Optimus to ramp up first to 10 thousand–20 thousand units: Why does JPMorgan prefer robot components?
Tesla has started Optimus production, but internal deployment and a gradual ramp-up still lie ahead before external sales. JPMorgan is more positive on component demand generated by research and development, pilots and data collection, while remaining cautious about sentiment-driven gains in robot manufacturers.
Original source：J.P. MorganFirst published on Unhidea：Original source date：
Key takeaway
Optimus’s near-term commercial opportunities are more concentrated in components needed for research and development and pilots, while large-scale deployment of complete robots is constrained by dexterous hands, real-world data and the pace of gradual mass production.
Production has started. Why is it still below some investors’ expectations?
Musk said Optimus production has started in Fremont and that Tesla could deliver robots to other companies before the end of 2027. Before external sales, production will first ramp up gradually to approximately 10 thousand–20 thousand units. Early units will mainly perform dangerous and repetitive tasks internally, with consumer uses coming later.
This pace is broadly consistent with the regional and global investors’ expectation described in the report of producing 10 thousand–20 thousand units in 2027, but below the optimistic scenario of approximately 50 thousand units formed by some domestic investors from supply-chain information, extrapolated at approximately 1,000 units per week. The gap in expectations primarily concerns the speed of mass production, rather than whether production has begun.
The key to the ramp-up is that the slowest component determines overall progress, and dexterous hands remain the main hardware bottleneck. As a result, improvements in component supply and system capabilities still stand between the start of production and stable, large-scale deliveries.
Why is near-term demand more visible for components than for complete robots?
Following research at the China International Industry Fair, JPMorgan concluded that near-term monetization will primarily come from “ecosystem demand,” including research and development, pilots and data collection. These activities already require sensors, 3D vision, data tools, reducers and dexterous hands, giving the relevant parts of the supply chain greater demand visibility than batch deployment of complete robots.
On that basis, the bank remains cautious about sentiment-driven gains in robot manufacturers such as Unitree, UBTECH and Estun, and prefers Hengli Hydraulic and Sanhua Intelligent Controls, two companies in Tesla’s robot supply chain. Among component makers, its top pick is Leader Harmonious Drive Systems, followed by Orbbec, while it also flags profitability risk at the latter.
This selection approach does not treat all robot suppliers as equal beneficiaries. It prioritizes parts of the supply chain directly connected to current research and development and pilot procurement. Clearer demand remains distinct from profitability, and Orbbec’s profitability risk is a constraint in the bank’s ranking.
What hurdles remain on the path from demonstration to practical use?
Beyond hardware, investor feedback in the report shifts attention to the move from simulation into real-world settings, with investors believing that real-world data will remain a bottleneck. Dexterous hands affect whether robots can be manufactured smoothly, while data from practical applications concerns how they work on site; the two impose different constraints.
Musk said an unscripted Optimus demonstration would be shown within the next 1–2 months, and said the next-generation AI5 chip in 2027 should significantly boost computing power and support more complex behavior. The demonstration relates to actual behavioral capabilities, while the chip relates to computing power; these developments will determine how quickly robot uses can expand beyond the mass-production ramp-up.
