[News Source] As learned from 36Kr, Tesla announced via its official Weibo account recently that its third-generation humanoid robot is about to make its official debut. This fully restructured product has set a mass production target, aiming to achieve a breakthrough of 1 million units in annual output, marking a crucial step for the humanoid robot industry from laboratory prototypes to commercial large-scale deployment.

[Source of Core Data and Information] The core technical parameters, mass production plans, cost targets and other information in this article are comprehensively derived from Tesla’s official announcements, 36Kr reports and public technological industry analysis materials. Among them, the production capacity plan, autonomous learning capabilities and other core information are all disclosed by Tesla’s official Weibo account, while technical details and commercial layout refer to public industry interpretations and disclosures of Tesla’s supply chain-related information[superscript:2].

Reconstructed from First Principles, Unlocking Core Breakthroughs of Humanoid Robots

The most core upgrade of Tesla’s newly announced third-generation humanoid robot (Optimus V3) lies in its “comprehensive redesign based on first principles” — unlike the previous two generations of prototypes focused on technical verification, this product completely breaks free from the constraints of the existing supply chain system, with an 80% update rate of components, achieving dual subversive upgrades in hardware and intelligent systems[superscript:2].

At the hardware level, the third-generation robot has significantly improved mobility and operational accuracy. The number of joints in the whole body has increased to 37, and the hand retains a 22-degree-of-freedom dexterous hand design, with an operation error controlled at the sub-millimeter level. It can complete delicate tasks such as picking up eggs and precision sorting of batteries, and even catch a thrown tennis ball with one hand and place it accurately, with dexterity approaching that of humans[superscript:1][superscript:4]. At the same time, its walking speed has increased by 30% compared with the previous generation to 1.2 meters per second, enabling stable walking on a 15-degree slope and recovery of balance within 0.3 seconds after being subjected to thrust interference, completely solving the problem of unstable walking of the previous generation. Its load capacity has doubled to 20 kilograms, and its battery life has been extended to 8-10 hours, which can meet the needs of all-weather industrial operations[superscript:1].

The breakthrough at the intelligent level is even a core highlight of this robot, and also the core support for Tesla’s official emphasis on “learning new skills by observing human behavior”. The third-generation robot is deeply integrated with Tesla’s FSD autonomous driving technology and empowered by the Dojo supercomputer, which greatly improves autonomous learning efficiency and environmental perception capabilities[superscript:1]. It is equipped with an FSD chip homologous to Tesla’s automobiles, and cooperates with 8 pure vision cameras to achieve 360-degree environmental perception. It can complete accurate environmental recognition without lidar, process 1,000 frames of high-resolution images per second, and its response speed is faster than the average level of humans[superscript:4].

More importantly, through end-to-end neural network deployment, this robot can independently learn new skills by observing human demonstrations, receiving verbal descriptions or watching videos, without complex programming. The training cycle for a single action has been significantly shortened from 48 hours to 2.5 hours[superscript:1][superscript:3]. Moreover, the skills it has learned can be synchronized to all the same models of robots around the world through the cloud, realizing “collective evolution”, greatly improving the efficiency of scene adaptation, and truly getting rid of the dependence of traditional robots on customized programming[superscript:1].

Million-Unit Production Is Not an Empty Promise: Tesla’s Three Pillars of Confidence

Tesla’s bold goal of “1 million units per year” is not just a strategic declaration, but based on three pillars of confidence: technology reuse, supply chain integration and production line transformation. The core is to solve the industry pain points of humanoid robots being “difficult to manufacture, expensive to produce and hard to mass-produce”[superscript:3].

First, the reuse of automotive industry technology and production lines shortens the mass production ramp-up cycle. Tesla is transforming the Model S/X production line at its Fremont factory into a dedicated production line for Optimus, planning to start the transformation in the second quarter of 2026 and officially launch production by the end of the year[superscript:2][superscript:3]. Relying on mature mass production experience in the automotive manufacturing field, there is no need to build a production line from scratch, which can quickly achieve production capacity ramp-up. The initial plan is to produce 50,000 to 100,000 units per year, increase to 500,000 units in 2027, and sprint for the long-term goal of 1 million units per year[superscript:1][superscript:2].

Second, vertical supply chain integration + domestic empowerment greatly reduce costs. By reusing supply chain resources such as automotive batteries, motors and chips, R&D and manufacturing costs are amortized; at the same time, it is deeply bound to China’s core supply chain. At present, the localization rate of Optimus core components has exceeded 60%, and leading Chinese enterprises such as Tuopu Group, Sanhua Intelligent Controls and Green Harmonic cover key links. Among them, Sanhua Intelligent Controls monopolizes 70% of the actuator supply, and Green Harmonic has broken the monopoly of Japanese harmonic reducers, greatly reducing the cost of core components[superscript:3]. Thanks to these measures, the target unit cost of the third-generation robot has been reduced to less than 20,000 US dollars, a decrease of more than 70% compared with the 70,000 US dollars of the first-generation prototype, breaking the cost bottleneck for the commercialization of humanoid robots[superscript:1][superscript:3].

