July 21, 2026, (Inside AI) — Tesla is poised to report its first quarterly cash burn in over two years on Wednesday, as heavy spending on artificial intelligence and robotics collides with a recovering auto business. The electric-vehicle maker’s free cash flow is expected to swing to a negative $3.3 billion for the second quarter, according to LSEG data, driven by surging capital expenditures on AI infrastructure, data centers, and manufacturing capacity.
The spending underscores CEO Elon Musk’s strategic pivot from carmaker to “physical AI” powerhouse, betting on robotaxis and humanoid robots to justify Tesla’s lofty valuation. But with capex projected to hit $25 billion this year, outpacing cash generated by core automotive and energy operations, investors are demanding tangible proof that these bets are building a durable competitive moat.
“As capex more than doubles and free cash flow turns negative, investors are increasingly focused on evidence that Tesla’s spending is strengthening its physical AI moat,” Morgan Stanley analysts wrote in a note.
The earnings report lands at a critical juncture. Tesla’s robotaxi service, launched in Austin, Texas, in April 2025, remains confined to just four cities—Austin, Dallas, Houston, and Miami—far short of Musk’s prediction that it would cover half the U.S. population by end of 2025. In January, Tesla promised expansion to seven new cities in the first half of 2026, but that has not materialized.
On Tesla’s investor-relations site, frustration is boiling over. The most-voted question ahead of the call, submitted by a retail investor, was: “What is keeping Tesla back from accomplishing these short-term goals that they've set for themselves?”
Nine of the top 10 questions center on AI-driven initiatives—robotaxis, Optimus humanoid robots, and Full Self-Driving technology. Another investor asked: “Why has growth of robotaxi vehicles stalled? When will we see Cybercab start customer rides?”
Tesla has begun manufacturing the Cybercab, a purpose-built robotaxi without a steering wheel or pedals, but deployment has been delayed. Musk himself warned the production ramp would be “agonizingly slow,” a phrase that now echoes as a refrain for the entire physical AI strategy.
The cash burn marks a sharp reversal from recent quarters, when Tesla’s automotive cash generation helped fund its AI ambitions. Now, even a record vehicle-delivery quarter—1.7 million units expected for 2026, up 3.9% from last year—may not be enough to offset the spending spree. Higher oil prices have boosted EV sales, especially in Europe, but analysts caution that the auto rebound alone cannot bridge the gap.
Barclays analysts noted that a stronger automotive business would help generate needed cash, but the second quarter’s vehicle-sales surge likely falls short. Deutsche Bank analysts flagged additional profitability headwinds: the elimination of upfront Full Self-Driving software purchases earlier this year and low interest-rate financing in May. Wall Street expects automotive gross margin excluding regulatory credits to slip to 18.1% from 19.2% in the prior quarter, per Visible Alpha data.
The spending trajectory mirrors broader industry dynamics where AI infrastructure costs are ballooning. A 2024 McKinsey report on AI infrastructure noted that hyperscaler capex for AI could exceed $200 billion annually by 2027, driven by data center buildouts and specialized hardware—a trend Tesla is now riding. Yet, unlike cloud giants that monetize AI through enterprise services, Tesla’s payoff hinges on consumer and fleet adoption of autonomous systems, a path littered with technical and regulatory hurdles.
Musk’s history of missed deadlines—from Full Self-Driving timelines to robotaxi rollouts—adds to the skepticism. The Cybercab, unveiled with fanfare, remains a prototype in the wild, while competitors like Waymo continue to expand their own driverless services. Waymo now operates in multiple cities, including San Francisco and Phoenix, and has begun charging for rides, underscoring the gap between promise and execution.
Still, Tesla’s AI bets are not without foundational work. The company’s Dojo supercomputer, designed to train neural networks for autonomous driving, represents a homegrown attempt to reduce reliance on Nvidia chips. Progress on Optimus, though early, hints at a future where humanoid robots could reshape manufacturing and logistics. But these are long-term plays, and the market’s patience is finite.
The earnings call will be a litmus test for Musk’s ability to convince investors that the cash burn is a bridge to dominance, not a pit of diminishing returns. With profit expected at 50 cents per share, up from 40 cents a year ago, the numbers may offer a thin cushion. But the real story will be written in the answers to those retail investors’ questions—and whether Tesla can finally turn its AI narrative from science fiction into quarterly reality.
