Skip to content

SBTC’s AI-powered EV battery disassembly research wins top prize at NVIDIA GTC 2025

Cleantech

4 April 2025

The Swiss Battery Technology Center (SBTC) has received international recognition for its pioneering research into automated EV battery recycling, winning first place at the NVIDIA GTC Poster Award 2025 in San Jose. Francesco Sarno of the SBTC has earned first place at the NVIDIA GTC Poster Award 2025 in San Jose, recognising a project that aims to revolutionize electric vehicle (EV) battery recycling through automation, artificial intelligence, and digital simulation. | © Switzerland Innovation Park Biel/Bienne

The Swiss Battery Technology Center (SBTC) has received international recognition for its pioneering research into automated EV battery recycling, winning first place at the NVIDIA GTC Poster Award 2025 in San Jose.

Francesco Sarno of the SBTC has earned first place at the NVIDIA GTC Poster Award 2025 in San Jose, recognising a project that aims to revolutionize electric vehicle (EV) battery recycling through automation, artificial intelligence, and digital simulation. | © Switzerland Innovation Park Biel/Bienne

EV battery disassembly is a complex and high-stakes task, complicated by the diversity of battery designs and risks associated with charged components. SBTC’s goal is to create a fully automated robotic system capable of dismantling batteries efficiently and safely, while recovering valuable raw materials and supporting the development of a circular economy.

To address the lack of publicly available datasets, the research team developed synthetic data using NVIDIA Omniverse. Digital twins of battery components were created and deployed in realistic virtual environments, enabling the training of AI models for object detection, segmentation, depth estimation, and 6D pose calculation.

The team also integrated reinforcement learning (RL) within the simulated environment to teach robots the necessary disassembly skills. These models were then successfully transferred from simulation to real-world scenarios using ROS2, a robotics middleware framework. Results include high-precision component identification, with pose estimation accuracy within three millimeters and detection performance exceeding 96%.

The project builds on previous work by the SBTC, detailed in 2024 and highlighted in our article Western Switzerland at the forefront of the battery technologies of tomorrow. The Center, hosted at the Switzerland Innovation Park Biel/Bienne, continues to position Western Switzerland as a leading hub for next-generation battery technologies and recycling solutions.

The award from NVIDIA GTC, a key global conference in AI and high-performance computing, further validates SBTC’s contribution to sustainable technology and positions it as a reference point for advanced robotics in battery recycling.