Tesla Doesn't Want to Build Cars Anymore
Tesla did not 'go all in on robots' with a wave of the hand: it killed the Model S and X not because robots won, but because the EV profit story can no longer be told. A 4.6% operating margin cannot support a trillion-dollar valuation. A $25 trillion pie can.
- 01THE EARNINGS — The Q4 call: revenue down, profit halved — and, on the same call, the end of Model S/X production announced.
- 02THE RETOOLING — Fremont lines converted to Optimus and Cybercab production; $20 billion in capex switches tracks.
- 03THE GAP — Musk sets summer mass production; supply chains whisper 15,000 units; the company admits output is roughly zero.
WHAT HAPPENEDWhat Happened
On January 28, 2026, Tesla's Q4 earnings call: full-year 2025 revenue of $94.8 billion, down 3% year on year; net income of $3.8 billion, down 46%; operating margin shrunk from a peak of about 17% to 4.6%. On the same call, Musk announced the end of Model S and Model X production — the two lines at the Fremont factory will be converted to build the Optimus humanoid robot and the Cybercab.
Killing two flagship models is unheard of in automotive history. But the numbers explain everything: when selling cars earns a 4.6% margin, while robots in Musk's telling are a '$25 trillion market opportunity', the production lines must make way for the new story. This is not a heroic amputation; it is sacrificing the rook to save the king.
THE MARGIN COLLAPSE4.6%: the end of the EV profit story
Tesla was once the world's most profitable carmaker, with operating margins near 17% that Toyota and Volkswagen could only look up at. In 2025, that number is 4.6%. Under the triple squeeze of price wars, tariffs and peaking demand, electric vehicles have gone from money printers to a thin-margin business.
The trend is crueller still: $94.8 billion in revenue is still shrinking, and $3.8 billion in net income has halved. When 'selling cars' is no longer sexy, capital markets will not award a high valuation for a 4.6% margin. Tesla must tell a new story — and Optimus and the Cybercab are its two protagonists.
THE PRODUCTION GAP15,000 units versus zero: the production Rashomon
At March's Abundance Summit, Musk confirmed Optimus V3 would 'begin mass production in the summer'; supply-chain rumours put 2026 production at about 15,000 units, with weekly output near 1,000 by end-September and a long-term Fremont target of 1 million a year. But on the July Q2 earnings report, Tesla itself admitted: the first-generation line is being installed, cumulative output is still roughly zero, and the first machines are going to the internal 'Optimus Academy' for training data.
15,000 units in words, roughly zero on the books. That is the state of Tesla-style production: targets always grand, deliveries always 'next year'. The specs look beautiful — 173cm tall, 57kg, 22 degrees of freedom in one hand, the AI5 chip, Grok voice, a target price of $20,000-30,000 — but specs are not products. Shipments are.
THE BET$20 billion: the most expensive pivot in history
Tesla's 2026 capex exceeds $20 billion, a record, going to six new or converted production lines and Giga Texas's Cortex AI training cluster (about 85,000 Nvidia H100/B200 GPUs). In essence, this money switches all of a car company's capital expenditure onto the AI and robotics track.
The logic is called 'dual-robot platform economics': the Cybercab and Optimus share the same AI5 chip, FSD vision stack and Cortex compute. In the first half of 2026, within weeks of each other, both crossed from 'pre-production' into 'delivery' — the Cybercab joined Austin's public robotaxi fleet on September 4. One platform, two products, amortising the same R&D cost.
THE SHIFTWhat Changed
SIGNAL JUDGMENT — ANALYSISThe carmaker is dead; long live the robot
When a car company's $20 billion in capital expenditure all flows to AI and robotics, and the lines of its flagship models are converted into robot lines, the sentence 'Tesla is a car company' is already past tense. Only one question remains: can the new story's delivery speed outrun the old story's collapse?
WHAT REMAINS UNCERTAINWhat Remains Uncertain
EVIDENCEEvidence
The S/X discontinuation and line conversion come from official disclosure on the January 28, 2026 Q4 earnings call; the 15,000-unit figure is a supply-chain rumour (not official), and Q2 acknowledged cumulative output of roughly zero.
PRIMARY SOURCESPrimary Sources
- Tesla Q4 2025 earnings call (2026-01-28)
- Tesla Q2 2026 earnings: Optimus production line being installed
- FinancialContent/MarketMinute: Analysis of the $20 billion AI gamble
- markets.financialcontent.com: Original article
https://markets.financialcontent.com/prnews.pressrelease/article/marketminute-2026-2-2-teslas-20-billion-ai-gamble-inside-the-pivot-from-automaker-to-robotics-titan
RELATEDRelated Intelligence
- Agility Robotics Debuts on Nasdaq as AGLT, First U.S. Public Pure-Play Humanoid Maker
- Agility Robotics Closes SPAC Deal, Set to Trade on Nasdaq as AGLT on November 5
- Uber Books $59.8 Billion in Q3 and Puts a Number on Autonomy: 700 Driverless Vehicles by Year-End
- South Korea opens its first humanoid robot data factory: 12 robots do the work first, and bank the data