing solar power and running AI compute in space”, Lu reminded.
Many other solar-in-space satellites are soon to join it.
