Nvidia Unveils AI Science Software Suite and Vera Rubin Supercomputers at ISC 2026

Nvidia Brings Agentic AI to the Lab

Nvidia opened the ISC High Performance 2026 conference in Hamburg, Germany, on Monday by unveiling a suite of AI-powered software tools aimed squarely at accelerating scientific discovery. The announcements span everything from chemistry and materials simulation to the processing of telescope data in the search for dark matter and dark energy, signalling a deliberate push to embed AI agents into the heart of scientific computing.

Three New CUDA-X Releases

At the centre of the announcement are three new CUDA-X software releases: cuPhoton, DAQIRI, and an update to NVIDIA ALCHEMI. cuPhoton is a reference code for processing the enormous datasets produced by telescopes and dark energy surveys, developed in collaboration with Princeton and Harvard universities. DAQIRI is a new library focused on data acquisition for experimental science. ALCHEMI provides AI-powered tools for chemistry and materials discovery using machine-learning interatomic potentials, which Nvidia says can generate simulation data up to 1,600 times faster than traditional methods.

Nvidia frames these tools as part of a broader move toward agentic AI in research, where AI agents work alongside simulation and data processing. The company argues this requires a new scientific computing stack that connects agents, simulation, and AI, powered by its Vera Rubin and Grace Blackwell platforms along with Quantum InfiniBand networking.

The Vera Rubin Platform

Alongside the software, Nvidia detailed its Vera Rubin platform, which delivers more than 7 exaflops of AI performance and 5 petaflops of native double-precision computing in a single rack. Major research institutions, including the Leibniz Supercomputing Centre, the National Energy Research Scientific Computing Center, and Los Alamos National Laboratory, are building next-generation supercomputers on the platform.

"Scientific discovery is now a race between the complexity of the world's greatest challenges and the computing systems built to solve them," said Jensen Huang, Nvidia's founder and CEO. "NVIDIA Vera Rubin is a new instrument for science — a rack-scale supercomputer that brings simulation, AI and data processing together to help researchers and industries design and discover faster than ever."

A Record European Buildout

The conference, running through 26 June, also saw Nvidia announce a record 35 new AI supercomputers in development across Europe, the continent's largest single-year expansion in supercomputing infrastructure. Vera Rubin NVL4-based systems are expected to be available from global system manufacturers in the fourth quarter of this year, with Dell among the first to ship hardware built on the new architecture. Taken together, the announcements underline Nvidia's ambition to position itself not just as a chip supplier but as the foundational platform for the next era of computational science.