Tesla (3Q25 Minutes): The release and mass production of Optimus 3 have been delayed to 2026.
The following are the minutes of the Tesla FY25Q3 earnings call organized by Dolphin Research. For an interpretation of the financial report, please refer to "Tesla: Relying on the Stars and the Sea, It's Time to Pay for Faith Again"
Tesla Minutes Summary:
1. Automotive Business:
The existing capacity will reach an annualized production rate of 3 million units within 24 months (possibly sooner), with the core increment coming from the Cybercab model, which will start production in the second quarter of 2026. This model is optimized for full autonomous driving, with no steering wheel or pedals, focusing on minimizing operational costs.
On the demand side, no additional stimulus is needed. The core selling point is "unsupervised autonomous driving that can be fully operated via mobile phone," which will drive users to purchase proactively. In the short term, profit margins will not be sacrificed, and in the long term, economies of scale and technological premiums will continue to enhance profitability.
1. Robotaxi Business Progress:
a. Robotaxi operational data continues to improve, with Austin achieving operations without safety drivers, accumulating over 250,000 miles; the Bay Area still has safety drivers, with over 1 million miles driven.
b. It is expected that safety drivers will be fully removed in Austin by the end of the year, and expansion will continue to 8-10 metropolitan areas, including Nevada, Florida, and Arizona.
c. The core algorithms and architecture of Robotaxi and the consumer version of FSD are consistent, with only minor functional differences (such as the consumer version's autonomous parking mode, which is not necessary for Robotaxi).
2. FSD Progress:
a. To date, the cumulative mileage of users using supervised FSD has reached 6 billion miles, with excellent safety performance.
b. The core obstacle to deploying unsupervised FSD is regulatory approval, and we are currently communicating with regulatory agencies in regions such as China and Europe;
c. On the technical side, version V14.1 has been opened for U.S. users to experience, and users can activate it by selecting "Advanced Software" in the vehicle settings, with the hardware equipped with the HW4.0 chip.
d. FSD V14 adopts a new software architecture, prioritizing safety. The initial version may have issues with operational smoothness, which is planned to be optimized in version V14.2.
e. Subsequent FSD version upgrades will include inference capabilities (expected to be implemented in versions V14.3 or V14.4), enabling intelligent decision-making such as optimal parking space selection and automatic drop-off and parking, while enhancing reinforcement learning through the Tesla Reality Simulator, with parameter scale increasing by an order of magnitude.
f. AI5 Chip: For the next-generation AI 5 chip, both TSMC and Samsung will be commissioned for production. Our AI 5 chip features a revolutionary design. By deeply optimizing the entire software and hardware architecture, we have removed redundant modules such as traditional GPUs and image signal processors from the design. AI 5 is expected to achieve a 2-3 times performance improvement and a 10-fold breakthrough in cost-effectiveness.
3. Humanoid Robot Progress
a. Prototype Launch and Mass Production Plan: The Optimus V3 prototype will be showcased in the first quarter of 2026 (expected in February-March), with a highly realistic appearance, resembling a "human in a robot suit"; mass production is planned to start by the end of 2026, with a production line capable of producing 1 million Optimus units annually, aiming to start mass production around the end of next year.
However, reaching an annualized capacity of 1 million units will require a ramp-up process, as the speed of progress will depend on the slowest, most challenging, and "uncooperative" component among the 10,000 unique parts. Ultimately, achieving an annual production of 1 million units is certain.
Looking further ahead, we will eventually launch Optimus 4, with an annual production target of 10 million units; Optimus 5's annual production could even reach 50 million or 100 million units.
The hardware design will not be frozen before mass production starts and will be continuously iterated and optimized during production, as some manufacturing challenges can only be discovered and resolved during the mass production phase.
b. Core Challenges in Launching Optimus:
The core challenges are concentrated in three dimensions: first, hand dexterity, as the design of human hand freedom, muscle strength, finger length, etc., is highly complex, and the engineering difficulty of the forearm and hand actuators exceeds the sum of the rest of the robot;
second, supply chain and manufacturing, as humanoid robots have no existing supply chain, requiring Tesla to vertically integrate and produce core components, with the goal of achieving a scale of 1 million units annually, presenting significant manufacturing challenges;
third, large-scale training, as real-world AI technology from the automotive side needs to be transferred to robots, while completing multi-dimensional model training for general mobility, specific tasks, and voice interaction. Currently, the Optimus robot can autonomously move at the Palo Alto engineering headquarters and respond to visitor navigation requests.
