Why is Tesla dismantling the Model S/X production lines to push forward the mass production of Optimus?

At the end of July, Tesla's 10 millionth all-electric vehicle globally rolled off the final assembly line at its Fremont, California plant. At this very same facility, the production lines of Model S and Model X, which once defined the high-end electric vehicle segment, ceased operation not long ago. In just 46 days, the old equipment on the automotive production lines was dismantled to make way for the Optimus humanoid robot production line.

Back in January this year, Elon Musk stated that Tesla would stop producing Model S/X and instead use the idle space at its California plant to manufacture the Optimus robot. When news of the dismantling broke, the first reaction from the outside world was astonishment. After all, the Model S is Tesla's first "truly independent vehicle" in the full sense, marking its transition from a manufacturer that modified Lotus chassis Roadsters to a mature automaker, and it was also the starting point for Musk to prove to the world in 2012 that "electric vehicles can beat fuel cars". Tearing down its production line seemed like an active erasure of a piece of history.

But if you expand the timeline, and piece together Tesla's ups and downs in the capital market over the past five years, the encirclement it has faced in the Chinese market, and Musk's obsession with the identity of an "AI robotics company", you will find that this is not an impulsive decision, but a math problem that has been calculated countless times.

2023: When the "automaker" narrative reached its limit

To understand the recent developments in Fremont, we have to go back to 2023 first.

In the summer of that year, Adam Jonas, a star analyst at Morgan Stanley, released a highly controversial report. His core view was straightforward: Tesla's future valuation should no longer be anchored by vehicle sales volume, but recalculated based on the "option value" of its AI business. He derived a large independent valuation for Tesla's "AI business", covering the Dojo supercomputer, FSD full self-driving, and the Optimus robot which was not taken seriously by most people at that time.

The report was dismissed by many people at the time. After all, Tesla in 2023 was caught in an unprecedented price war in the Chinese market, with its gross margin falling from a peak of over 25% in 2022 to around 15%. Musk cut prices while sending internal letters talking about the "severe economic environment", which hardly looked like the calm demeanor of an AI giant.

But the underlying logic of star analyst Adam Jonas hit the nail on the head: Wall Street votes with its feet far faster than people imagine. Fund managers have long been examining Tesla from a split perspective — they look at the profit statement and balance sheet for automotive business performance, but fill in AI figures in their valuation models. A price-earnings ratio of over 60 times would have been halved if compared with true automotive giants such as Toyota and Volkswagen; but this figure has held up because the market is "pre-paying" for a future that has not yet been realized.

In the valuation framework recently constructed by Alexander Potter, an analyst at Piper Sandler, he covered 17 product lines of Tesla and believes that the core business value is $400 to $420 per share. His judgment is quite straightforward: at a price of $400 per share, we believe investors get Optimus for free. The long-term value of Optimus and its supporting "inference as a service" business may eventually exceed the sum of all Tesla's current core businesses. But he did not forget to add a note: the execution risk in the short term remains very high.

The trouble is that this "pre-payment" has an expiration date.

Counting from Musk's announcement in 2016 that "full self-driving will be realized within two years", FSD has missed its launch schedule no less than four times. The release date of Robotaxi was pushed from 2024 to 2025, and then from 2025 to "coming soon". Every delay erodes the market's patience, and every missed deadline makes the AI valuation floating high above increasingly illusory. Andrew Percoco, an analyst at Morgan Stanley, voiced the general sentiment of institutional investors: Strong automotive and energy deliveries have improved near-term fundamentals, but we still believe that Robotaxi and Optimus will be the main drivers of the stock. Stronger investment confidence will require clearer evidence that Robotaxi is being scaled up, as well as more tangible Optimus verification points to support the return on investment of high capital expenditures. He gave Tesla a "neutral" rating with a target price of $417 — not too far, not too close, precisely reflecting a wait-and-see attitude of "wanting to see more evidence".

Capital needs a new anchor: a tangible "physical AI carrier" with visible, verifiable physical progress. This is the macro context for Optimus to make its debut. It is not a sci-fi project that popped up out of nowhere, but an inevitable "narrative relay baton" after the FSD and Robotaxi narratives became fatigued.

Musk understood this perfectly. In every earnings call and every investor day after 2023, his statements were systematically shifting: "Tesla is not a car company, we are an AI and robotics company." He may have said this a hundred times, but before 2023, most people took it as "Musk bragging again"; after 2023, when the pressure in the Chinese market became too real to ignore, people began to realize that this was probably not bragging, but him building public opinion groundwork for a fundamental strategic shift.

The rapid development of Chinese automakers,

Turned Tesla's "second growth curve" from an elective course to a required course

If the erosion of patience in the capital market is the internal cause, the changes in the landscape of the Chinese market are the most direct external driving force.

