Microsoft, NVIDIA and Infineon launch Physical AI growth lab for startups

Defining Physical AI

Physical AI represents the convergence of three distinct technologies: foundation models, simulation, and robotics. While standard AI often lives in a digital environment, physical AI moves that intelligence into the real world. This allows machines to interact with physical objects and navigate complex environments like factories or logistics networks.

The goal of this technology is to move from digital simulation to reality. By combining the reasoning power of large models with the movement capabilities of robotics, companies can create autonomous systems. These systems are designed to perform tasks that require both high-level planning and precise physical execution.

The Physical AI Growth Lab aims to bridge this gap for early-stage companies. It provides a structured environment where startups can refine their prototypes into market-ready solutions. By focusing on the intersection of artificial intelligence and industrial expertise, the program seeks to turn abstract concepts into functional machines.

The Partnership Stack

The accelerator is built on a foundation of four major partners, each providing a different layer of the technology stack. Microsoft provides the cloud infrastructure through Azure, offering startups USD 350,000 in credits. This allows companies to train and host their models on a global enterprise network without high upfront costs.

NVIDIA contributes the software and simulation tools necessary for development. Startups will gain access to NVIDIA SDKs, open models, and the benefits of the NVIDIA Inception program. These tools help founders simulate how a robot will behave in a virtual space before they ever build the physical hardware.

Infineon provides the physical hardware foundation that makes the systems actionable. Their contribution includes the semiconductor stack, which consists of power semiconductors, microcontrollers, sensors, and connectivity solutions. These components are the literal building blocks that turn intelligence into motion in the real world.

Who Can Apply

The program is specifically designed for startups based in Europe, the Middle East, and Africa (EMEA). It targets companies that are building for these regions and have a working prototype or a minimum viable product (MVP). The accelerator is looking for companies that have been founded within the last three years.

There are specific structural requirements for applicants to ensure they are ready for scaling. Teams must include at least one technical co-founder and can consist of up to 20 people. The program is equity-free, meaning startups do not have to give up ownership in exchange for the resources and mentorship provided.

The selection process will choose 10 to 15 startups to participate in the program. These companies will receive direct introductions to enterprise customers and connections to a network of over 200 investors. They will also benefit from mentoring provided by more than 170 deep tech experts.

Market Context and Timing

The timing of this launch aligns with a significant shift in the European industrial landscape. It is estimated that Europe will capture roughly 22.5% of the worldwide physical AI market by 2026. This growth is supported by regulatory clarity from the EU AI Act and Germany's High-Tech Strategy 2025.

The initiative is scheduled to take place between January and April 2027, utilizing a mix of remote and on-site sessions in Munich. This period allows startups to build a technical foundation and secure early enterprise customers. The goal is to help these companies become category leaders before the market consolidates around a few dominant platforms.

By providing access to capital, infrastructure, and industry experts, the partners aim to ensure the next generation of physical AI is developed in the EMEA region. This collaboration seeks to strengthen Germany's position as a hub for the adoption of autonomous robotic systems. The program highlights how real acceleration works by bringing together technology partners and established companies.