Third, scenario-based deployment first, with demand supporting mass production scale. Tesla has formulated a “industrial first, household follow-up” deployment strategy for the third-generation robot. The first batch of products will be prioritized for deployment in Tesla’s own factories, undertaking repetitive tasks such as component handling, battery assembly and quality inspection, which is expected to increase factory yield by 15% and reduce work-related injuries by 90%[superscript:1]. At present, it has won an order for 20,000 logistics sorting robots from Amazon. In 2026, it will expand to commercial scenarios such as catering and retail. After being launched to the public in 2027, it will gradually enter C-end fields such as home services and care. In the long run, it will also cover scenarios that are difficult for humans to access, such as mineral mining and disaster rescue. Diversified scenario demand provides solid support for million-unit mass production[superscript:1].

More Than Tesla’s Transformation, It’s a Paradigm Shift for the Industry

For Tesla, the mass production of the third-generation humanoid robot is a crucial step in its transformation from an automaker to an AI robot company — Elon Musk has clearly stated that Optimus will carry more than 80% of Tesla’s long-term value, and is expected to promote Tesla’s transformation into a technology giant with a valuation of 25 trillion US dollars, whose value will far exceed the existing automotive business[superscript:1][superscript:2]. If the million-unit mass production target is successfully achieved, Optimus will become Tesla’s third core business after automobiles and energy storage, with commercial potential exceeding 10 trillion US dollars[superscript:3].

For the entire humanoid robot industry, Tesla’s entry and mass production breakthrough will completely restructure the industry’s development paradigm. For a long time, the humanoid robot industry has been trapped by problems such as immature technology, high costs and difficult mass production, mostly staying in the stage of laboratory prototypes or small-batch pilots. Relying on its advantages in mass production capacity, AI technology accumulation and cost control in the automotive industry, Tesla has built an insurmountable moat in the industry[superscript:1].

More importantly, the debut and mass production plan of the third-generation robot will inject a strong catalyst into the humanoid robot industry chain, driving the rapid development of related fields such as core components, AI algorithms and human-computer interaction[superscript:1][superscript:2]. If million-unit mass production is realized, it will profoundly change the labor structure of industrial production and social services, replacing humans in high-intensity, high-risk and repetitive work. A silicon-based productivity revolution led by Tesla is quietly kicking off[superscript:1][superscript:3].

Opportunities and Challenges Coexist: The Commercialization of Humanoid Robots Still Has a Long Way to Go

Although Tesla’s plan is full of imagination, the commercialization of the third-generation humanoid robot still faces multiple challenges. At the technical level, problems such as insufficient durability of dexterous hands and the need to improve tactile perception accuracy still need to be broken through — in actual tests, the service life of dexterous hands is only about 6 weeks, the accuracy of transmission components will decrease significantly after long-term use, and the resolution of tactile sensors is still an order of magnitude lower than that of human skin[superscript:4].

At the mass production level, the supply chain supporting facilities still need to be improved. The stability of large-scale mass production of core components such as dexterous hands and the efficiency adaptation after production line transformation are all problems that Tesla needs to solve[superscript:2]. In addition, issues such as the safety and reliability of robots, as well as ethical norms and industry standards, will gradually become prominent with large-scale deployment, which need to be explored and solved jointly by enterprises and the industry[superscript:2].

However, it is undeniable that Tesla’s entry has broken the development stalemate of the humanoid robot industry. From the first-generation walking prototype in 2022, to the second-generation completing factory tests in 2023, and now the third-generation targeting mass production, Tesla has completed the leap of humanoid robots from concept to commercialization in only four years[superscript:1].

With the official debut of the third-generation humanoid robot and the gradual advancement of the million-unit mass production plan, humanoid robots are moving from “science fiction” to “reality”. Tesla’s ambition is not only to create a subversive product, but also to define a new industrial track — when silicon-based labor gradually replaces carbon-based labor, the production and lifestyle of human society may usher in an unprecedented transformation.

References: 1. 36Kr Report “Tesla’s 3rd-Gen Humanoid Robot to Debut, with Annual Production Expected to Reach 1 Million Units”; 2. Disclosures from Tesla’s Official Weibo Account; 3. Public Industry Analysis and Supply Chain Information Disclosures[superscript:1]-[superscript:4].