From 2023 to 2025, a quiet power transfer took place in the global new energy vehicle industry. BYD overtook Tesla for the first time in the fourth quarter of 2023 with 526,000 all-electric deliveries, taking the global top spot in all-electric vehicle sales. Although Tesla regained the quarterly first place by cutting prices in early 2024, the equation "Tesla = king of electric vehicles" has been broken, and possibly permanently.

What is more fatal is not a single competitor, but a qualitative change in the competitive landscape. Over the past decade, Tesla had three core advantages: the ultimate efficiency of its three-electric system, pricing power under the direct sales model, and brand premium brought by its "tech vibe". These three advantages were systematically dismantled and caught up with by Chinese automakers between 2023 and 2025.

The three-electric system? CATL's Qilin battery, BYD's Blade Battery, and XPeng's 800V high-voltage platform have at least matched or even surpassed Tesla at the parameter level. Pricing power? When the BYD Qin Plus was priced as low as 79,800 yuan and the Zeekr 007 was positioned directly against the Model 3, Tesla was forced to keep cutting prices, but the price cuts did not bring linear growth in sales, instead eroding its most precious gross margin. Brand premium? The "tech vibe" is rapidly becoming a "standard feature" rather than a "selling point" in front of the smart cockpits, urban NOA (Navigate on Autopilot) systems, and battery swapping networks of China's new EV startups.

An unavoidable fact is that Tesla, in China — its largest and once most profitable market — is being pushed into a situation where it wins acclaim but fails to achieve satisfactory market performance. For the whole of 2025, Tesla's share in China's new energy vehicle market has fallen below 5%, and while the absolute sales volume of Model 3 and Model Y remains considerable, their growth rate has lagged behind the overall market across the board.

This means that Tesla's path of relying on continuous growth in its automotive business to support its market capitalization is being called into question.

If you were Elon Musk, the situation you face would be as follows:

First, the gross margin of the automotive business is declining, the sales growth rate is slowing down, and competitive pressure will only keep growing;

Second, the valuation the capital market gives you is extremely dependent on the imagination of "non-automotive businesses";

Third, you happen to have a project in hand with extremely high technical homology, a potential market size possibly larger than that of automobiles, and almost no mature competitors — Optimus.

CITIC Securities gave a concise judgment in a recent research report: Tesla has both leading AI large model technology and large-scale manufacturing capacity, and the company is at the top tier of the global embodied intelligence industry chain, and it is firmly optimistic about the mass production and application prospects of Tesla's robots.

Once this logical chain holds, "whether to develop robots" is no longer a multiple-choice question, but a mandatory question that must be answered.

Phasing out S/X is just a simple arithmetic problem

Let's go back to the dismantled production line in Fremont.

The total annual delivery volume of Model S and Model X in 2025 is roughly estimated to be between 60,000 and 80,000 units, accounting for less than 4% of Tesla's total global deliveries. The S/X production line at the Fremont plant had a designed capacity of about 100,000 units per year back then.

A production line with a designed capacity of 100,000 units level that only achieves 60%-70% of its designed output, in a manufacturing enterprise that calculates "output per unit area" to the extreme, what does this mean? It means waste. Huge waste.

On July 23 Beijing time, Tesla released its Q2 2026 earnings report, which disclosed that the Optimus (humanoid robot) production lines are being advanced in parallel at Fremont, California and Austin, Texas, with planned annual capacities of 1 million units and 10 million units respectively.

This timing in itself speaks volumes. From the announcement of production shutdown in January to the launch of the new production line in July, only half a year passed. Musk himself said: If we can shut down a production line, dismantle the entire line, install a brand new production line, and put it into operation within four months, that will be an astonishingly fast speed. I don't think any other company on Earth has ever done that before.

A dedicated production line that maintains vehicles of the complexity level of Model S/X has an estimated annual fixed maintenance cost, special tooling depreciation, and dedicated material supply chain management cost of between 200 million and 300 million US dollars. More critically, it occupies the most central workshop location in the Fremont plant.

The Fremont plant is the "parent body" of all Tesla's manufacturing logic: the trial production of 4680 batteries is here, the process verification of the next-generation platform is here, and the initial trial production of Optimus is also here. In other words, every square meter of this plant is an extremely scarce resource. The roughly 20,000 to 30,000 square meters of space occupied by the S/X production line, if converted into the mass production potential of Optimus — even if the initial weekly output is only 100 units, at a long-term pricing of 25,000 US dollars per unit, the annual output value can reach the level of 1.3 billion US dollars. Moreover, the Optimus production line does not need to be built from scratch with great effort, because it overlaps with the electric vehicle production line in a large number of links.

