Investors are left with unanswered questions as the company leans on car sales to help fund its expensive AI ventures

Tesla beat expectations with its delivery numbers earlier this month, but there's far more for investors to consider.

Tesla's artificial-intelligence ambitions depend on a car business that already pays the bills, and the company's latest delivery report gives shareholders a reason to examine that business more closely.

Elon Musk's company delivered 486,532 vehicles in the third quarter, more than Wall Street had been expecting. The result is welcome news for a company spending heavily on robotaxis and humanoid robots, because car sales help fund those bets. Yet shareholders still need evidence that the billions Tesla is committing will eventually pay off.

The stronger-than-expected deliveries make only the first part of that case. Deliveries fell 2.1% short of the record 497,099 vehicles Tesla delivered a year earlier, when U.S. buyers rushed to claim the $7,500 federal EV tax credit before it expired at the end of September 2025. More importantly, a vehicle count reveals little about the prices customers paid, the costs of making those cars or the cash left after investment.

Tesla's second-quarter results show how much a delivery count leaves out. Operating cash flow rose 85% from a year earlier to $4.7 billion, but capital expenditures more than doubled to $5.8 billion. That left free cash flow, the cash remaining after investment, at negative $1.1 billion. Tesla generated substantial cash, and its investment program consumed even more. Revenue rose 26%, yet operating income fell 57% to $398 million, leaving Tesla with about 1.4 cents of operating profit for every dollar of sales. Higher sales alone cannot show that Tesla's expansion is becoming easier to finance.

Those cash-flow figures cover all of Tesla, including energy storage and services. The company does not report how much of its capital spending goes to the car business, which needs its own money for factories, new models and service centers. Automotive gross profit, what car sales earn before research, overhead and new investment, therefore cannot be treated as a budget for AI.

The size of Tesla's spending plans raises the stakes. In its latest quarterly filing, the company said it expects capital expenditures to exceed $25 billion in 2026. Having spent $8.28 billion in the first half, Tesla would need to spend more than $16.7 billion in the second half to meet that forecast.

The commitment to heavy spending extends beyond this year. On Tesla's July earnings call, CFO Vaibhav Taneja said capital spending would grow over the next two to three years. He cited expansion of the robotaxi fleet, production capacity for Optimus, Tesla's humanoid robot, a semiconductor plant, solar manufacturing capacity and AI-computing infrastructure, alongside continued investment in the car business.

Tesla has considerable resources. It ended June with $43.5 billion in cash, cash equivalents and short-term investments. On Sept. 29, it arranged $30 billion in credit facilities, replacing a $5 billion revolving credit line, which works much like a corporate credit card. The largest piece is a $20 billion three-year term loan that Tesla can take out later. That is borrowing capacity, and Tesla has used none of it so far: No loans were outstanding under the new facilities on Sept. 29, and the company said it does not currently plan to draw on them for the rest of the year. The agreements give Tesla a cushion if its investment needs grow or its operations weaken.

A cushion is not a funding plan, however. For the car business, I would start with the economics behind the delivery number. Factory efficiency, pricing and vehicle mix, meaning the balance between cheaper and more expensive models, determine how much profit each sale brings in. Discounts and low-rate financing can attract buyers, but they also cut into that profit, so investors need to know what they cost before concluding that demand will last.

Cash flow also deserves a look across several quarters. Tesla delivered 22,141 more vehicles than it built in the third quarter, and selling cars from inventory releases cash that Tesla cannot keep drawing on indefinitely.

Energy storage broadens the picture. Tesla deployed 13.7 gigawatt-hours of battery storage in the third quarter, up from 12.5 gigawatt-hours a year earlier. Profitability has moved the other way. In the second quarter, the energy business's gross margin - the share of revenue left after production costs - fell to 20.4% from 30.3% the year before. More batteries installed does not automatically mean more profit, so investors should track both.

The same scrutiny should apply to the businesses Tesla is spending on. Running a robotaxi fleet requires vehicles, charging, maintenance, cleaning, insurance and remote staff who monitor the cars and step in when they get stuck. The financial case depends on enough paying rides to cover those costs and eventually recover the money spent building the fleet.

Fleet ownership also changes the timing. A car sold to a customer brings in cash on delivery; a car Tesla keeps for its robotaxi service earns its money gradually, one ride at a time. Rapid fleet growth could require a lot of money upfront. Ideally, I would want to see each car spend more of its day carrying paying riders, while costs per car fall as the fleet grows.

Optimus raises a different set of questions. Tesla is installing its first Optimus production lines in Fremont factory space that once built the Model S and Model X, and it said in its second-quarter update that the initial robots would collect training data for its AI and help develop new functions. For now, then, the robots serve Tesla's own research. Investors should look for evidence that Optimus can do dependable work customers will pay for, along with what each robot costs to build and keep running. A robot that impresses in a demonstration still has to work reliably for enough hours to justify its price and upkeep.

The Oct. 21 earnings report offers the next opportunity to connect these questions. I will be watching margins, operating cash flow, capital spending and management's explanation of how the expansion will be funded. Clearer evidence about the economics of robotaxis and robotics would help investors assess the eventual payoff.

The delivery beat showed that buyers still want Tesla's cars. Tesla's next task is to prove those sales will generate enough cash to fund its AI ambitions, and that the money it spends on them will reward shareholders.

-Jurica Dujmovic