A strange kind of pivot: Tesla reported its first-ever drop in annual revenue even as it announced what amounts to a high-stakes bet on a very different future — one run by robotaxi fleets, humanoid machines and in-house chips.
The headline numbers are blunt. Tesla’s fourth-quarter revenue slid to $24.9 billion, and full-year revenue for 2025 fell to $94.8 billion from $97.7 billion the year before — the company’s first annual decline. Quarterly net income plunged about 61% to $840 million, leaving full-year profit at roughly $3.8 billion, down sharply from the prior year.
Those figures arrive with a dramatic change of emphasis from Elon Musk and the company’s leadership. After two years of sliding auto sales and mounting competition from rivals such as BYD — which has overtaken Tesla in global EV deliveries — the company is spending to reinvent itself.
Dollars and direction
Tesla told investors it plans to more than double capital expenditures this year, targeting roughly $20 billion in spending on factories, robots, driverless rides and the compute needed to power all of it. That comes after capex fell to about $8.6 billion in 2025.
Part of the shift is immediate and concrete: production of the older Model S and Model X will end, and the Fremont lines that made them are being repurposed to build Optimus humanoid robots. The company also confirmed a sizable, recent $2 billion investment in xAI — the startup behind the Grok chatbot — underscoring Musk’s stated aim to fuse Tesla’s physical products with advanced AI software.
Tesla still depends on cars for the bulk of revenue — automotive still accounts for roughly 70% of the business — but those auto sales fell about 10% in 2025 as competitors rolled out newer models and pressure mounted on pricing and margins.
The big bet: robots, robotaxis and chips
What Tesla calls “physical AI” now sits at the center of its pitch to investors. That includes Optimus humanoid robots, a planned expansion of robotaxi services in U.S. cities, and even a long-range ambition to build a large domestic semiconductor facility — Musk’s so-called TeraFab — to avoid supply constraints and geopolitical risk.
On the investor call, Tesla’s CFO said the capex will be spread across factories and infrastructure projects, including battery and storage refineries, AI compute clusters, and work tied to the Cybercab and Semi. Musk has repeatedly framed Optimus and robotaxis as the source of most future value, but he and the company were careful to call Optimus still very much in R&D, with meaningful volume unlikely until later in the year at the earliest.
This is not happening in a vacuum. The wider tech world is likewise pouring money into agentic and multimodal AI — developments from companies like Google and Microsoft illustrate a fast-moving landscape where cloud models, image generation and task-oriented agents are part of the race to commercialize AI at scale. For context on how platforms are evolving, see coverage of Google’s AI Mode and Microsoft’s MAI-Image-1.
Why investors and customers might be nervous
There are stacked risks. Tesla’s car lineup is increasingly aged compared with rivals launching new EV models; brand perception has taken hits tied to Musk’s high-profile political activity and to safety questions around partially automated driving systems. Regulators and consumer groups are watching closely as Tesla removes human safety supervisors from some robotaxi pilots and continues to push toward driverless operations.
Competition looms in both arenas: Waymo and Baidu expand robotaxi services, chipmakers and robotics firms (from legacy chip foundries to startups) chase the same talent and capacity Tesla wants. Building a full semiconductor fab is a multiyear, multibillion-dollar endeavor and would be a new line of business entirely.
There are also reputational hazards tied to the xAI investment; the Grok chatbot has drawn scrutiny for outputs that mirrored Musk’s own polemic takes and for problematic deepfake content. Tying Tesla’s future to an AI firm that courts controversy raises corporate governance and public-relations questions.
Pockets of resilience
Not everything in the report is bleak. Energy storage revenue grew strongly last quarter, and Tesla’s gross margins in the fourth quarter posted notable improvement versus a year earlier — a sign the company can still squeeze efficiencies from the business that remains core today. Some Wall Street bulls point to robotaxis and scale-effect opportunities as justification for tolerating near-term profit declines.
Stock reaction has been mixed; the shares swung around the earnings release and investor commentary, reflecting the market’s uncertainty about whether Tesla can execute a sweeping transformation while shrinking its historical cash cow.
Tesla’s gamble is straightforward: accept weaker auto-led results now while investing aggressively in an AI-driven ecosystem that, if it works, could remake the company. The danger is equal parts technical difficulty and timing — ambitious factories and robot fleets only deliver if the software, sensors, supply chains and regulators all fall into place.
Either way, the Tesla that dominated headlines as a pure electric-car maker looks different this week than it did last. The company has signposted a future that places physical machines and compute at the center of its identity — and investors, regulators and consumers will be watching closely to see whether reality can match the vision.