In the same week NVIDIA lost the man who sold the world on physical AI, Europe’s three biggest chipmakers stood up in Munich and said the chips inside the next robots should be European.
Speaking at the electronica trade fair on November 9, the chief executives of Infineon, STMicroelectronics and NXP said Europe could compete with the US and Asia in robotics, autonomous machines and intelligent industrial systems.
In Infineon CEO Jochen Hanebeck’s words: “for me, industrial AI or physical AI is, first of all, a must win for Europe. There are so many topics where we feel that we are second to other regions.”
Rather than trying to compete directly with NVIDIA‘s dominance in data-centre accelerators, Hanebeck and other European semiconductor manufacturers are focusing on physical AI. Physical AI, a branch of artificial intelligence that enables AI-powered machines to act within the physical world, uses sensors, power electronics and microcontrollers to connect AI to physical machinery.
NVIDIA dominates high-end graphics processing units (GPUs), holding an estimated 75% to 85% market share, according to analyst estimates. While GPUs are seen as the brains of AI, they are expensive and power-hungry, which limits their use inside machines.
Europe doesn’t have a major domestic firm designing GPUs that comes close to competing with NVIDIA. However, the EU does specialise in legacy industrial semiconductors, automotive chips, and semiconductor equipment that are essential to localised applications and physical AI.
Where Europe Already Wins
A microcontroller (MCU) acts as a tiny, self-contained computer on a single integrated circuit chip that processes and controls physical devices. Unlike power-hungry GPUs, MCUs handle the fast, local decisions a machine has to make on the spot, such as when to fire a motor or read a sensor.
Here the picture flips. European firms, largely absent from GPUs, dominate microcontrollers. Infineon has been the world’s largest MCU supplier since 2024 and held 23.2% of the market in 2025, up from 21.4% a year earlier. NXP is second.
MCUs serve as the bridge between high-level machine learning trained in data centres and real-world physical AI in robotics, industrial automation and smart devices. They do this at a fraction of the cost of GPUs, and without the delay of sending data to the cloud and back.
Europe’s Circulatory System
European firms are also strong in the power semiconductor space. Power semiconductors convert, regulate and switch electrical energy. They are used in power grids, data centres, televisions, phone chargers, refrigerators, etc.
In physical AI, power semiconductors act as the muscle, turning raw power into motion and computation. The high efficiency of current power semiconductors is essential to AI infrastructure: without them, a large share of energy would be lost and costs would soar.
Infineon and STMicroelectronics are both major suppliers of automotive and industrial power semiconductors, while NXP’s strength lies in automotive processors, MCUs and in-vehicle networking. Infineon alone held 17.4% of the global market for power discretes and modules in 2024, more than double its nearest rival, onsemi, according to Omdia.
Brussels Picks Its Battles
The EU Chips Act, which entered into force in September 2023, aims to double Europe’s share of global semiconductor production from roughly 10% to 20% by 2030, while reducing strategic dependencies that were first felt amid the 2020 supply chain disruptions.
The Act set a minimum of €43 billion in policy-driven investment through 2030, which Brussels expects private money to match. The European Court of Auditors has called the 20% target “very unlikely” to be met. One of its complaints was that European chipmakers mostly need mature production technology, while the original Act pushed cutting-edge manufacturing.
In June 2026, the European Commission proposed a Chips Act 2.0 designed to reinforce Europe’s existing strengths, including mainstream chips, while building capacity in cutting-edge technology. The proposal still has to pass the European Parliament and EU governments.
STMicroelectronics, meanwhile, is building a €5 billion silicon carbide campus in Catania, Sicily, backed by a roughly €2 billion grant from Italy that the Commission approved in May 2024. The plant will serve electric vehicle and industrial markets and is expected to reach full capacity in 2032.
Money Follows the Machines
The world is racing to build troves of humanoid factory workers. While most of this tech is still in piloting phases, or being exhibited more for spectacle, advances in the sector hint at a huge shift forward in robotics and automation.
China has emerged as a frontrunner in general-purpose robotics, but European companies are attracting serious money. On November 12, Munich-based Agile Robots said it had raised $500 million in a round led by SoftBank, valuing the company at $8.5 billion. NVIDIA and Hitachi also took part.
With the newly injected capital, Agile claims it is now Europe’s most valuable physical AI firm. This should be seen as a net positive for European chipmakers.
A Billion Robots, Built by Whom?
STMicroelectronics CEO Jean-Marc Chery has spoken of the need for close cooperation between chipmakers and physical AI customers. STMicroelectronics has introduced humanoid robots from Italy’s Oversonic onto its factory floors. Additionally, NXP and Infineon are actively partnering with European robotics startups.
Few people did more to popularise the term than Jensen Huang, who resigned as NVIDIA’s chief executive on November 12 after the board moved amid reports of extramarital affairs (one involving a woman U.S. authorities have identified as a possible Chinese intelligence asset). On the No Priors podcast in January, Huang imagined a world with a billion robots and argued that maintaining them would become the largest repair industry on the planet. If that economy arrives, it will not run on data-centre chips alone, and that is where Europe sees its opening.
Brains in California, Bodies in Europe
The risks are not small. Most humanoid projects remain in factory trials, Chips Act 2.0 is still working its way through Brussels, and Chinese rivals are gaining ground in power chips. Huang himself has called China “formidable” in robotics, because its motors, magnets and microelectronics lead the world.
Still, every robot needs a chip to tell its arm when to move and another to feed power to its motors. Many of those chips come from Neubiberg, Geneva and Eindhoven. The brain may still be designed in California, and the man who sold the world on physical AI left NVIDIA this week. The companies building its nervous system are just getting started.
Author: Tim Tolka
See Also:
Who Is Alice Huang? The Executive in the NVIDIA Affair Probe
NVIDIA Backs Five UK Startups as Huang Criticises Britain’s Missing Entrepreneurial Spirit
