Elon Musk’s companies increasingly resemble the components of a single industrial system rather than a collection of unrelated ventures. SpaceX now encompasses Starlink and, after a series of transactions, xAI and X; Tesla is moving beyond electric cars into energy storage, autonomous machines and humanoid robots. Mr Musk himself has acknowledged the growing overlap between Tesla and SpaceX, keeping speculation about an eventual combination alive.
The industrial logic is stronger than it once was. The legal and political logic is not. Tesla’s deep exposure to China sits awkwardly beside SpaceX’s growing importance to America’s defence, intelligence and space infrastructure. Any merger would also be one of the most complicated related-party transactions in corporate history, forcing public shareholders to decide how much Tesla and privately controlled SpaceX are really worth.
The most likely outcome, therefore, may not be a dramatic merger at all. Mr Musk may first build an increasingly integrated economic ecosystem through cross-shareholdings, supply contracts, shared infrastructure and common technology. By the time a formal combination becomes necessary, much of the merger may already have happened in practice.
For most of their lives Tesla and SpaceX shared little beyond a founder. One built electric cars; the other rockets. Their customers, regulators, economics and capital requirements were radically different. A merger between them would once have looked like the sort of conglomerate-building that American capitalism spent the 1980s trying to dismantle.
That objection is becoming harder to sustain.
Tesla is no longer merely an electric-car manufacturer. Its ambitions increasingly lie in autonomous driving, artificial intelligence, energy storage and humanoid robotics. SpaceX, meanwhile, is no longer merely a launch company. Starlink has turned it into a communications network; its acquisition of Mr Musk’s artificial-intelligence interests has pushed it further towards computing and digital infrastructure.
Seen from Mr Musk’s perspective, the pieces are beginning to fit together.
Tesla supplies batteries, energy-storage systems, motors, manufacturing expertise and, eventually perhaps, millions of autonomous robots. SpaceX supplies launch capacity, satellite communications and access to space. xAI supplies foundation models and computing. Starlink provides a global network. A future Optimus robot constructing infrastructure at Starbase, operating a warehouse connected by Starlink and powered by Tesla batteries would not obviously belong to one company rather than another.
Mr Musk has begun to acknowledge this convergence. Asked recently about combining Tesla and SpaceX, he noted that there was “more and more overlap”, while declining to discuss a transaction that would require formal procedures. That was not a merger announcement. But neither was it the straightforward denial that would have ended the speculation.
There is precedent for taking such possibilities seriously. Mr Musk has already been consolidating his corporate empire. X and xAI were combined; those assets subsequently became part of the broader SpaceX orbit. Tesla has also invested in SpaceX, while commercial transactions between the companies are becoming more substantial.
The direction of travel is therefore clear even if the destination is not.
The case for combination
Conglomerates are usually justified badly. Managers speak of “synergies”, which often means that shareholders are being asked to finance the chief executive’s enthusiasm for empire-building. Investors can diversify cheaply by themselves; companies rarely need to do it for them.
Mr Musk’s collection of businesses is somewhat different.
There are genuine complementarities between artificial intelligence, robotics, energy, communications and space infrastructure. AI data centres require enormous quantities of electricity and storage. Autonomous machines require computing and connectivity. Space operations require robots, batteries and sophisticated manufacturing. Satellite networks generate strategic advantages for vehicles and machines operating far from terrestrial infrastructure.
The underlying system might eventually resemble a vertically integrated technology stack:
energy provides power; semiconductors and data centres provide computation; artificial intelligence provides decision-making; robots provide physical labour; Starlink provides communications; Starship provides transportation beyond Earth.
Viewed this way, Tesla and SpaceX are not two unrelated companies but two ends of the same emerging machine.
There is also a capital-allocation argument. SpaceX may generate large cash flows from Starlink while requiring enormous investment in Starship and future space infrastructure. Tesla’s energy and robotics businesses may demand similarly vast sums. Combining them could allow Mr Musk to allocate capital across projects without negotiating repeated related-party agreements.
For a founder who thinks in decades rather than quarters, that flexibility would be attractive.
Yet the same logic contains an important counterargument. Almost all these benefits can be obtained without a merger.
SpaceX can buy Megapacks from Tesla. Tesla can purchase Starlink services. Optimus robots can work at SpaceX facilities. xAI can license models to Tesla. Joint ventures can build data centres. The companies can invest in one another.
Corporate law does not require common ownership merely because two businesses cooperate.
The question for Tesla’s independent directors would therefore not be whether collaboration makes sense. It plainly does. It would be why shareholders should bear the risks of a full combination when contracts can capture much of the benefit.
That is where the merger becomes difficult.
The China problem
The greatest obstacle may not be antitrust law but geopolitics.
SpaceX has become increasingly important to the American state. It launches national-security payloads, provides satellite communications and supports some of the most sensitive infrastructure used by America’s defence and intelligence establishment. Its relationship with Washington is therefore fundamentally different from that of an ordinary commercial company.
