SpaceX and Tesla Are Discussing a Merger. The Numbers Don't Make It an Easy Sell.

SpaceX and Tesla have formed special committees to explore combining the two companies, Elon Musk confirmed this week, adding a new and genuinely consequential item to a list of Musk ambitions that already includes Mars colonization and a humanoid robot workforce. Musk floated the idea during SpaceX's third-quarter earnings call, telling investors "there's a lot of merit to combining" the two firms because "there's a lot to be gained from cooperating on developing superintelligence software."

That's the official pitch. The unofficial one, laid out by analysts, lawyers, and longtime Musk-watchers in the days since, involves voting control, a multibillion-dollar pay package, and a regulatory window that may be closing.

The case Musk is making

Musk's public justification leans on his long-running Mars narrative. He's argued that sustaining human life off-world will require robots capable of operating with minimal oversight, since the communication lag between Earth and Mars makes real-time remote control impractical. That requires sophisticated AI. Tesla supplies two pieces of that puzzle through self-driving systems and humanoid robots; SpaceX, in Musk's framing, supplies the large language models and the data centers to run them on.

The timing isn't an accident. Both companies released earnings this week, and the numbers give some shape to what a combined AI business might look like.

SpaceX reported $7.56 billion in revenue from what the company is now calling its "superintelligence" division. Within that, "SI solutions and infrastructure revenues" jumped to roughly $7 billion, more than triple the prior quarter. Part of that jump came from folding in Cursor, the AI coding tool, which added $608 million. Gil Luria, an analyst at DA Davidson, told The Verge that most of the remaining revenue likely comes from renting out compute capacity to Anthropic and Google, with Grok, Musk's chatbot, contributing comparatively little.

That number might look impressive next to a company like Anthropic, whose leaked S-1 showed $16.3 billion in revenue in the first half of this year from its frontier AI models. But the comparison is misleading. SpaceX's press release touted "increased Grok adoption across enterprise customers and consumers," without breaking out what Grok itself actually generated. If that figure were strong, Musk would likely be citing it directly. As it stands, SpaceX's AI business looks less like a model developer and more like a compute landlord, closer to CoreWeave than Anthropic.

Tesla's quarter, meanwhile, was rough: a second consecutive quarter of declining profits. Musk's attention has visibly shifted away from the core car business toward robots and self-driving taxis. A recent demo of a Roadster with hopping capability generated headlines but little else, since it's years from production, expensive, and not guaranteed to ship at all. Musk didn't elaborate on the merger talk during Tesla's call.

Terafab and Macrohard: the track record so far

SpaceX and Tesla already have some financial overlap. SpaceX bought $131 million worth of Cybertrucks from Tesla and has held Tesla bonds in the past. More relevant to the merger conversation are two joint AI ventures announced earlier this year: Terafab, a planned chipmaking facility in Texas with Intel attached as a partner, and Macrohard, an AI project aimed at simulating software products already made by Microsoft.

Neither has produced much since their announcements. SpaceX's S-1 filing from May described both projects as being "in the very early stages," with financial terms, intellectual property rights, and the length of the collaboration all still unresolved. Intel's CEO reaffirmed the company's commitment to the partnership earlier this month, but the filing notes Intel isn't obligated to stay involved with Terafab, and a definitive agreement may never materialize. An Intel spokesperson didn't respond to a request for comment on whether anything had been formally signed.

On the SpaceX earnings call, Musk said Terafab would reach volume production in the second half of 2028, with a first test chip arriving in mid-2027. Reuters noted dryly that Musk "has a history of setting timelines that his companies have missed." When an analyst asked whether Musk had actually placed orders for the specialized machinery needed to produce cutting-edge chips, he didn't answer. Neither Terafab nor Macrohard appears in the companies' most recent 10-Q filing.

The real argument for merging

Beyond the AI mission statement, there's a more mechanical case for combining the companies, one that financial writer Matt Levine has laid out in detail. All of Musk's companies compete for the same scarce supply of Nvidia chips. A merger would eliminate that internal competition and, in theory, let one unified AI system drive the cars, operate the robots, and run everything else under a single roof. Levine has also pointed out that in the current market, invoking Nvidia is close to a magic password for raising capital, and framed the kind of pitch a combined company could make to lenders this way: offer to build data centers in space, and if that doesn't happen, let creditors seize the car factories and terrestrial data centers instead. Luria added that if Musk builds infrastructure at the pace he's promised, there could be enough spare capacity to sell off profitably.

Levine's broader point is that having one person run two massive, overlapping AI-adjacent businesses as separate public companies creates a structural mess of conflicting interests. Merging them, in his telling, would let Musk make resource decisions, whether chips go to robots or data centers, based on what's best for the whole enterprise rather than pitting Tesla against SpaceX.

Why Musk specifically benefits

The governance details matter here. Musk holds majority voting control at SpaceX through supervoting shares and serves as its CEO, CTO, and chairman. At Tesla, his voting stake is roughly 20 percent. A stock-for-stock merger with SpaceX as the acquirer, which most observers assume given its larger valuation and Musk's greater control there, would hand him effective voting control over the combined company. Ann Lipton, a law professor at the University of Colorado Boulder, told The Verge this is essentially Musk's only remaining path to that kind of control: "He can't change it to a dual-class entity now that it's publicly traded. This is the only way to do it."

Both companies are incorporated in Texas, which Lipton noted makes shareholder lawsuits unusually difficult to bring. Securities law, she said, largely comes down to disclosure: as long as the terms are accurately disclosed to shareholders before a vote, federal requirements are largely satisfied. State-level fiduciary duty claims remain theoretically possible if the deal is structured unfairly given Musk's conflicts, though Lipton pointed to Shari Redstone's merger of Paramount and CBS as proof such deals can clear that bar.

Then there's the pay package. A change of control at Tesla would void the operational milestones tied to Musk's 2018 compensation plan and trigger accelerated vesting of shares that otherwise wouldn't vest until 2028, letting him sell sooner. According to Bloomberg, the combination could add $110 billion to Musk's net worth. If SpaceX values Tesla at around $2 trillion, Musk would receive roughly $6 billion in shares on top of that accelerated vesting.

There's also a calendar pressure point. Bloomberg, citing people familiar with the matter, reported a push to announce a deal by the first half of 2027 so it can close while the current administration is still in office, a regulatory environment that's been notably permissive in part because of Musk's own involvement with the Trump administration's DOGE initiative. With roughly two years left in Trump's term, and AI market sentiment showing signs of nervousness following episodes like the CoreWeave selloff, the window may not stay open indefinitely.

Whether the math actually works

For SpaceX shareholders, the case is less clean. A deal anchored around a $2 trillion Tesla valuation would function as something close to a merger of equals, meaning significant dilution for existing SpaceX investors. Tesla's core business is stagnant, and its AI ambitions are costly. Fortune's Shawn Tully estimated that a combined SpaceX and Tesla could spend close to $50 billion on capital expenditures in the fourth quarter alone, far more than their current businesses generate.

None of this is finalized. The special committees are still in the discussion phase, and SpaceX's own S-1 filing acknowledges that a formal agreement on Terafab and Macrohard, let alone a full corporate merger, may never materialize. What's clear is that the financial logic for shareholders and the personal logic for Musk point in somewhat different directions, and whichever one wins out will likely determine whether this deal comes together before that 2027 deadline matters at all.