Sure, let’s smoosh SpaceX and Tesla together, who cares

SpaceX and Tesla have formed special committees to discuss a merger, and Elon Musk said on SpaceX's earnings call that there is a lot to gain from cooperating on what he calls superintelligence software. Bloomberg reports a push to announce a deal by the first half of 2027 so it closes under the current administration.

This would fold cars, robots, rockets, satellites, data centers and a fast-growing AI compute business into one company. For shareholders, for antitrust, and for how AI infrastructure gets financed, that's a very big deal.

The industrial logic isn't crazy, and we should give it its due. Self-driving, humanoid robots, language models and data centers all fight over the same scarce chips and engineers. Columnist Matt Levine's point, cited by The Verge, is that one person running two giant AI-hungry companies is a standing conflict of interest, and merging them would at least let resources go where they're most useful. Pairing hard assets like factories with a data-center buildout could also make borrowing easier.

But look at what the numbers say about the AI business itself. SpaceX's superintelligence division reported $7.56 billion in Q3 revenue. Cursor contributed $608 million after being folded in, and an analyst at DA Davidson told The Verge most of the rest was likely renting compute to Anthropic and Google, with Grok contributing less. Grok's share wasn't broken out. That looks more like a neocloud than a frontier lab, and it's a business with heavy capital needs. SpaceX is about to borrow another $40 billion for Nvidia chips, and Fortune estimates the two companies could spend close to $50 billion on capex in Q4 alone.

Then there's governance, which is where we land hardest. Musk holds over 50 percent of the vote at SpaceX through supervoting shares, versus about 20 percent at Tesla. If SpaceX buys Tesla in a stock deal, he gains control of the combined company in a way he can't get at Tesla directly, as a law professor quoted by The Verge put it. Both companies are in Texas, where shareholder lawsuits face steep barriers. A change of control could also trigger a payout under his Tesla pay package if the price reaches $2 trillion or more, plus accelerated vesting.

The joint AI projects that would justify the merger remain thin on paper. SpaceX's own S-1 said Terafab and Macrohard were in very early stages, with key terms unsettled, and neither appears in the latest quarterly filing. Musk now says Terafab volume production is due in the second half of 2028.

Our read: concentrating this much AI compute, physical infrastructure and voting power in one entity, with weak shareholder recourse, is exactly the kind of scale that deserves scrutiny. If the superintelligence story is real, it should survive a hard look at the terms.

Tesla reported a second consecutive quarter of declining profits as it shifts focus toward robots and robotaxis. The companies already have ties, including SpaceX's purchase of $131 million of Cybertrucks and joint projects Terafab, a planned Texas chip fab with Intel loosely involved, and Macrohard, an AI software project announced in March.