Forrester has published research arguing that the next major advance in artificial intelligence will come from physical AI systems rather than humanoid robots. The report identifies physical AI as one of the key emerging technologies likely to shape the next decade.
The research examines how systems that can perceive, reason and act in real-world settings are being used across manufacturing, logistics, mobility and energy. It argues that these systems are pushing AI beyond software and screen-based applications into machines and autonomous platforms operating in dynamic physical environments.
According to the report, physical AI combines four elements: modelling, perception, reasoning and action. Forrester distinguishes this approach from traditional automation, which relies on scripted behaviour in controlled settings.
That distinction matters in environments where conditions change quickly. Physical AI systems use world models, sensor data and reasoning engines to adjust their behaviour safely as circumstances shift, allowing them to operate in settings ranging from factory floors to public roads and industrial sites.
Beyond humanoids
A central finding is that humanoid robots are only one part of a much wider market. Although they attract much of the public attention around robotics and AI, many of the most commercially useful deployments are appearing in other formats designed for specific operational tasks.
These include mobile robots, collaborative robots, autonomous vehicles and inspection systems. The same advances in AI models and sensing that have driven interest in humanoids are also improving these systems, often in forms that are less expensive and more durable.
The research also points to growing use of shared perception and reasoning models across different machines and applications. Companies are increasingly applying the same underlying models to multiple tasks, helping systems switch roles, recover from faults and continue operating for longer periods.
Another feature highlighted in the report is fleet-level coordination. Managing groups of machines together can help organisations maintain output even when one unit is offline for maintenance or fails during operation.
Industry impact
The report presents physical AI as a broader shift in how artificial intelligence is deployed in business operations. Rather than supporting digital decision-making alone, these systems directly affect the physical environment, with implications for safety, regulation, workforce planning and capital spending.
For industries facing labour shortages, hazardous sites or highly repetitive tasks, that shift could change where automation investment is directed. The value proposition, the report suggests, extends well beyond machines that resemble humans to a range of specialised systems built to handle narrow or difficult tasks in real-world conditions.
This assessment comes as companies continue testing where AI can improve operational resilience and productivity. In Forrester's view, physical AI offers a path to more adaptable automation because it can respond to unstructured settings rather than relying entirely on fixed routines.
Paul Miller, Vice President and Principal Analyst at Forrester, said the conversation around the technology has become overly focused on humanoid robots.
"There's no question that humanoid robots are impressive, but they're not the most important part of the story. The real shift is the growing capability of physical AI - systems that can perceive what's happening around them, reason about next steps, and safely act in the real world. Those capabilities benefit humanoids, but they matter far more for the many other forms of physical automation that are cheaper, more durable, and already delivering practical value at scale," Miller said.