4. Energy Business
a. Strong Demand for Energy Storage
Demand for Megapack and Powerwall remains strong, with orders booked until next year; the Megablock product has received high recognition from customers and will start shipping from the Houston factory in 2026.
Demand in the AI and data center fields has grown significantly, with Megapack becoming an important choice for large-scale data centers and utility companies due to its advantages in enhancing grid reliability and alleviating grid constraints.
U.S. residential solar demand has surged due to policy adjustments, with a new leasing product to be launched in the first half of 2026. Residential solar panels produced at the Buffalo factory will be delivered in the first quarter of 2026, with industry-leading aesthetic design and performance, demonstrating Tesla's commitment to American manufacturing.
b. Energy Business Facing Tariff Pressure, Shanghai Megapack Factory Partially Avoids Tariff Costs
The energy storage business is currently facing increased competition and tariff pressure, as battery cell procurement mainly comes from China, with tariffs having a higher cost impact on this business than on the automotive business. In the third quarter, the combined tariff impact on both businesses exceeded $400 million, with each business accounting for about half. After the capacity ramp-up of the Shanghai Megapack factory, it has started supplying non-U.S. markets, effectively avoiding some tariff costs.
1. Review of Core Financial Information
2. Detailed Content of the Conference Call
2.1 Key Information from Executive Statements
Elon Musk's Speech:
1. Strategic Turning Point and Leadership in Artificial Intelligence
We believe that Tesla is at a critical turning point as we bring artificial intelligence into the real world. I must emphasize that Tesla is the absolute leader in real-world AI, unmatched by anyone. Our cars have the highest "intelligence density" among all AI, and this advantage continues to grow.
Key business developments include: the large-scale launch of Full Self-Driving (FSD) and Robotaxi, which will fundamentally change the nature of transportation; the energy storage business (Powerwall, Megapack) can significantly enhance energy output efficiency through grid energy buffering; the Optimus humanoid robot is expected to become the largest product in history, with human-like movement and interaction capabilities.
2. Full Self-Driving and Capacity Expansion
We are at the starting point of a large-scale expansion of Full Self-Driving (FSD) and Robotaxi, which will fundamentally change the nature of transportation, with momentum like a "shockwave." Since we have a clear understanding of achieving unsupervised full self-driving, I am now confident in expanding Tesla's capacity, with the goal of expanding future production "as quickly as reasonably possible."
3. Impact of the Energy Business
We are having a significant impact on the energy sector through battery storage, particularly Megapack. By using batteries for energy buffering, we can effectively double the energy output of the grid without building new power plants. We have announced the Megapack 4 plan, which will integrate substation functions and directly output about 35 kV of voltage, greatly enhancing deployment capacity and speed.
4. Potential of the Optimus Humanoid Robot
Optimus is on the verge of a major breakthrough, with the potential to become "the greatest product ever." We have all the core elements needed to achieve it: real-world AI, excellent electromechanical engineering capabilities, and large-scale production capabilities. We look forward to showcasing Optimus V3 in the first quarter of next year, with unprecedented realism, looking like "a person in a robot suit."
CFO's Speech:
1. Quarterly Performance
The third quarter performed excellently on multiple levels, setting new records in delivery volume and deployment volume, as well as achieving breakthroughs in total revenue, energy business gross profit, energy profit margin, and free cash flow. This achievement is due to customers' continued confidence in the company's products and the relentless efforts of the Tesla team.
2. Delivery Performance in Various Regional Markets
All regions showed strong growth: Greater China: 33% quarter-on-quarter growth, Asia-Pacific: 29% quarter-on-quarter growth, North America: 28% quarter-on-quarter growth, Europe, Middle East, and Africa: 25% quarter-on-quarter growth
This growth momentum is mainly due to the continued hot sales of the new Model Y, fulfilling the company's promise of "2025 being the year of Y," with the launch of the new Model Y, long-range version, performance version, and recently the standard version in North America and Europe, Middle East, and Africa.
3. Robotaxi Business Progress
Robotaxi services are currently operating in Austin and most Bay Area cities. Since its launch, the service area in Austin has tripled and plans to continue expanding. Unlike competitors, Tesla's Robotaxi fleet integrates well into the operating environment, as it does not require additional sensors, making it unobtrusive. It is expected that with the large-scale experience of the supervised version of FSD, vehicle demand will increase significantly.
4. FSD Adoption Status and Plans
About 12% of the existing fleet are paying FSD customers, with the scale still small but progressing well.
