Commentary: Elon Musk’s AI side gig should keep its distance from Tesla

NEW YORK: The all-in view of Tesla was summed up in a line from a report this summer by one of the more all-in analysts covering the company: “The car is to Tesla what the video game chip is to Nvidia. The car is to Tesla what selling books is to Amazon.”
Morgan Stanley’s Adam Jonas was arguing that Tesla is morphing from a maker of electric vehicles (EVs) to an artificial intelligence and robotics powerhouse. On this reading, a Tesla isn’t just a box-on-wheels with a battery but an AI-delivery vehicle.
CEO Elon Musk also touts this thesis, declaring with typical bravado: “If somebody doesn’t believe Tesla is going to solve autonomy, I think they should not be an investor in the company.”
An asterisk on that statement might be in order. This weekend, the Wall Street Journal reported that xAI, Musk’s artificial intelligence side gig to his social media side gig, X, has floated the possibility of licensing its models to aid Tesla’s autonomous driving ambitions in exchange for a revenue-sharing deal.
Musk dismissed the story over the course of several tweets, including calling it “nonsense”.
On the other hand, only five months ago, Musk accused Reuters of “lying” when it reported that Tesla had scrapped plans for a sub-US$30,000 EV model, only for the company to soon after announce a pivot from a very cheap EV toward robotaxis.
Plus, Musk has already polled that body of impartial sages known as his followers as to whether Tesla should invest US$5 billion in xAI (the majority answered in the affirmative). 
That flapping sound you hear is the wind blowing through a resplendence of red flags.
Musk has a habit of treating his various separate companies as parts of a unified whole. The most egregious example was Tesla’s acquisition of SolarCity in 2016, whereby the EV company bought a spiralling rooftop solar business that was chaired by Musk, owed him (and SpaceX) money and was run by his cousin.
Even the analyst Jonas later characterised that deal as a “controlled detonation”.
When Musk first bought Twitter, which became X, he brought in engineers from Tesla to review its code. A few months ago, he redirected thousands of Nvidia chips destined for Tesla to xAI, explaining it away as a logistical hiccup.
If Tesla were to now actually invest billions in xAI, and/or share revenue with it, the potential conflict of interest could hardly be clearer. X has struggled as a business since Musk’s takeover, with advertising revenue under pressure and Fidelity, one of the co-investors, marking down the value of its stake enormously.
Musk has indicated that the shareholders of X would own 25 per cent of xAI, though it isn’t clear if that would be individually or collectively through X itself. Either way, any infusion of cash or revenue potential from Tesla would be useful not only to xAI but to the broader X project.
Such a deal would also make a mockery of Tesla’s justification for awarding a gigantic options package to Musk and then pushing for its reinstatement after a Delaware court struck it down.
Chairwoman Robyn Denholm exhorted shareholders to vote yes by arguing, in part, that it would keep Musk’s roving eye focused on Tesla: “We want those ideas, that energy and that time to be at Tesla, for the benefit of you, our owners,” she wrote.
This followed Musk’s unsubtle threat to take his AI vision elsewhere if he didn’t get a bigger stake in Tesla (after he had sold off a chunk during the takeover of Twitter).
If Tesla were to end up investing in, or contracting with, another Musk-controlled venture that literally has AI in its name, that rather suggests at least some of his AI ideas went elsewhere, options or no. Beyond the conflicts and governance-trolling, though, there’s a more fundamental problem here. 
Jonas, like similarly bullish analysts, ascribes only a fraction of his Tesla valuation to the making and selling of EVs, even though this accounts for the vast majority of the company’s revenue and profit. The only way to remotely justify the almost US$700 billion market cap is to believe that Tesla isn’t a mere automaker but instead an AI pioneer.
Yet if that is the case, if Tesla is going to “solve autonomy,” why would it consider outsourcing core AI functions or feel the need to invest billions of dollars into a startup launched only last year?
Automakers do partner with other companies for certain technologies. But Musk has been saying for years that Tesla is already on the cusp of cracking vehicle autonomy under its own steam. Putting money, or revenue, into xAI undercuts that narrative and would divert some of the promised value away from Tesla’s shareholders to Musk’s own ventures.
Perhaps these are mere proposals and polls that stay just that. Still, there was another line in that Jonas report that sticks, telling investors to prepare for the company’s new strategy “to more conspicuously connect Tesla with Elon’s other controlled enterprises such as SpaceX/Starlink, X and xAI”.
Autonomy is, after all, about just doing what you want.