Overview
- A joint announcement Monday, November 16, 2026, said IBM, NVIDIA and Fermi National Accelerator Laboratory will colocate IBM’s 120‑qubit Nighthawk processor with NVIDIA DGX B200 accelerated classical systems inside Fermilab’s Superconducting Quantum Materials and Systems (SQMS) center.
- Fermilab will test cryogenic interconnects and system integration so multiple quantum processors can be linked and supported by growing pools of classical compute for control and decoding.
- NVIDIA unveiled a new decoder called TCBP‑JOBS and said it can cut some research decoding tasks from 12 weeks to 45 minutes, a headline claim the partners describe as an upper bound that applies to specific workloads.
- The collaboration focuses on speeding error correction, evaluating scalable modular architectures, and running scientific applications in high‑energy and nuclear physics that need tight quantum‑classical workflows.
- The effort builds on prior ties—IBM’s 2024 SQMS partnership and NVIDIA’s CUDA‑Q work—and will test whether close coupling of high‑throughput chips and supercomputer‑class GPUs can make practical progress toward fault‑tolerant, modular quantum systems.