What is physical AI and why everyone is talking about it now

Think of physical AI as a symbiosis of artificial intelligence and robotics that enables machines to perceive their surroundings in real time, learn, and act autonomously. Unlike ChatGPT or Claude, which "live" in a data center, physical AI operates directly in the physical world — through humanoid robots, autonomous carts, drones, or surgical assistants.

According to an April 2026 report by MarketsandMarkets, this represents a convergence of several key technologies: edge AI computing (processing data directly on the device without sending it to the cloud), multimodal perception (cameras, LiDARs, radars, and tactile sensors), and generative AI models that enable robots not only to execute programmed movements but also to improvise in unexpected situations.

Nvidia's Jensen Huang has called physical AI "the next wave of artificial intelligence" — and Nvidia is betting heavily on it. Its platform includes the Omniverse simulation engine, the GR00T framework for humanoid robots, and Jetson Thor computing modules that allow complex AI models to run directly onboard a robot. At CES 2026, Nvidia demonstrated how robot manufacturers around the world use its stack.

A market set to grow seventeenfold in six years

The specific numbers from the report are impressive:

• 2025: $0.89 billion
• 2026: $1.50 billion
• 2032: $15.24 billion
• Compound annual growth rate (CAGR): 47.2%

For comparison: the European generative AI market (i.e., the models and services themselves like ChatGPT) is forecast to grow from $16.5 billion in 2025 to just under $203 billion by 2032. Physical AI comes from a smaller base, but its 47% annual growth rate is among the highest in the entire technology sector.

The two fastest-growing segments will be industrial robots (56.7% CAGR) and advanced Level 3 autonomy (60.8% CAGR). Level 3 refers to robots that handle complex tasks without constant human supervision — from manipulating fragile objects to navigating in a dynamic warehouse environment.

Defense, medicine, and digital twins: three growth engines

The report identifies three main drivers pushing physical AI forward.

Defense and security

Armies around the world are investing in autonomous systems — from inspection drones to robotic scouts to logistics robots that supply troops in the field. Boston Dynamics has delivered over a million units of its Spot robot for inspection and security purposes, including military applications. Physical AI here means not only greater efficiency but also protecting human lives in dangerous situations.

Healthcare

Robotic assistance in medicine has long ceased to be science fiction. Diligent Robotics with its humanoid Moxi has achieved over 1 million deliveries of materials in hospitals and saved healthcare staff over 575,000 hours of routine work. Physical AI in surgery, rehabilitation, and elderly care is among the segments with the greatest growth potential — especially in the context of Europe's aging population.

Digital twins and manufacturing

Factories are increasingly creating digital copies of their operations, where they test changes before deploying them in reality. Physical AI connects these virtual models with real robots — the robots learn in simulation and transfer acquired skills to the real world. According to MarketsandMarkets, this is the fastest-growing software segment of the entire market.

BMW deployed humanoids in Leipzig. Europe isn't falling behind

While Asia (primarily China, Japan, and South Korea) holds a 50.4% share of the global physical AI market in 2026, Europe is catching up fast. A key milestone was the February deployment of Figure 02 humanoid robots at the BMW plant in Leipzig — the first commercial use of physical AI in European automotive manufacturing. The robots assist with sheet metal handling and battery component production.

For the Czech Republic, this is not a distant future. With the strongest industrial base in the EU (industry's share of GDP around 30%) and the presence of car manufacturers such as Škoda Auto and TPCA, the Czech Republic is ideally positioned for the adoption of physical AI. Moreover, the European AI Act and the EU Machinery Regulation 2023/1230 create a clear regulatory framework that should make investment decisions easier for companies.

European leaders in the physical AI supply chain include ABB (Switzerland), Festo and Bosch Sensortec (Germany), STMicroelectronics (Switzerland), and Infineon Technologies (Germany). All of these develop sensors, actuators, and control systems without which no smart robot could function.

Who will dominate the market? Nvidia has a lead, but competition is intensifying

According to MarketsandMarkets' methodology, the "stars" of the market include, alongside Nvidia, Qualcomm (edge AI processors), ABB (industrial robotics), and Texas Instruments (semiconductors). In the startup and smaller player segment, companies like Figure AI, Agility Robotics (Digit), and Boston Dynamics (Atlas, Spot) are scoring points.

A landmark case is the collaboration between Amazon and Agility Robotics: the humanoid Digit has already handled over 100,000 containers in Amazon warehouses and demonstrated a positive return on investment. Similarly, Figure 02 at the BMW Spartanburg plant helped produce over 30,000 vehicles.

The biggest challenge remains interoperability — the ability of robots from different manufacturers to work together in a single operation. There are no unified standards yet, which complicates the deployment of mixed robotic fleets for companies. That is precisely why initiatives such as MIPI Physical AI BoF (which we recently covered on jarvis-ai.cz) are working toward interface standardization.

What this means for companies — and for employees

Physical AI doesn't just mean more efficient factories. For Czech manufacturing companies, it represents a way to remain competitive at a time when labor costs are rising and the shortage of skilled workers is deepening. Robots aren't taking jobs — they're taking over routine, physically demanding, and dangerous tasks, while people shift to activities requiring creativity, decision-making, and fine motor skills.

MarketsandMarkets' analysis also identifies risks: high initial investments, long hardware replacement cycles, and persistent public distrust. Yet the 47% annual growth trajectory suggests the sector is entering a phase where experiments become a standard part of industrial equipment.