NVIDIA Rolls Out GPU Tools and Partnerships Across European Industry

A cluster of announcements from NVIDIA and European industrial partners covers accelerated simulation, new software libraries, and AI collaborations across semiconductor, quantum computing, life sciences, and manufacturing. imec benchmarked a 40x speedup in lithography simulation using Synopsys S-Litho on NVIDIA hardware, and Infineon achieved a 12x speedup in 3D breakdown simulation for power devices. NVIDIA released cuDSS, a generally available CUDA-X library that offloads sparse linear system solving to GPUs; Applied Materials achieved up to 10x speedup in its Ginestra platform using it, ZEISS SMT achieved a 4x speedup in an EUV lithography plasma simulation, and Dassault Systèmes reports 8-10x speedup in Abaqus electromagnetic solvers. Mistral and NVIDIA announced a collaboration to bring AI to industrial engineering workflows, including faster simulation, open benchmarks for physics AI, and agentic AI evaluations for CAE and EDA. NVIDIA and the Baker Lab released RFD4-Proteina, an open protein design model that generates designs roughly 13x faster than its predecessor across more than 10 protein design tasks. NVIDIA also released Ising, a family of open models for quantum hardware calibration and error correction, with IQM, SEEQC, and AWS among early adopters. L'Oréal integrated NVIDIA ALCHEMI and Orbital Industries' OrbMol-v2 molecular simulation model into its cosmetic ingredient screening workflow, with OrbMol-v2 reported as five times faster and 21% lower error than the next-best model of its kind.