April 15, 2026

Physical AI refers to various robotic systems that can perceive, understand, and act in the real physical world, rather than simply processing data digitally. It includes autonomous vehicles, industrial automation systems in factories, drones, smart home devices, and other types of robots. As more of these devices are designed and produced, an important question arises: how will the growth of physical AI affect power generation?

How Advancements in AI Lead to Growth of Robotic Devices

Robotic devices have existed for a long time, but traditionally they have been programmed to perform strictly defined tasks. For example, a welding robot in a car assembly plant operates based on predefined coordinates. However, in many cases robots must respond to real-world conditions, which can vary. For instance, a robot operating at a construction site must determine its movements based on the actual layout of the site and the location of materials. This capability can be achieved using modern algorithms and models running on advanced chipsets embedded in the robot.

Robot Power Consumption

The power consumption of a robot depends on both its electric motors, which enable movement, and its electronic components, such as chipsets, sensors, and lighting. Incorrys estimates that, depending on the type of robot and its functions, power usage ranges between 0.5 and 3 kW—similar to that of a household dryer.
Incorrys analyzed three scenarios of robot power consumption:

• Average utilization of 30% (robot operates 30% of the time), average power: 2 kW
• Average utilization of 50%, average power: 2 kW
• Average utilization of 60%, average power: 2.5 kW

Figure 1 shows energy consumption for different numbers of robots under these scenarios.

Figure 1. Robot Power Consumption

In 2025, total U.S. utility-scale electricity generation was about 4,430 trillion kilowatt-hours (kWh). This is comparable to the estimated power consumption of 300 million robots under Scenario 3. According to the International Federation of Robotics (IFR), the number of industrial robots in the United States was around 380,000 and continues to grow at approximately 5% per year.

In addition to industrial robots, future growth may include household robots (currently in development), construction robots, defense and security systems, search-and-rescue robots, and entertainment robots. These estimates do not include robotic transportation systems such as self-driving cars.

It is plausible that after 2040, most households will own at least one such device. For comparison, there are approximately 89 million residential clothes dryers in the United States, consuming an estimated 66–71 terawatt-hours (TWh) of electricity annually. However, dryers operate only about 2.5% of the time, whereas robots—especially in industrial or construction settings—can operate up to 24 hours per day.

Conclusion

While these scenarios are highly speculative, the analysis above suggests that the energy industry must be prepared for potentially rapid growth in electricity demand driven by physical AI.

See Also:

Gasoline Prices Rise: Will EV Sales Increase?

References:

Owen-Hill, Alex. “Saving Money on Robot Energy Costs.”, RoboDK,  June 8, 2023, https://robodk.com/blog/saving-money-on-robot-energy-costs/

Soori, Mohsen. Arezoo, Behrooz. Dastres, Roza. “Optimization of energy consumption in industrial robots, a review.”, Cognitive Robotic, Volume 3, 2023, Pages 142-157, https://www.sciencedirect.com/science/article/pii/S2667241323000174?via%3Dihub

“The Energy Diet of Humanoid Robots.”, Mobius Market Research, Nov 17, 2024, https://research.mobiusriskgroup.com/p/the-energy-diet-of-humanoid-robots.

“Electricity in the United States.”, Energy Information Administration, 2025, https://www.eia.gov/energyexplained/electricity/electricity-in-the-us.php

“Record 380,000 Robots Working in US Factories.”, International Federation of Robotics (IFT),  September 2024, https://ifr.org/downloads/press2018/2024-SEP-24_IFR_press_release_World_Robotics_2024_-_USA.pdf