Regarding Robotaxi, robots, the Chinese supply chain, and talent, after researching Tesla, Morgan Stanley provided ten observations

Tesla is accelerating its robotics business, and Morgan Stanley has provided ten observations after its research.

According to the Wind Trading Desk, in Morgan Stanley's latest research report, it maintains an "overweight" rating on Tesla, giving a target price of $410, which represents about a 17% upside from the current level, and reiterates it as the top stock pick in the North American automotive industry.

This report is based on the Morgan Stanley team's investigation of Tesla and other automation and robotics companies in the San Francisco Bay Area, with the core of the report focusing on Tesla's developments in Robotaxi, robotics, supply chain and talent, and supercomputing. Here are the ten key insights they gathered:

Robotics Project: The Future is Here, but Time is Needed

1. Austin's autonomous vehicle fleet is about to launch, but with cautious scale

The autonomous vehicle fleet in Austin is about to go into operation, but Tesla is taking a cautious approach, with an initial scale of only 10 to 20 vehicles. This will be an invitation-only public road test, equipped with a large number of remote operators to ensure safety—as a Tesla representative said: "We can't mess this up." The specific launch date has yet to be determined, but preparations are underway.

2. The Optimus humanoid robot is set to debut

Tesla's head of investor relations, Travis Axelrod, stated that the company hopes to allow more investors to visit its small-scale manufacturing facility in the fourth quarter of this year, suggesting that there may be events related to the Optimus robot before the end of the year.

3. Drones? Possibilities exist but are not a priority

Tesla acknowledges that managing capacity is one of the biggest bottlenecks the company faces. According to Morgan Stanley analysts, the technological leap from autonomous vehicles to autonomous robots is comparable in scale to the leap from autonomous vehicles to autonomous aircraft/drones. While Tesla does not rule out the possibility of entering the drone field, the company currently has too many projects underway.

Cost Control, Talent Competition, and Supply Chain Localization

4. Cost control is key

Tesla has set a target cost of $20,000 per unit for the Optimus humanoid robot, while the target cost for the robotaxi is $30,000 per unit, the latter being a heavier piece of equipment. Tesla emphasizes, "You can't separate scalability from cost," indicating that mass production is crucial for cost reduction.

5. Ambitious production expansion plans

Tesla plans to commercialize Optimus as early as mid-2026. Currently, the Palo Alto lab can only manually produce "a few dozen" robots per batch. The production line on the second floor of the Fremont factory will have a monthly capacity of 1,000 units, with the second-generation production line targeting a monthly output of 10,000 units, and the third-generation production line targeting a monthly output of 100,000 units. Tesla plans to produce "thousands" of Optimus robots by the end of this year.

6. The popularization of household robots will take several years

Tesla acknowledges that intelligence and cost "still have a long way to go" to unlock the true potential of humanoid robots. The neural network of Optimus is larger than that of cars because it needs to handle more degrees of freedom and more open-ended tasks. This means that humanoid robots for home use may still take years to realize7. The Talent War is Becoming More Intense

With the rise of startups, the talent war is becoming increasingly fierce. Tesla acknowledges that the talent market is "exceptionally competitive." Additionally, Tesla emphasizes that China has a large pool of highly educated talent in applied sciences and manufacturing, believing that "having a large number of cost-effective technical workers is a significant advantage for China."

8. Challenges for Domestic Production in the U.S., China Controls the Supply Chain

Tesla points out that, in terms of core next-generation technologies, China controls the supply chain, and there is still a lot of work to be done to regionalize the supply chain in the U.S.

Dojo Supercomputer and Computing Power Demand

9. Dojo Supercomputer Still Has Room for Development

Tesla plans to launch the D2 supercomputer in 2026, which is expected to be more efficient than the D1 and cheaper than similar NVIDIA solutions, but it still cannot compete with NVIDIA "one-on-one" in terms of competitiveness. The D3, expected to be launched in 2028, is anticipated to narrow the performance gap with ASICs. As Tesla considers increasing computing power by ten times, acquiring cheaper chips could become a significant breakthrough.

10. Computing Power Demand Will Grow Exponentially

As the scale of neural networks expands, Tesla's computing demands will increase. In terms of inference, the next-generation AI5 will have more than ten times the TOPS (trillions of operations per second) compared to the current AI4, indicating that Tesla's demand for computing power will continue to grow