On May 27, CNBC reported that, according to a Tesla employee and others familiar with the talks, the EV-maker and rocket and AI purveyor are weighing a merger. Prior to that story, speculation about the potential tie-up was already running rampant. Wedbush Securities analyst Dan Ives put the chances for a combo at 80%, adding that the game plan was already in place for fusing operations at Elon Musk’s two biggest holdings. Long-time Tesla investor Ross Gerber, citing that Musk had already folded xAI into SpaceX, stated that this new gambit would advance his vision of running one big company amounting to a kind of Berkshire Hathaway of AI-driven tech. As of today, betting site Kalshi displays 52% odds that a mega-deal will happen by May of next year.

Though it’s impossible to predict if so many Wall Streeters are making the right call, one thing’s for certain: Capitalizing on the incredible buzz surrounding the pending SpaceX IPO as a strategy for rescuing stricken Tesla makes perfect sense for Elon Musk. At an expected market cap of $1.75 trillion, SpaceX stock looks vastly overpriced (and, as I’ve written, an IPO prominent analysts are saying they’d avoid). So Musk could marshal its inflated shares as currency to pay big for Tesla, even making the deal at its current market cap, a number that’s also over the top based on any conventional metric. Without SpaceX as an acquirer in the wings, Tesla looks highly vulnerable to a major selloff, given that it’s somehow maintaining a gigantic, even expanding valuation as its profits dwindle. “It’s been my intuition for a long time that this has to happen,” says David Trainer, CEO of research group New Constructs. “It’s the only way to bail out Tesla shareholders. It’s what Tesla investors have been expecting for a long time,” and in his view, the anticipated grand exit that’s been bolstering its stock.

A SpaceX-Tesla union would mark by far the largest merger of all time. Remarkably, the two enterprises’ valuations are about equal, since Tesla’s market cap of $1.65 trillion sits just a shade behind SpaceX’s anticipated debut at $1.75 trillion. Let’s assume the most likely scenario that casts SpaceX as the buyer. To clinch the deal, it would need to issue a batch of new shares equivalent to 94% of the current number, reflecting the ratio between the two valuations ($1.65 trillion divided by $1.75 trillion). And that’s assuming SpaceX doesn’t have to pay a premium as almost always required in a takeover. SpaceX’s count would nearly double from the 4.1 billion shares its S-1 IPO filing effectively forecasts as of the debut, to 8 billion.

If SpaceX can do the deal around the same price where its stock’s expected to open following the IPO, slated for mid-June, the combined SpaceX-Tesla would emerge sporting a valuation of $3.4 trillion. It would edge today’s levels for Amazon ($2.9 trillion) and Microsoft ($3.2 trillion) to rank as the world’s fifth most valuable publicly-traded company, trailing only Apple and Google ($4.6 trillion each), Nvidia ($5.2 trillion), and Saudi Aramco ($6.7 trillion).

The problem: The behemoth wouldn’t be making any money, and the chances it can deliver for investors is beyond remote

While the $3.4 trillion valuation is breathtakingly big, the combined profits generated by the two potential partners wouldn’t even qualify as shockingly small––based on recent results, they’d be negative. In the past four quarters, Tesla posted $3.9 billion in GAAP net earnings. That’s down from $15 billion in 2023, and $7.0 billion in 2024. Even the official figure overstates its core profitability. In the year ended in Q1, Tesla booked $1.6 billion from sales of regulatory credits, a category that’s already dwindling and Musk admits will probably disappear, plus proceeds from shrinking its cache of Bitcoin. So its real “core” earnings are more like $2.3 billion. But let’s stay optimistic and use the official GAAP number at $3.9 billion.

What about SpaceX’s contribution? It registered a loss of $4.94 billion last year, and the trend’s downwards: It made a tiny but still positive $791 million in 2024. Its Starlink “connectivity” mobile and broadband sector that collects subscription revenues from service provided by its galaxy of orbital satellites is highly profitable. But the space side’s in deficit, and AI’s the biggest burden, imposing nearly $9 billion in operating losses in the past five quarters. At today’s numbers, a pro-forma SpaceX-Tesla would show a GAAP yearly earnings of around minus $1 billion.

