Google DeepMind has selected 15 European robotics startups for its first accelerator program, a three-month initiative that gives founders access to Google's AI infrastructure, hands-on mentorship from DeepMind engineers, and up to $350,000 in cloud computing credits. The program takes no equity. The cohort kicked off in London this week.
The companies come from across Europe: UK, France, Germany, Switzerland, Italy, Sweden, Denmark, Greece, Norway, and Romania. They are early-stage, they all have venture capital backing, and their products are not theoretical. They are building things that weld metal, sort waste, inspect ocean floors, test medical devices, construct homes, and navigate brain tissue.
The key thing Google gets from this is data and relationship capital. Each startup that builds on Google's Gemini Robotics models generates real-world data that feeds back into improving those same models. This is a well-established pattern: Google funds the builders, the builders produce data, and that data makes Google's models better. The startups get infrastructure that would otherwise cost them millions to access.
The broader picture is that robotics funding globally has exploded. Global robotics and physical AI startups raised $27.6 billion in 2025, more than double the $13.7 billion invested in 2024. European robotics startup equity investment more than doubled in 2025 to around €1.45 billion. Labor shortages, aging populations, and rising wages are pushing operators in manufacturing, logistics, and healthcare to accelerate automation decisions they might otherwise have delayed.
For business operators, a few companies in this cohort deserve attention. Danu Robotics (UK) is automating waste sorting using AI robotic systems, which touches recycling compliance and circular economy obligations increasingly written into procurement contracts across Europe. AUAR (UK) is deploying robotic micro-factories directly on construction sites to bring down homebuilding costs, a direct response to the housing shortage hitting every major European market. Staer (Sweden) uses existing cameras and sensors inside facilities to build 3D spatial maps that give robots a shared view of an environment and operators real-time visibility into their physical operations, without replacing any existing hardware.
ROBEAUTE (France) is the most striking entry: microrobots designed to travel through brain tissue to diagnose and treat neurological conditions. That one is years away from broad commercial deployment, but it signals how far the ambition in this cohort reaches.
What Google DeepMind is doing here reflects a wider move by large AI companies to stop just selling software and start embedding themselves into the physical world's supply chains. Agile Robots, a Munich-based company with over 20,000 robotics installations worldwide, announced a separate research partnership with DeepMind in March 2026, with the goal of deploying Gemini-powered robots in electronics manufacturing, automotive, data centers, and logistics.
For business operators, the practical read is this: the companies that will offer you AI-powered robotic services in the next three to five years are being built right now, and several of them are receiving direct support from one of the best-funded AI research labs in the world. Watching which of these 15 startups raise their next funding round, and in which sector, is a reasonable way to track where deployable robotics technology is actually heading.