Key changes in the robotics sector: Elon Musk has reassigned the head of autonomous driving to the Optimus project, aiming to evolve the robot’s “brain” with FSD technology. In the short term, this may disrupt progress expectations, but in the medium to long term, technological empowerment is expected to drive industry breakthroughs. Current stock prices reflect low expectations, necessitating tracking of technological catalysts.
1. Strategic Direction
– Core Action: FSD head Ashok has been reassigned to the Optimus project.
– Technical Logic: Reusing algorithms for visual perception and motion control from autonomous driving to enhance the robot’s environmental recognition and decision-making capabilities.
2. Market Dynamics
– Short Term: Personnel adjustments may raise concerns about project timelines, putting emotional pressure on the market.
– Medium to Long Term: The integration of technologies will enhance the intelligence level of robots, opening up valuation space.
3. Stock Price and Expectations
– Position: The stock has retraced to the low levels of February to April, with optimistic expectations fully digested, revealing a safety margin.
– Catalyst: Upgrades to the Optimus algorithm and evolution of the brain; observe new developments from the new team in the next two to three months.
4. Strategic Recommendations
– Focus on core components such as actuators and sensors, and targets that collaborate with autonomous driving technology, tracking project technological progress.
Several clues:
1. This year’s order rhythm has indeed changed, which is related to adjustments in the robotics division’s leadership affecting product training and design details. The current head, Ashok, is dissatisfied with the efficiency of hardware and its corresponding model training, and adjustments will be made at the model and hardware levels.
2. Since FSD has utilized world models in autonomous driving, we speculate that Ashok is likely to introduce training methods based on world models. However, there has been no relaxation in the number of training samples, so the overall training orders are unlikely to see a significant decline; the main change is in the order rhythm. Additionally, we speculate that the overall machine calibration is unlikely to change, as previous video data and training would become largely ineffective; thus, hardware adjustments will be more about fine-tuning.
3. The latest procurement and engineering personnel indicated at the procurement meeting that this year’s orders are only temporarily delayed, with normal deliveries expected in Q2. Orders for Q3 and beyond may decrease or increase, but this matter “doesn’t matter” (exact words); the important thing is to finalize training this year to enable mass production starting next year; the goal of commercialization by Q4 next year remains unchanged, so suppliers should generally prepare for the capacity corresponding to 50,000 robots.
4. On the specific hardware level, we understand that changes in the dexterous arm are greater than those in the joint module, including minor adjustments in overall design (previous factory spy photos have reflected this) and adjustments to internal lightweight components, with core lightweight component suppliers currently conducting lifespan tests, with orders expected this year.
Feedback from the Hangzhou Humanoid Robot Exhibition:
Today at the exhibition, we visited several companies focused on dexterous hands, lead screws, lightweight materials, reducers, etc., including Lingxin Qiaoshou, Qiangnao, Lingqiao Intelligent, Beifang Machinery, and Weiyuan. Through this exhibition, we believe that humanoid robots still have potential changes in the following areas:
1. Improvement of dexterous hand grip strength: Currently, most dexterous hands using rope drive solutions have a grip strength of 50N. To increase grip strength, using a linkage method may lead to a decrease in degrees of freedom. Methods to enhance grip strength include: increasing motor power density, enhancing rope strength, and switching to direct drive micro lead screws.
2. Increased penetration of micro lead screws: Micro lead screws can simultaneously meet the demands for high degrees of freedom and large loads in dexterous hands. Currently, there are few players in high-precision micro sliding and ball lead screws. Considering that sliding lead screws have lower transmission efficiency and ball lead screws have weaker load capacity, the penetration of micro roller lead screws is expected to increase in the future.
3. Advancement in lightweight materials: Many manufacturers are considering using materials like PEEK to achieve lightweight robots, thereby improving endurance. Currently, the strength of engineering plastics like PEEK is mostly between HRC25-35, and stronger lightweight materials are still to be discovered.
4. Trends in cycloidal pinwheel reducers need observation: Cycloidal pinwheel reducers, with their simple structure, high load capacity, and high reduction ratio, are expected to replace planetary reducers in the lower limbs of robots. However, this trend still requires further observation and validation.
