Tesla used its first-quarter earnings call to commit to $25 billion in capital expenditure this year and to put Intel at the heart of its in-house chip ambitions, surprising analysts who had read early-quarter spending patterns as a signal that Elon Musk was quietly pulling back.
The headline came from Craig Trudell, Bloomberg's global autos lead, who told Bloomberg Technology on 23 April 2026 that Tesla had gone beyond a number flagged only months earlier. "The numbers looked so good because maybe there was a moment where everyone wondered whether or not Tesla was actually going to go forward with this sort of high-wire act of spending the $20 billion plus in capex that was outlined just a few months ago. For them to actually only have roughly two and a half billion dollars of capex in the first quarter, that sort of sent a signal potentially that spending wasn't going to be quite at the pace that was flagged. Instead, Musk hops on the call and says 'oh no, we're going ahead with that and then some.'"
The market reaction was punishing. Tesla shares fell sharply after management confirmed the company would enter negative free cash flow for the balance of the year as it funds Robotaxi, the Optimus humanoid robot and Terrafab, Musk's semiconductor manufacturing project. "There's been a lot of patience with being able to sort of look past the idea that a lot of these projects have not earned meaningful return and may not for the time being," Trudell said. "We've gotten to a place now where this is a company that is primarily a car business, where a couple of years now they have actually seen declines in sales and declines in their position in their most important segment." He added that both the Robotaxi ramp and Optimus timeline are slipping, with "a bit of incremental indication that maybe next year" for Robotaxi being "a bit far off."
The chip announcement was the earnings call's wild card. Musk said Terrafab will spend $3 billion on a pilot line in Texas and will use Intel's 14A process — news that sent Intel shares sharply higher. Bloomberg semiconductor reporter Ian King described the logic of the spend in industry terms. "This is a long-established practice in the semiconductor industry where you make what's called a pilot line, and this is your way of showing: does my manufacturing technology work, and do my designs work? $3 billion, I know that's a lot, but in the chip industry it really isn't very much at all. It's about 10% of what it would cost to build a leading edge full-size fab."
Intel's read-across for investors is direct. "They don't have enough of their own work, their own demand for their own designs to be able to fill their factory. So they need outside customers," King said. "This raised the possibility that guess what, Elon Inc. is going to be a customer of Intel in some way. We don't know whether he's just going to license that and do it himself, or whether he's going to actually help Intel fill its fabs."
Pierre Ferragu of New Street Research, who rates Tesla a buy with a $600 price target, told Bloomberg Terrafab is best understood as a "five to ten years out" enabler and framed the $25 billion through the lens of vertical integration. "TSMC is now charging on average a 65% and growing gross margin. So if you have your own manufacturing capabilities in-house, whether you work magic against TSMC on unit economics or not, you are actually working magic on your own unit economics."
On the question of whether the capex is justified given sliding delivery numbers, Ferragu pointed to scale. "Tesla this year is set to generate about $15 billion of operating cash flow. So the negative free cash flow of Tesla this year might be $10 billion if they manage to spend fast enough of the 25 billion, and they have a cash balance today of about $45 billion. We're talking things that at the scale of Tesla is actually very reasonable." He added that once Robotaxi scales, "there is a broad consensus that it's a multi-trillion-dollar opportunity. So if the cost to get into the race and win it is $25 billion this year, it's a good price."
Musk pushed back when Ferragu's framing was put to him on the call. According to Bloomberg Technology host Ed Ludlow, Musk rejected the idea that Terrafab is primarily a unit-economics play. "He got asked this on the call: 'are you doing Terrafab, the initiative to manufacture chips at scale, because of unit economics?' And his response was no. He consistently says, 'because I don't believe that TSMC and Samsung can match the supply that Elon Inc. needs to fulfil those long-term ambitions.'"
The Tesla-SpaceX convergence question also resurfaced. Asked whether the Terrafab news added to the case for an eventual merger of the two Musk companies, Ferragu said only that "it's going to become more and more challenging not to do it." He cited Optimus robots remaining cloud-connected, potential Starship-based space data centres and cross-pricing between Tesla and SpaceX as structural reasons. "There are a lot of good reasons for the merger to happen. Now, does that mean it's going to happen? I wouldn't bridge that gap yet."
For investors, the call crystallised a binary. Either the Robotaxi and Optimus timelines compress back into Tesla's guidance and the $25 billion unlocks a multi-trillion dollar market, or the negative free cash flow window lengthens, the car business continues to erode, and the capital cycle gets measured against a company whose core product is no longer gaining share.