In one week Elon Musk did three things that the news treated as three separate stories. On 15 October he showed a Tesla Roadster that flew. On 20 October he said the boards of Tesla and SpaceX had formed special committees to study merging the two companies. And behind both sits a SpaceX plan for up to a million satellites that would work as data centres for AI. I think those are one plan. That is my view, not a fact, so I will keep what was reported apart from what I think.
What Was Reported
The Roadster demo was held at a SpaceX test site in Texas. The car used ten nozzles fed by high pressure air to flip, spin and fly over a row of parked vehicles, with nobody inside. It is not a certified aircraft, and reports say production is set for 2028.
On the earnings call, Musk said there is a lot of merit in combining the companies. He also said it would need to clear a high hurdle, a majority vote of Tesla shareholders, and no deal has been agreed. Reports say SpaceX had already absorbed xAI in February.
Tesla's third quarter was a little softer. It delivered 486,532 vehicles, down about 2 percent from last year's record, though ahead of what analysts expected, and revenue was roughly flat.
And Musk reportedly wrote on X that unless there is a major world war, SpaceX probably gets to more than 100 trillion dollars within ten years.
How I Read It
First, Tesla needed a moment. Sales slipping a little and revenue standing still is not a crisis, but it is not a story either. A car that flies, with nobody in it, five days before merger news, is a story. I do not think it is a product, since it is years from production and nowhere near certified. I think it is a signal, and it worked, because people started talking about him with trust again instead of doubt.
Second, look at where it happened. A SpaceX site, with thruster technology from the rocket side of his world. The demo quietly showed why the two companies belong under one roof.
Third, and this is the part I care about, 100 trillion dollars is not a number you reach by selling rockets. You reach it by being the place where AI lives. AI needs huge amounts of power, land and permission, and on the ground it keeps running into neighbours, councils and power limits. In orbit there are far fewer people to ask. There are still licences and treaties, but they were written for a different kind of sky. In my piece on Musk I said the rules were the bottleneck. Orbit looks like the way round them.
The Part Nobody Asks About
At least four countries, the US, China, India and Russia, have destroyed a satellite with a missile in tests. Others have jamming and lasers. A million satellites doing work that more and more of the economy depends on would be the most tempting target in the sky, and a lever over whoever owns them.
Musk seems to know it. He reportedly added a condition to his own forecast: unless there is a major world war. That tells me he understands the biggest threat to the plan. For it to work, nobody can be willing to switch it off, so the world has to agree, or at least be unable, to shoot. That is not an engineering problem.
A Rare Bet
Not everyone thinks it works. One portfolio manager called orbital data centres a next decade event, and a Fortune piece argues they will not solve AI's power problem soon. I could be wrong about the whole plan. But I would call it a rare bet. It is a bet that the sky stays peaceful, that countries with missiles let it happen, and that one company with cars, rockets and AI can carry all three.
Three stories on the news. One plan, if I am reading it right. The flying car buys trust, the merger puts the pieces under one roof, the satellites are the point, and the sky has to stay quiet for any of it to work.