We are working with regulatory agencies in China, India, and other regions to seek approval for FSD operations.
5. Financial Performance of Each Business Segment
Automotive Business:
Revenue increased by 29% quarter-on-quarter, consistent with delivery volume growth, while regulatory credit revenue decreased quarter-on-quarter. Excluding credits, the automotive profit margin increased from 15% to 15.4%, mainly due to improvements in material costs and increased production volume leading to improved fixed cost absorption efficiency.
Energy Storage Business:
Record highs in deployment volume, gross profit, and profit margin, but facing significant tariff pressure (currently all battery cells are procured from China). The capacity increase at the Shanghai Gigafactory helps avoid tariffs, as the factory mainly supplies non-U.S. market demand. In the third quarter, the total tariff impact on the automotive and energy businesses exceeded $400 million, with both businesses equally affected.
Service and Other Business
Significant quarter-on-quarter improvement, mainly due to progress in the insurance business and the development of service center business
Robotaxi operating costs are included in this business segment, along with paid supercharging, used car business, parts and merchandise sales, and other businesses.
6. Cost and Expense Analysis
Operating expenses increased quarter-on-quarter, mainly due to: restructuring and related expenses (aimed at cost reduction and efficiency improvement and AI chip design integration), expenses related to specific legal cases, incremental costs for shareholder meeting preparation, increased employee-related expenses, especially R&D expenses due to stock awards granted to AI project employees, and this trend is expected to continue.
7. Other Financial Indicators
Other income decreased quarter-on-quarter, mainly due to a significant reduction in gains from changes in the market value of Bitcoin holdings (Q3 was $80 million, Q2 was $284 million) and foreign exchange changes. Cash flow was approximately $4 billion, setting a new high. The total cash and investment at the end of the period exceeded $41 billion
8. Capital Expenditure Outlook
Capital expenditure for 2025 is expected to be approximately $9 billion
It will increase significantly in 2026, mainly to support the development of existing businesses and investment in AI-related plans (including the Optimus project)
2.2 Q&A
Q: What are the latest Robotaxi metrics? Fleet size, cumulative mileage, completed rides, intervention rate, and when will safety drivers be removed? What obstacles remain for deploying non-regulatory FSD to customer vehicles?
A: In terms of Robotaxi business deployment, we are steadily advancing the removal of safety drivers. The company expects to achieve fully driverless operations in most of Austin's operating areas by the end of this year. We plan to achieve this goal in at least part of Austin in the coming months. For safety reasons, we are highly cautious about deployment, as any accident could become global headline news, so we choose to take a more prudent approach.
We plan to launch Robotaxi services in about 8-10 metropolitan areas by the end of the year. The specific expansion progress will depend on regulatory approval in each region, and we have already submitted relevant applications in Nevada, Florida, and Arizona, which are publicly available.
In terms of actual operational data, our driverless fleet in Austin (no safety driver in the driver's seat) has accumulated over 250,000 miles. In the Bay Area, due to regulatory requirements, we still have safety drivers in the driver's seat, but the mileage has exceeded 1 million miles.
It is worth emphasizing that the total mileage of customers using the supervised version of FSD has recently reached 6 billion miles, which is an important milestone. We will gradually remove in-car safety drivers according to the plan, starting with Austin.
Q: In the current AI boom, what is the demand and backlog situation for Mega Pack, Powerwall, solar, or energy storage systems? Does Tesla plan to power other large-scale data centers?
A: Demand for Megapack and Powerwall remains strong, with orders booked until next year. Our Mega Block product has received positive feedback from customers and will start shipping from the Houston factory next year.
With the development of AI technology, the demand for energy storage products in the data center field has grown significantly. Large-scale data centers and utility companies increasingly recognize the value of Megapack in enhancing grid reliability and alleviating power supply pressure.
In the U.S. market, policy changes have driven a surge in residential solar demand. We have launched a new solar leasing product, and this trend is expected to continue into the first half of 2026. A new leasing product will be launched in the first half of 2026, and residential solar panels produced at the Buffalo factory will be delivered in the first quarter of 2026, with industry-leading aesthetic design and performance, demonstrating Tesla's commitment to American manufacturing.
Q: What existing challenges does Optimus face in going to market?
A: Bringing Optimus to market is an extremely difficult task. It must be clear that this is by no means easy. Currently, Optimus can autonomously operate 24/7 at our Palo Alto engineering headquarters, providing guidance services for visitors. However, I do not intend to underestimate the difficulty involved.
