人形机器人板块点评：马斯克关于人形机器人近期深度访谈要点一览

Investment Rating - The industry investment rating is "Positive" (maintained) [6] Core Insights - Initial production of humanoid robots will be "painfully slow," with capacity ramp-up following a stretched S-curve, as Tesla phases out Model S and Model X to focus on building a production line with an annual capacity of one million units. Achieving an annual capacity of ten million units likely requires the introduction of a more mature fourth-generation model [1] - The ultimate cost target for manufacturing a single Optimus robot is approximately $20,000 [2] - Tesla faces three core bottlenecks in humanoid robot development: real-world AI technology, scalable manufacturing capabilities, and the dexterity of robotic hands, which is emphasized as the most complex electromechanical engineering challenge [2] - The construction plan for the "Terafab" factory, which will exceed the scale of any existing Tesla superfactory, aims for a monthly wafer production capacity of over one million pieces. Significant AI computing power will be deployed in space within the next 36 months [3] - The core bottleneck for computing power development on Earth is not chip production but power supply, with space data centers being the most economically viable solution [3] - Tesla has developed a world simulation model to create an Optimus Academy, where thousands of robots will self-train in real-world scenarios to master various task skills [4] - The pricing gap between Tesla's Optimus and Chinese competitors is significant, with the R1 humanoid robot from Yushu Technology pre-selling for as low as $4,900. However, Musk clarifies that the two are not comparable products due to Optimus's advanced design and capabilities [5] Summary by Sections Production Capacity - Initial production will be slow, with a focus on building a million-unit capacity line [1] - Achieving ten million units will likely require a new model [1] Cost Targets - Target manufacturing cost for Optimus is around $20,000 [2] Development Bottlenecks - Key challenges include AI technology, manufacturing scalability, and hand dexterity [2] Computing Power - "Terafab" factory will have a monthly capacity exceeding one million wafers [3] - Space data centers are seen as a viable solution for computing power bottlenecks [3] Training and Competition - Optimus Academy will facilitate self-training for robots [4] - Significant pricing differences exist between Tesla and Chinese competitors [5]