Tesla Is Killing the Model S and X to Make Room for Optimus. Figure Is Already on the Factory Floor.

Tesla will end Model S and Model X production this quarter and repurpose part of its Fremont plant for humanoid robots. The bet is enormous. The execution risk is bigger, and a quieter rival has already shipped what Tesla is still promising.

Tesla confirmed on its Q4 2025 earnings call that the Model S sedan and Model X SUV will exit production in Q2 2026, with the freed-up Fremont capacity redirected toward building the Optimus humanoid robot. The two flagship vehicles accounted for a small slice of Tesla's 2025 output. For the full year, Tesla delivered roughly 1.64 million vehicles, of which approximately 50,000 were classified as "other models" (Model S, X, and Cybertruck combined) — meaning S and X together represented well under 3% of the total. The financial sting is small. The strategic message is not. Elon Musk framed it bluntly: the company is winding down the legacy car programs and "moving into a future based on autonomy."

That sentence is the whole story. Tesla is no longer a car company that happens to build robots; it is positioning itself as a robotics company that still happens to build cars. The aspiration is staggering — Musk has floated annual Optimus capacity of one million units, with a Gen 3 reveal slated for Q1 2026 and a new hand design meant to fix the dexterity problem.

Here is where the gap between narrative and reality opens up.

The numbers Tesla doesn't lead with

Tesla promised 10,000 Optimus units in 2025. It delivered a few hundred. The 2026 target sits at 5,000 — already a climbdown — and even that assumes a smooth conversion of a high-volume automotive line into a humanoid assembly line, a manufacturing problem nobody has solved at scale.

More telling was Musk's own admission: Optimus "is not in usage in our factories in a material way," and the robots that exist are there mostly so the system "can learn." For a company whose valuation leans heavily on the robot becoming a labor product, that is a remarkable concession. Repurposing Fremont is a capital and engineering bet placed on a machine that, by the CEO's account, does not yet do useful work.

Figure already did the boring part

While Tesla reorganizes its factory around a promise, Figure quietly executed the unglamorous version of the same vision. Its Figure 02 robot worked an 11-month deployment at BMW's Plant Spartanburg, handling more than 90,000 sheet-metal parts across roughly 1,250 hours and contributing to the production of over 30,000 X3 vehicles. BMW's expansion to European factories in early 2026 — using Hexagon Robotics' AEON humanoid at Plant Leipzig — confirmed that real industrial customers are now willing to put humanoid robots on active production lines, not just in labs. Figure's Spartanburg data is feeding into its next-generation Figure 03 platform, with further BMW deployments under evaluation.

That is the distinction that matters. Figure has a paying industrial customer and logged hours on a real line. Tesla has a target, a stage reveal, and a factory it intends to convert. One company is iterating on deployments; the other is still iterating on learning.

The risk for Tesla isn't ambition — it's sequencing. Tearing out a working car line to chase a million-robot future only pays off if Optimus crosses from "can learn" to "can work" fast. Figure's BMW hours suggest that crossing is hard, slow, and won validation by doing it, not announcing it.

Fontes

• https://blog.robozaps.com/b/tesla-model-s-optimus-robot-factory-conversion
• https://newmarketpitch.com/blogs/news/humanoid-robotics-figure-vs-tesla