The biggest engineering challenge lies in the development of the hand and forearm. Creating a mechanical hand as dexterous as a human hand is extremely difficult. The human hand is a sophisticated biological structure, and after in-depth research, we found that the configuration of four fingers and a thumb, the different degrees of freedom of each finger, differentiated muscle strength, and finger length have all undergone precise evolution. Therefore, the hand and forearm of Optimus (with actuators mainly concentrated in the forearm area) constitute a huge electromechanical engineering challenge, technically more complex than all other parts of the robot combined.
To achieve the practical value of a general-purpose robot, it must have such precise hands, as well as mature world AI technology and large-scale production capabilities. Manufacturing only a few hundred robots would be meaningless. We must achieve a manufacturing scale comparable to or even higher than that of automobiles, such as an annual production of 1 million units. This presents a huge manufacturing challenge, as the entire humanoid robot field has not yet established a mature supply chain system. Unlike the automotive and computer industries, we must deeply vertically integrate and independently produce most of the components.
Third, large-scale training is needed to transfer real-world AI technology from the automotive side to robots, while completing multi-dimensional model training for general mobility, specific tasks, and voice interaction. Currently, the Optimus robot can autonomously move at the Palo Alto engineering headquarters and respond to visitor navigation requests.
Based on these challenges, I believe Tesla has unique advantages in manufacturing technology, world AI, and dexterous hands. This is our core development path. However, I have a fundamental concern: if I do not have enough voting control, will I be excluded from the decision-making level in the future after establishing this robot army? This is actually the only reason I care about voting rights. The core of the issue is that I must ensure sufficient influence (not absolute control) over this robot army, otherwise, I will not be able to proceed with this project with peace of mind.
Q: Can you provide the latest information on the $16.5 billion chip deal with Samsung in Taylor? Considering the importance of semiconductors to the future of autonomous driving and Tesla AI, what confidence do you have that Samsung can complete AI chip production according to Tesla's timeline and achieve relatively better yield and cost than TSMC?
A: Regarding the chip cooperation with Samsung, we highly recognize its performance. Samsung is currently responsible for producing our AI 4 autonomous driving chips and has demonstrated excellent manufacturing capabilities. For the next-generation AI 5 chip, I need to clarify an important decision: we will commission both TSMC and Samsung for production.
Our AI 5 chip features a revolutionary design. By deeply optimizing the entire software and hardware architecture, we have removed redundant modules such as traditional GPUs and image signal processors from the design. This minimalist design allows the AI 5 chip to achieve a 40-fold performance improvement over AI 4. Due to our full-stack understanding of the autonomous driving system, we can precisely define the functions required by the chip while decisively discarding unnecessary components.
In terms of manufacturing layout, both foundries will produce in the U.S.: TSMC in Arizona and Samsung in Texas. It is worth mentioning that the Samsung factory is equipped with slightly more advanced equipment. Our strategic goal is to achieve sufficient AI 5 chip capacity, even "excess supply," as surplus capacity can be flexibly allocated to data center use.
Compared to NVIDIA, we have a unique advantage. Tesla only needs to meet its own needs, allowing chip design to pursue extreme simplicity. By simplifying the complex interconnections between logic modules, we have significantly improved energy efficiency. AI 5 is expected to achieve a 2-3 times performance improvement and a 10-fold breakthrough in cost-effectiveness.
Although these expectations still need to be verified through actual mass production, we believe that this design philosophy focused on specific needs will bring significant competitive advantages. Our existing data centers already use both AI 4 and NVIDIA chips, and the surplus capacity of AI 5 will be deployed in this mode in the future.
Q: Why not prioritize providing replacement incentives for Hardware 3 users, but instead promote the upgrade from Hardware 3 to Hardware 4? How will you ensure the rights of Hardware 3 users in the future?
A: In terms of hardware upgrade strategy, we have not abandoned the HW3 platform. Over the past year, we have provided customers with the option to transfer FSD functionality to newly purchased vehicles and offered more attractive promotional schemes for FSD users during specific periods.
We fully recognize the importance of early users, who are crucial to our development. It is worth noting that my daily commuting vehicle is a model equipped with HW3 hardware, and I use the FSD function every day. We are committed to continuing to support this group of users.
Currently, the focus is on breakthroughs in unsupervised FSD technology, and a streamlined version of V14 adapted for Hardware 3 will be launched later, expected to be implemented in the second quarter of 2026. Hardware 3 users are the core early users of FSD, and the company will continue to ensure their user experience through software optimization.
Q: How soon can we see autonomous Tesla Semi trucks?
