Tesla Just Spent $4.73 Billion on Capex. Here’s Where Every Dollar Went.

Tesla‘s (TSLA +4.04%) capital expenditures (capex) totaled $4.73 billion in Q3, more than double the $2.25 billion it spent a year earlier. That’s a substantial increase, although considerably less than the roughly $7.5 billion Wall Street anticipated.

Tesla also reduced its full-year capital spending forecast from more than $25 billion to about $20 billion. But that’s still more than twice the $8.53 billion it spent in 2025. So where is all that money going? Let’s take a closer look.

AI infrastructure and autonomous vehicles

Tesla has been rapidly expanding its artificial intelligence (AI) computing infrastructure at Gigafactory outside of Austin, Texas. This is what’s training the neural networks powering full self-driving, Cybercab autonomous vehicles, and Optimus humanoid robots. And it requires thousands of specialized processors, servers, networking equipment, cooling systems, and electrical infrastructure.

Tesla is also investing in Cybercab manufacturing and its robotaxi fleet. Unlike vehicles sold directly to consumers, Tesla robotaxis require up-front investment in vehicles, charging infrastructure, and maintenance infrastructure before generating transportation revenue. The potential payoff is significant.

Image source: Getty Images.

Estimates from S&P Global Market Intelligence and Visible Alpha suggest Tesla’s robotaxi business could generate about $42.4 billion in annual revenue by 2030. At a hypothetical 20% operating margin, that would produce roughly $8.5 billion in annual operating profit.

Tesla could also generate recurring revenue from its full self-driving software. At $99 per month, 5 million paying subscribers would generate about $5.9 billion annually.

Of course, these projections depend on regulatory approvals, technological improvements, and customers adopting the services in substantial numbers. While I’m optimistic on all three fronts, there are no guarantees.

The Optimus advantage

Tesla is also investing in manufacturing capacity for Optimus, its humanoid robot designed to perform physical tasks in factories and other workplaces. In Q3, Tesla began production of Optimus at its Fremont, California, facility, where it’s converting manufacturing space previously used for the Model S and Model X vehicles.

The company has discussed eventually producing as many as 1 million robots annually at Fremont. But getting there will require significant investment in production lines, motors, sensors, batteries, and specialized manufacturing equipment. And that stuff isn’t cheap.

Tesla Chief Executive Officer Elon Musk has previously suggested that Optimus could eventually cost about $20,000 to $30,000 per robot to manufacture at scale. At a hypothetical selling price of $30,000 per robot, annual sales of 500,000 robots would generate $15 billion in revenue, while 1 million robots would generate $30 billion.

Energy storage is already producing results

Unlike robotics and robotaxis, Tesla’s energy storage business already has an established customer base. In Q3, the company deployed 13.7 gigawatt-hours of energy storage products, up about 10% year over year. Tesla also began production at its new Megafactory in Brookshire, Texas, which has a planned annual capacity of 50 gigawatt-hours.

$407.46+$15.84 (+4.04%)

NASDAQ · As of 4:00 PM ET 10/27/26

If you’re unfamiliar, these large battery systems help utilities store electricity, stabilize electrical grids, and provide backup power. So as you can imagine, demand from AI data centers is creating another potential growth opportunity.

Tesla reported its highest-ever energy storage order rate and backlog during the quarter. It also began shipping Megapacks using domestically manufactured lithium-iron-phosphate battery cells from Nevada.

Semiconductor manufacturing adds another expense

Tesla is also pursuing its ambitious Terafab semiconductor project in Texas, which will support future AI computing requirements. Manufacturing advanced chips requires expensive fabrication equipment, specialized facilities, and substantial technical expertise.

Tesla hasn’t disclosed a complete cost breakdown for these plans, so at this point, we don’t know how much of its projected capital spending will ultimately go toward semiconductor manufacturing.

Can Tesla afford all this spending?

Tesla generated $6.26 billion in operating cash flow during Q3. After capital expenditures, free cash flow reached $1.53 billion. The company also finished the quarter with about $45.4 billion in cash and investments. However, net income declined 43% to $787 million, while revenue remained essentially unchanged at $28.1 billion.

The revised $20 billion annual capital spending forecast suggests Tesla is becoming more selective about how quickly it deploys capital. That’s encouraging from a cash management perspective, but slower investment could also delay product commercialization.

Still, Tesla’s long-term potential remains solid. Robotaxis, Optimus, and AI software could eventually generate billions in additional revenue, while energy storage is already expanding. But spending billions doesn’t guarantee those businesses will become profitable.

Indeed, the next few years will determine whether Tesla’s investments produce returns large enough to justify their cost. But given Musk’s history of turning big ideas that were once considered impossible into reality, I remain bullish.

Source: www.fool.com — article syndicated from the publisher’s feed; all rights remain with the original publisher.