Tesla and SpaceX Weigh a $3.5 Trillion Merger Built Around AI

Investors have joked for years that Elon Musk really runs a single large enterprise with a few different stock tickers. On Tuesday, that joke moved a step closer to something real. In the course of SpaceX's third quarter earnings call, Musk revealed that the boards of SpaceX and Tesla have each put together special committees of independent directors to look into a merger. At current market values, the two together would be worth about $3.5 trillion.

It is not a done deal, and Musk made sure to say it may never become one. Yet his explanation of why they are looking at it tells you where both firms see the money being made: in artificial intelligence, the chips behind it and the robots that do the work.

What Musk Actually Said

The news came out in the Q&A with analysts. Musk said the committees were set up at shareholders' request, and that as the biggest individual shareholder in each company he cannot be involved. "Because I'm a conflicted party, I can't actually be in the discussions of the special committees," he said.

Musk did not want to overstate the case. Asked about the likelihood of a deal, he replied, "I think that would be overstating it." He added that investors would have the last word: "At the end of the day, the shareholders have to be supportive of this." But he left no doubt as to his own view. "I think it would be a good idea," he said, arguing there is "a lot of merit to combining the companies" and much to be had from working together on AI software, chip design, robotics and connectivity.

Still, there are no terms in place, and Tesla's proxy statement, filed on Monday, has no merger proposal for shareholders to put to a vote.

Why AI Is Driving the Conversation

The figures SpaceX put out that day help explain the timing. The company's third quarter results, as filed with the Securities and Exchange Commission, show $14.6 billion in revenue, a 161 percent increase over the prior year, and net income of $4.4 billion, its first quarter of operating profit since it went public in June. Starlink is no longer the top earner. With $7.6 billion, the AI side of the business, rebranded as Super Intelligence, has surpassed the connectivity unit at $5.2 billion and the launch division at $1.8 billion.

SpaceX has not held back on spending to build that operation. In August it closed its purchase of the AI coding firm Cursor at an implied $60 billion, had 1.7 gigawatts of AI computing capacity on hand by the end of September and put $19.6 billion into capital projects during the quarter alone.

Tesla, for its part, has staked its future on self-driving software, humanoid robots and custom AI chips. The pair are already at work on a joint chip plant, Terafab. Seen that way, a merger would be little more than a formal stamp on a relationship that already shares engineers, silicon and ambitions.

What Each Side Brings

SpaceX has more to offer than rockets. The numbers tell the story: 13.6 million Starlink consumer subscribers, the first orbital flight of Starship, which carried 26 next generation Starlink satellites, and a deal announced this month for 800 MHz spectrum licenses that could extend its connectivity reach further into phones. Then there is Starmind, Musk's plan to put AI data centers in orbit.

From Tesla come the cars, a sizable energy storage business, self-driving software, the Optimus humanoid robot and years of know-how making custom AI chips for its vehicles. In the pitch they are easy to combine: AI models trained on SpaceX compute, running on Tesla silicon in Tesla robots and cars, all tied together by Starlink. It is now up to the committees to decide whether that is worth the trouble of merging two such different public companies.

The Hurdles Are Real

The governance side is not simple. With a majority of the voting power at SpaceX, Musk would have an easy path on that side. Tesla is another matter. A merger there would need the support of a majority of all outstanding shares, not merely the ones that vote, a hurdle Musk himself has called a high one.

Independence is an issue too. A number of directors have connections to both firms or to Musk personally, and Tesla's board has been under the microscope for years, not least in the litigation over Musk's pay package. The price will be only part of how any deal is judged; how it was negotiated will count for as much.

There is also the sheer scale of it. At around $3.5 trillion, the combined company would be one of the world's most valuable businesses, spanning rockets, satellites, cars, robots, an AI model business and a coding tool with millions of users. Regulators in Washington and abroad would almost certainly want a close look.

What Investors Heard

Markets took the news in stride, but with interest. SpaceX stock jumped in after-hours trading on the back of its results, and Tesla posted a small gain. Some longtime bulls were enthusiastic; Gene Munster of Deepwater Asset Management posted on X after the call, "Believe me, shareholders would support this."

Tesla is due to report its third quarter numbers on Wednesday, and the merger will probably come up again. Investors will be listening for any word on timing, the committees' advisers, or how Tesla's board intends to manage the conflict of interest.

Is One Musk AI Empire Good for Silicon Valley?

The prospects are plain to see. A merged company would be in a position to bring chip design, manufacturing and AI training compute under one roof, and it could move faster on projects such as Terafab without having to negotiate between two boards. Tesla's cars and robots might well run on models trained in SpaceX data centers, while Starlink connects them anywhere on Earth. Bay Area engineers and suppliers would have one very large customer with deep pockets for talent, software and chips, and for Taiwan's chip and hardware supply chain it could mean bigger, longer orders. And with a single balance sheet, financing the large data center and chip projects both firms say they need could be simpler.

But the concerns are no less concrete. Putting so much AI infrastructure, connectivity and physical robotics under one controlling shareholder raises questions about competition, national security and corporate governance. Tesla shareholders who bought into a car and energy company may find themselves in a far more complicated business with different risks, not least the heavy capital SpaceX is putting into AI. Given the record of related-party deals between Musk's companies, starting with Tesla's purchase of SolarCity, critics will look closely at every term. Then there is the matter of talent: engineers who joined a rocket or car company may not care to work inside a conglomerate whose priorities are set by its AI ambitions.

Some controversy is already taking shape. Supporters argue that the two are in effect a single technology platform and that a merger would simply remove the friction and conflicts that exist today. Skeptics say independent boards are there precisely to shield shareholders from a founder's preferred outcome, and that a merger could deepen the very conflicts it is meant to resolve. Some investors will also want to know whether a $3.5 trillion valuation is backed by real earnings or by enthusiasm for AI.

The signals to watch are specific. What does Tesla say on its Wednesday call? Do the special committees hire independent advisers and set out a timeline? Will a shareholder vote be added through Tesla's proxy or annual meeting? How will regulators react if talks progress? And can SpaceX grow its AI revenue fast enough to justify the spending and the valuation any deal would rest on?

Where the Hardware Comes From

At Sparknify we have been tracking the move from AI software to physical AI. In Physical AI Puts Taiwan Back at the Center of Tech Investment, we made the case that robotics and AI infrastructure rely as much on chipmaking and supply chains as they do on code. A Tesla and SpaceX combination centered on chips, robots and data centers would be the largest bet yet on that idea, and a good deal of what it needs would still have to be built by partners from Hsinchu to San Jose.

One way or another, the message of Tuesday's call is plain: the next stage of AI will be measured in gigawatts and factories, not only in models.

Subscribe to Sparknify for Silicon Valley tech news and the people behind it.