Tesla, by contrast, has one of the deepest Chinese footprints of any major American technology manufacturer.
Its Shanghai factory has been central to its global manufacturing system. The company relies extensively on Chinese suppliers and operates within China’s regulatory and data environment. Tesla’s success there has long been one of its greatest strategic achievements.
Inside a merged company, it could become a strategic liability.
American officials would reasonably ask whether an organisation trusted with sensitive military communications and classified space programmes should simultaneously operate one of its most important industrial bases under Chinese jurisdiction.
Chinese officials might ask the inverse question: whether a company deeply integrated into America’s defence establishment should be permitted broad access to Chinese manufacturing, data and infrastructure.
These are not questions that clever lawyers can necessarily solve.
One possible answer would be to separate Tesla’s Chinese operations before any combination. The Chinese business could become an independently governed subsidiary, a joint venture or even a separately listed company. Tesla might retain economic exposure through licensing, branding or minority ownership while insulating the American parent from some political risks.
Such a restructuring would be extraordinarily complicated. Mr Musk has also denied that a Chinese separation is currently under discussion. But if geopolitical rivalry between Washington and Beijing continues to deepen, the corporate architecture of his empire may eventually have to reflect it.
In that sense Tesla’s Shanghai factory could become the most consequential obstacle standing between Tesla and SpaceX—not because it is unsuccessful, but because it has been too successful.
Whose empire is it anyway?
Then comes corporate governance.
Mr Musk can reorganise privately held companies with considerable freedom. Tesla is different. Its outside shareholders own real economic claims and possess legal rights that cannot simply be subordinated to the strategic interests of the Musk ecosystem.
A Tesla-SpaceX transaction would pose a deceptively simple question: what are the two companies worth?
Suppose SpaceX were valued generously and Tesla conservatively. Tesla shareholders could argue that Mr Musk was transferring their assets into a company over which he exercised greater control on unfair terms. Reverse the valuations and SpaceX shareholders would have reason to complain.
There is no neutral market price for a large privately held company such as SpaceX. Nor is Tesla easy to value. Its automotive business might justify one valuation; its promised future in robotaxis and humanoid robots another.
Billions—perhaps hundreds of billions—of dollars could depend on assumptions about technologies that do not yet generate meaningful profits.
Any deal would therefore require independent directors, outside advisers, fairness opinions and probably a shareholder vote. Litigation would be almost inevitable.
Mr Musk's history with Tesla compensation and governance means courts would scrutinise the process closely.
This problem is more profound than a procedural inconvenience. It exposes a tension at the centre of Mr Musk’s corporate universe.
Mr Musk increasingly appears to manage his companies as pieces of a single long-term project. Yet Tesla’s shareholders invested in Tesla, not automatically in that broader project.
What may be economically rational for “Musk Inc.” is not necessarily fair to every shareholder in Tesla.
The company that does not yet exist
This suggests that the eventual architecture may be more subtle than SpaceX simply swallowing Tesla.
A holding company could make greater sense.
Under such a structure Tesla, SpaceX, Starlink, xAI and other operations could retain separate legal identities while sharing capital, technology and strategic direction. The arrangement would resemble Alphabet more than a conventional industrial merger.
Separate subsidiaries could help isolate regulatory liabilities, accommodate outside investors and, crucially, place sensitive businesses behind different governance walls.
Whether such a structure would actually satisfy American national-security officials or Chinese regulators is uncertain. But organisationally it would better reflect what Mr Musk appears to be building.
For investors, therefore, the most important signals may not be announcements about a merger.
Watch instead for less dramatic developments: Tesla increasing its SpaceX stake; related-party transactions growing from hundreds of millions to billions of dollars; shared AI and semiconductor infrastructure; Optimus becoming an important SpaceX customer or supplier; or Tesla restructuring its Chinese operations.
Any one of these could be explained independently.
Together they would suggest that the corporate boundaries are becoming increasingly artificial.
Merge slowly, then suddenly
Mr Musk has spent much of his career attacking industries organised into narrow silos. Cars, batteries, charging networks and software were combined at Tesla. Rockets, satellites and communications were combined at SpaceX. Artificial intelligence, social media and computing infrastructure are now being drawn into the same orbit.
The ultimate project may therefore not be Tesla, SpaceX or xAI individually. It may be the creation of an integrated industrial platform capable of producing energy, intelligence, machines, communications and transport.
That is an extraordinarily ambitious model of vertical integration—closer to the industrial empires of the late 19th century than to the specialised corporations favoured by modern finance.
Whether it produces extraordinary efficiencies or extraordinary governance problems remains to be seen.
A formal Tesla-SpaceX merger is therefore far from inevitable. It may even be unnecessary.
But that should not obscure the more important development.
The economic merger has already begun.
The legal one may merely be the last step.