The "overlap" here is not just "a little similar", but highly homologous.

Optimus has 28 body actuators (joints), and the core component of each actuator is a motor, a reducer, a sensor, and a controller. Translated into the language of automotive engineers, this is "small-size drive motor + precision reduction mechanism + position sensor" — which highly overlaps with the technical lineage of Tesla's motor drive systems, steer-by-wire systems, and electronically controlled brake systems.

Cross-verified based on information disclosed by Morgan Stanley and Tesla officials, the motor control algorithm and power electronics topology of Optimus have more than 70% technical homology with the traction motor and thermal management system of Tesla's vehicles.

But this does not mean that the road to mass production will be smooth, and Musk himself has a sober, almost ruthless understanding of this. When talking about the difficulty of mass production of Optimus, he said bluntly: Optimus will become the largest product in history, but also the most difficult to scale up mass production, because almost every component on the robot is brand new. Unlike electric vehicles that can rely on mature suppliers to provide tires, rearview mirrors, glass, and body panels, Optimus currently has no ready-made supply chain available, we have to build it from scratch, or turn to in-house production.

He used the term "S-curve" to describe the shape of the mass production ramp-up, and then added a meaningful sentence: Optimus will follow the normal S-curve of manufacturing growth, but since the parts are quite novel, the initial part of the S-curve will be quite flat and long. This means that dismantling the S/X production line is not "eliminating production capacity", but migrating the same group of engineers, the same set of supply chains, and the same manufacturing muscle from a product with diminishing marginal benefits to a product with potentially increasing marginal benefits. But this migration process is destined not to be a straight line.

Information from the supply chain confirms this. The lead screw on the robot's hand has been revised for almost 4 years before finalizing, and even the wire material inside the motor has to be remade. A person from Tesla's robot supply chain told the media that this dilemma of "every part has to be polished from scratch" is rarely encountered in traditional automobile manufacturing.

But the underlying logic of this arithmetic problem still holds. On one side is a luxury car with annual sales of 60,000 to 80,000 units, negative growth rate, and continuously shrinking profit margins; on the other side is a general-purpose robot whose BOM cost is expected to be pushed down to just over 20,000 US dollars, with long-term annual sales potentially reaching the million-unit level, and whose competition window has just opened. In the limited space of Fremont, the only correct answer to this multiple-choice question is to move the "old" out and bring the "new" in.

Musk put it more plainly: The real problem is that Optimus is a brand new product, with a brand new production line behind it, and there are more than 10,000 unique parts, none of which have really undergone large-scale mass production. The overall progress will eventually be dragged down by the luckiest, slowest, and least capable part among these 10,000 parts.

But he still chose to dismantle the S/X to make room for these 10,000 parts. This is not recklessness, but a precise calculation — not calculating "when Optimus will succeed", but calculating "when keeping S/X will start losing money".

Optimus Supply Chain: "We are betting against time"!

At the Tesla project progress meeting at the end of June 2026, Musk threw a tough statement in front of the Optimus team: he required the team to hit the production capacity target before the end of the year, otherwise he would fire the entire Optimus procurement team.

This is not the first time Musk has used tough language in an internal meeting, but when placed on the Optimus project, the weight is particularly different. Because the mass production of Optimus is not something that Tesla can complete on its own, it relies on the synchronous speedup of the entire supply chain. The guidance figures Tesla gave to suppliers are quite aggressive: suppliers are required to increase production capacity to 1,000 units per week in September, and to 2,000-2,500 units per week by the end of the year.

What does this figure mean? If the weekly output of 2,000 units can really be achieved by the end of the year, the annualized production capacity will be 100,000 units — in a brand new segment, from zero to 100,000 units, in only half a year.

But the real feedback from the supply chain is much more cautious. It was previously said that the weekly production capacity should reach 1,000 units in July, but now it has been postponed to September. The same group of supply chain insiders gave a more practical judgment: the year-end production capacity may not reach 100,000 units.

Even so, the actions of the suppliers have spoken for themselves. According to reports from *Caijing* and multiple media outlets, some Chinese suppliers have received component procurement orders for Tesla's third-generation Optimus robot. Suppliers stated that in order to meet Tesla's production capacity requirements, their factories have allocated 70% of their production capacity to the Optimus project, and are building new plants to expand capacity. This kind of "betting-style" production capacity tilt is not common in the manufacturing industry.

However, this gives the supply chain and the capital market more confidence than any order and expectation guidance, and it is not an exaggeration. The mass production of Optimus is not just a matter for one company, it affects an entire supply chain that is being built from scratch. These suppliers are willing to bet 70% of their production capacity on it, not because they see the current orders for Optimus, but because they believe Musk