This helps to manage variability and unpredictability in the external environment, in high-mix, low-volume production, as well as in public environments. Robots equipped with vision systems, for example, can analyze past tasks to identify patterns and optimize their operations for greater accuracy and speed.

Robot and chip manufacturers are investing in the development of dedicated hardware and software that simulate real-world environments. This physical AI allows robots to train themselves in virtual environments and operate by experience, rather than programming.

These generative AI projects aim to create a “ChatGPT moment” for physical AI. This simulation technology is expected to advance in both traditional industrial robot environments, as well as in service robot applications.

Robots in the shape of human bodies are also receiving a lot of attention. The goal is to create bipedal machines that can perform general-purpose tasks such as loading a dishwasher on their own or working independently on an assembly line.

Many manufacturers are currently focusing on humanoids that only perform single-purpose tasks. Most of these projects are being carried out in the automotive industry and the warehousing sector.

The challenges is to determine if and how humanoid robots can make an economically viable and scalable business case for industrial applications, especially when compared to existing options such as cobots.