If anything, the cash flow picture is even more alarming. In the S-1, SpaceX acknowledges that it will be deploying capex to build its AI data center footprint far in excess of the cash it generates from operations for an extended period before that franchise turns highly profitable. Last year, its free cash flow deficit reached $14 billion. Through Q1, Tesla was collecting around $1.5 billion a quarter more in cash from selling cars and batteries than it was spending on plant and equipment, mainly for robots and robotaxis still in development. But that formula’s reversing: For the final nine months of this year, it’s expecting to spend a total of at least $22.5 billion on capex, far more than the cash from operations delivered by the car and battery franchises. Analysts are predicting sharp increases from those levels as Tesla ramps spending on AI infrastructure for self-driving cars, Optimus and robotics facilities, and its Terafab project.

So both sides as stand-alones harbor gigantic investment needs that will far outstrip their ability to garner cash from their basic businesses in the years ahead. Those extra tens of billions have to come from somewhere. The S-1 suggests that SpaceX will float more stock and issue debt to fill the gap. That course would dilute shareholders and raise interest expense, curbing profits. Instead of contributing excess cash to ease that burden, adding Tesla––now embarking on a capex blowout––would make the scenario worse.

SpaceX and Tesla are already working closely together. The automaker invested $2 billion in xAI, bought by SpaceX in February, and sells the rocket-maker cybertrucks. SpaceX in the past two years has purchased over $700 million in battery storage systems for data centers from Tesla.

According to analysts, the deal’s rationale appears to hold that both companies count on riding the AI wave to greatness, and can roar even faster as a team. In its S-1, SpaceX calculates the total addressable market for AI at $26.5 trillion, 13-times the potential for connectivity and rockets combined. But as Trainer points out, huge TAMs attract big competition, and SpaceX-Tesla would face strong pricing pressure from the likes of Amazon and Google in AI, and everyone from T-Mobile to Vodafone in broadband and mobile.

As Trainer argues in an excellent, and highly critical, report on SpaceX prospects, it would need to achieve future numbers no company worldwide shows today in order to reward investors buying at a $1.75 trillion market cap. Specifically, using a discounted cash flow analysis, Trainer puts the bogeys for delivering cost-of-capital returns at $248 billion in net income, and $1.1 trillion in revenues by 2035. The sales number is 1.5x bigger than the top line at Amazon, the S&P leader in the category, achieved in the past four quarters. The profit requirement is $90 billion more than the biggest earner, Alphabet, made over the same span. As Trainer puts it, those targets are “really out of this world.”

But a merger with Tesla would far more than double the challenge for the original SpaceX shareholders. They’d go from owning 100% of the enterprise to 52%. What they’d get in return is Tesla’s just under $4 billion in current profits, a number that once again is boosted by ephemeral sales of reg credits, and is already falling fast. At Tesla’s current valuation of $1.65 trillion, SpaceX would be spending 420 times earnings, meaning that shareholders are getting just $2.3 in earnings for every $1000 they’re paying. The SpaceX investors would also be shouldering Tesla’s huge capex expenditures in addition their already immense outlays for building out AI.

I called Trainer to ask if indeed, it would require 2x the previous numbers for revenues and profits to ring the bell, and he said that’s indeed the case. What are the chances SpaceX-Tesla could achieve nearly $500 billion in profits and $2.2 trillion in revenue a decade from now? As Trainer told me, “It’s a kind of suspended disbelief, squared.” Musk’s proven a wizard at suspending belief to send Tesla and SpaceX valuations to the stars while their earnings remain stuck earthbound. Buying Tesla using SpaceX stock would solve the EV pioneer’s big time over-valuation problem by cashing out its shareholders. But it would only transfer that burden to SpaceX, and make SpaceX’s math problem far worse. Musk’s on the verge of forging a creation that must achieve feats so hard to believe that, at long last, even his today’s true believers may stop believing.

This story was originally featured on Fortune.com