A: Regarding the progress of the Tesla Semi autonomous truck, our production plan is proceeding according to the scheduled timeline. The factory building has been completed, and production equipment is being installed. The validation test fleet has been conducting road tests on real roads, and a larger-scale trial production is expected to be completed by the end of this year, with formal mass production starting in the first half of next year, capacity ramping up in the second quarter, and achieving large-scale production in the second half of the year.
In terms of technical application, although the current engineering team is mainly focused on perfecting passenger car autonomous driving technology, we confirm that the same technical architecture can be smoothly transferred to the Semi truck. Once we obtain sufficient actual operating data for the Semi, we can quickly advance the development of autonomous driving functions for this platform.
Regarding the possibility of replacing rail transport, we believe that trains have significant efficiency advantages in long-distance point-to-point transport. However, in the "last mile" of the logistics chain, including cargo loading and unloading and short-distance delivery, autonomous Semi trucks will provide a more optimized solution. We expect this to reshape the future transportation landscape, as Elon said, fundamentally changing people's perception of transportation.
Q: You mentioned that after gaining confidence in non-regulatory autonomous driving, you will quickly expand vehicle production. How should we view this issue in the context of your existing 3 million vehicle capacity? When do you hope to reach this production level? Does this require some degree of demand stimulation or incentives? Does this mean prioritizing production over short-term profitability for longer-term opportunities?
A: Regarding the capacity expansion plan, we have not yet fully reached the 3 million vehicle capacity level, but we expect to achieve an annualized production target of 3 million vehicles within 24 months, possibly sooner. This requires the coordinated cooperation of the entire supply chain system, and we will advance capacity construction at the fastest speed that we and our suppliers can bear. Based on the existing capacity, we are planning to build new factory facilities.
The most important capacity expansion will come from the Cybercab model, which is planned to start production in the second quarter of next year. This is a vehicle specifically designed for fully autonomous driving scenarios, with the main feature of eliminating the steering wheel and pedals, and the core design concept is to achieve the lowest comprehensive operating cost per mile.
Regarding market demand and profitability, I believe there is no need to stimulate demand by sacrificing profit margins. Autonomous driving technology itself is the strongest demand driver—when consumers realize they can safely use their phones to handle affairs in the car, this will generate huge market demand. In fact, this is the main reason for the current high frequency of traffic accidents, and autonomous driving technology will fundamentally improve driving safety.
On the technical side, I am confident in achieving full autonomous driving that surpasses human driving safety levels. We have released FSD version 14.1 and developed a comprehensive technical roadmap. Next, we will introduce inference capabilities to the vehicle, continuously optimizing the AI system through a powerful reinforcement learning simulator. We also plan to increase the model parameter scale by an order of magnitude, and these improvements will enable vehicles equipped with the AI 4 computing platform to exhibit intelligence levels close to biological intelligence.
Looking to the future, the AI 5 computing platform will bring a leap in performance by an order of magnitude. Such powerful computing power may even exceed the vehicle's own needs, leading us to consider the possibility of building a distributed computing network. When the global Tesla fleet reaches tens of millions or even hundreds of millions, with each vehicle providing 1 kilowatt of inference capability, it will form a distributed computing network with a computing power scale of 100 GW, becoming a highly valuable strategic asset for the company.
Q: When considering growth prospects, how do we define areas that truly belong to Tesla's core capabilities, and where do you draw the line to distinguish markets or AI applications that do not belong to Tesla's core capabilities?
A: Regarding Tesla's core capabilities, we believe that the company's core competitiveness is dynamically formed during the development process. Looking back at the development history, Tesla initially did not have any core capabilities but gradually built the existing capability system through the internal incubation of multiple startup projects. We have successively broken through battery pack manufacturing technology, developed home and utility-scale energy storage products, established a global supercharging network, and independently created chip design and AI software teams.
In the field of humanoid robots, the large-scale production of Optimus has revolutionary significance. We estimate that its annual productivity may reach five times the level of humans, and with 24/7 continuous operation capability, this will create enormous economic value. Currently, the Optimus engineering team brings together senior engineers from the automotive department and many new talents who have just graduated. Through close collaboration between the engineering and manufacturing teams, we continuously optimize design solutions to improve manufacturability. From the kung fu performance demonstrated by Optimus 2.5 to the upcoming Optimus 3, we are achieving both technical breakthroughs and large-scale manufacturing goals.
