Tesla's humanoid robot "Optimus" has been upgraded again?

Recently, a new "golden Optimus Prime" was unveiled, and the most amazing thing is its hand - the fingers are flexible and finely structured, almost like a real hand, and can make all kinds of delicate movements. Many people speculate that this may be the latest generation of the Optimus 3 prototype.

This wave of evolution directly "circles" the big guys. Salesforce founder Mark Benioff praised after reading it: "This is the beginning of the 'physical intelligence' revolution!" He also predicted that this robot may be qualified for jobs with an annual salary of $200,000 to $500,000 in the future, such as office tasks, factory operations, etc.

From "active" to "able to work"

Tesla's robot "hand" has quietly evolved for three generations, becoming more and more like a real person's hand.

The first generation: designed to imitate the human hand, with five fingers, it can do a total of 11 movements (professionally called "degrees of freedom"), of which 6 actions are actively controlled by the motor, and the rest are mechanically assisted.

By the third generation, progress was noticeable:

Many people look at robots and only care about "whether they can walk", but Jingtai believes that the hand is the "winner and loser" of the commercialization of humanoid robots.

Why?

Because most of the work does not rely on walking, but on "hands-on". Factory assembly: screwing screws, inserting parts; warehouse sorting: taking boxes, labeling; home services: opening doors, pouring water, sorting out items.

Tesla's dexterous hand has 22 degrees of freedom, indicating that it is not for performance, but to truly enter factories and homes and replace human labor.

Behind this are three strategic intentions: reduce manufacturing costs: in the future, robots must be highly integrated and mass-produced to sell for less than $20,000; Improve reliability: external drive + modular design, more convenient maintenance; Reuse Tesla genes: planetary gears, lead screws and other components, highly synergistic with the automobile manufacturing supply chain.

From "AI on wheels" to "AI on feet"

Just a few days ago, Tesla officially released the fourth chapter of the "Grand Plan", clearly announcing that the highlight of the future is no longer electric vehicles, but bringing AI into the real world - that is, robots.

Musk bluntly said on the social platform X on the same day: about 80% of Tesla's value in the future will come from Optimus robots, not from selling cars.

This is a heavy statement, which is equivalent to an official announcement: Tesla is fully transforming from an "electric vehicle company" to an "artificial intelligence + robot" giant. Cars are still "cash cows", but they are no longer "future". Robots are Tesla's "second curve" and ultimate goal.

The company also emphasized that they are building products and services that "make AI move" - so that smart algorithms no longer only exist on screens, but can walk, work, and interact with the real world.

Don't think it's a sudden turn. They have been quietly deploying in the field of AI and automation for almost two decades. Now, it's time to bring robots to center stage.

|Optimus Prime mass production plan

Musk revealed that Tesla plans to produce more than 1,000 Optimus Prime units in 2025, and by 2026, that number will soar to 50,000 units. The goal is clear: to make Optimus Prime an "all-purpose worker" in the factory - whether it is assembling parts, carrying goods, or doing all kinds of delicate work, it can be done.

Tesla hopes to keep the cost of each robot at around $30,000 (almost more than 200,000 yuan) so that factories are willing to purchase in large quantities.

But then again, this goal sounds cool and not easy to achieve. At present, Optimus Prime is still a prototype, and to achieve an annual production capacity of 50,000 units, it must cross two mountains. Large-scale mass production: from manual assembly to automated production lines, yield and consistency are difficult problems; Cost control: dexterous hands, drive systems, and sensors are all "big burners", how to reduce BOM costs?

This is like Tesla's Model 3 back then - the "capacity hell" is not to scare people, but to use engineering capabilities to "break through".

How to understand Tesla's "new valuation logic"?

1. Revaluation : From "car company" to "AI + robot platform"

At present, the market still values Tesla mainly based on "electric vehicle sales"; But in the future, Optimus' potential market is trillions (global labor replacement).

Suggestion: Treat TSLA as a "future technology platform" rather than a traditional car company; Pay attention to the progress of robots (such as mass production time, customer pilots); Hold for a long time, but avoid chasing higher in the short term (expected volatility).

2. Pay attention to the "physical intelligence" industry chain

Tesla's transformation will drive a number of "physical AI" related industries.

Dexterous hand/joint module: harmonic reducer, coreless motor (Tuopu Group, Sanhua Intelligent Control);

AI training infrastructure: NVIDIA, AMD, TSMC;

Sensor fusion: lidar, IMU, force control sensor;

Human-computer interaction: voice, vision, motion capture.

These "behind-the-scenes players" will continue to benefit from the wave of AI materialization.