What It Will Take to Build 400,000 U.S. Humanoids by 2030

NEW YORK -- U.S. humanoid robot production may total 300,000 to 400,000 units per year by 2030, says an estimate by JPMorgan Chase & Co. But achieving that level of production is going to take solving problems which are unrelated to the ability of the robots to walk around and do things in demos.

The forecast from the bank, reported Nov. 12, sees production growing to about 1.5 million humanoids by 2035. The projections are based on the assumption that the robots will begin to take care of repetitive tasks in manufacturing, warehousing and logistics.

However, there is a substantial gap between creating a robot that can carry out the required action and producing tens of thousands of robots that can repeat the process every day.

From Prototypes to Production

Humanoid manufacturers are now trying to develop their machines, test them on customers or get ready for large production runs. JPMorgan estimates that U.S. manufacturers of humanoids, other than Tesla, aim at producing 10,000 to 30,000 robots annually.

That's just the plan.

Large-scale production requires an uninterrupted flow of actuators, motors, sensors, and robotic hands. These parts need to endure repeated manipulation and loads as well as remain affordable for corporate clients.

Developing robotic hands constitutes a challenging task. Industrial grippers are built to handle certain items; yet humanoid hands have to manipulate diverse sized, shaped, and weighted objects. That means many actuators, sensors, and complex control system need to fit in a limited space.

Learning abilities of robots constitute another limitation. To manipulate differently-sized items under varied circumstances, robots need to adapt to changing environment. Gathering enough practical information is expensive and knowledge acquired in simulations often doesn't apply to the hardware.

Humanoid Economics

Cost-efficiency of humanoid robots cannot be assessed based on manufacturing prices alone.

According to JPMorgan estimates, each client can save around $100,000 yearly by using robots during multiple shifts. The exact amount will differ according to the purchasing price or rental fee of the machine, maintenance costs, energy usage, efficiency and the number of human managers required.

Such a robot with the ability to perform a few functions could prove useful to warehouses and factories designed for human labor. However, a machine that is frequently interrupted or works slower than normal automation might have little financial gain.

That is why customer performance tests would be more valuable than presentations in assessing manufacturers' development.

Profits for Manufacturers of Components

According to JPMorgan, by 2035, the market for some key humanoid components could reach $30 billion.

Companies that sell actuators, motors, machine vision systems, and other equipment could see their profits grow as manufacturers spend more money on development and production, even if they are not using humanoids extensively yet.

Companies that JPMorgan sees as beneficiaries include Timken, Robotis, Leader Drive, Orbbec and Cognex.

However, for humanoid manufacturers, production capabilities would be not enough to ensure their commercial success. They should also show that their robots can work economically and efficiently for long periods of time for the clients.

The actual JPMorgan research note was not available in the analyzed materials. Thus, the exact definition of their production numbers of projections remains unknown.