In terms of AI technology routes, Tesla and XAI take differentiated development paths. Tesla's AI model scale is relatively compact, focusing on real-world application scenarios, with model sizes only 5%-10% of Grok's. This design philosophy focused on specific fields allows our AI system to run efficiently on vehicle and robot platforms. Although the two have technical complementarities in areas such as voice interaction, they essentially advance artificial intelligence development from different dimensions: XAI is committed to exploring general artificial intelligence, while Tesla is deeply engaged in the real-world application of embodied intelligence.
Q: Returning to Austin, if you can remove the safety driver on your side, is the restriction in the Bay Area purely regulatory, or is it a market-to-market learning process? Similarly, in the eight to ten markets you mentioned, is it a step-by-step process to decide whether to place a safety driver or a safety driver in the passenger seat? Or is it really just the regulatory requirements of each market?
A: In terms of safety driver deployment strategy, our cautious attitude is mainly due to our own safety considerations, rather than purely being constrained by regulatory requirements. Even if regulatory agencies do not make mandatory requirements, we will still adopt a conservative plan of deploying safety drivers in the early stages of new markets. This obsession with safety stems from the emphasis on potential risks—even a one-in-ten-thousand probability of an abnormal situation could lead to serious consequences.
We choose to conduct about three months of safety driver testing in each new metropolitan area to comprehensively verify the system's adaptability to different urban environments. This helps us identify potential special challenges, such as complex road intersections or other unforeseen driving scenarios. Although from a technical perspective, we are fully capable of directly conducting fully unmanned operations, for absolute safety considerations, we prefer to take a more prudent approach. This observation period will provide us with sufficient data support, and once the system's stable operation is confirmed, we will remove the safety driver as planned.
Q: Is the development path you are taking in Robotaxi different from the version you provide to early adopters? When you push these new versions, are you looking for significant improvements in intervention rates? Or is this issue basically resolved? Is it more about adding features like parking and driving modes, or just overall comfort?
A: In the development path of autonomous driving software, we always adhere to the basic principle of safety first. When we release important new software architectures, the primary task is to ensure system safety, and then we will optimize the comfort experience. This is why we recommend most users wait for version 14.2, which will address many issues affecting comfort. In the initial version, the system may be safe but operate somewhat stiffly, which requires time to gradually improve.
Regarding product version strategy, the software we provide to ordinary users is basically consistent with the Robotaxi platform. The core algorithms and system architecture are completely the same, with the main differences only in some specific functions, such as the ordinary user version including personalized functions like autonomous parking space selection, which are not necessary for Robotaxi operations.
In terms of technical roadmap, we plan to introduce "inference" capabilities in versions 14.3 or 14.4. This will enable the vehicle to intelligently judge parking lot situations, such as automatically identifying the possibility of scarce parking spaces at the entrance and making reasonable decisions—first dropping off passengers at the entrance and then autonomously finding a suitable parking space. This inference capability based on 360-degree panoramic vision will surpass human performance.
The key challenge in achieving these functions is integrating complex inference capabilities into the onboard computer. Although inference is relatively easy to achieve on the server side, the vehicle needs to have real-time decision-making capabilities, which poses extremely high requirements for both hardware and software.
It is worth mentioning that the Tesla AI system has a significant advantage in intelligence density. Because all intelligence must be compressed into the AI 4 computing platform, we have achieved extremely high intelligence density. This technical accumulation will lay a solid foundation for the future AI 5 platform, and when computing power increases by an order of magnitude, we will achieve more powerful performance.
Q: I just want to coordinate the timeline for starting production next year with the current supply chain situation, and it sounds like there is still a lot of work to be done in terms of flexibility before you can really freeze the hardware design and start scaling production?
A: Regarding the production timeline and supply chain coordination for Optimus, our hardware design strategy has a continuous iteration characteristic. Even after officially starting production, we will continue to update the Optimus design in a rolling manner, as complex aspects that need improvement are often discovered during the manufacturing process.
According to the current plan, we will complete the demonstration of the production intent prototype in the first quarter of next year (expected in February-March). Subsequently, we plan to build the first Optimus production line with an annual capacity of 1 million units by the end of next year. It should be noted that achieving the goal of an annual capacity of 1 million units requires a gradual ramp-up process, and its progress will depend on the weakest link in the entire supply chain system.
Looking further ahead, we expect the Optimus 4 generation product to achieve a production capacity of tens of millions of units, while the Optimus 5 generation may reach a scale of 50 million to 100 million units. This development path, although challenging, shows great growth potential.
Risk Disclosure and Statement of this Article:Dolphin Research Disclaimer and General Disclosure
