HomeTech NewsNVIDIA, IBM and Fermilab Plan a Quantum Codesign Platform

NVIDIA, IBM and Fermilab Plan a Quantum Codesign Platform

NVIDIA, IBM and the US Department of Energy’s Fermi National Accelerator Laboratory have announced plans for a modular quantum-classical codesign platform that would place NVIDIA DGX B200 systems alongside IBM quantum computers, including machines powered by IBM Quantum Nighthawk. NVIDIA’s 16 November newsroom release said the company “announced plans for new collaborations” with IBM and Fermilab. IBM, the same day, said the three “intend to launch” the platform. Fermilab described it as a proposed, intended collaboration that “would” unite quantum processors, accelerated computing and the high-energy and nuclear physics community. This is a plan, not the opening of a finished facility.

The partners said researchers would use the platform to develop and evaluate error-correction codes, decoding algorithms and quantum-classical applications in high-energy and nuclear physics, and to study how classical systems can scale alongside multiple linked quantum processors. Fermilab said it would lead engagement with that scientific community and help coordinate access through its HEPCloud infrastructure. In a later phase, it said, its Superconducting Quantum Materials and Systems Center intends to add ultracoherent superconducting cavity hardware to the platform.

Anna Grassellino, Fermilab’s chief technology officer and director of the SQMS Center, called it an “intended” and “proposed” platform and “an important step toward the capabilities and community needed for a future DOE quantum computing user facility.” Jerry Chow, IBM Fellow and chief technology officer of quantum-centric computing, said: “With Fermilab and NVIDIA, we plan to explore how these real-world problems can help us shape the modular, interconnected architectures that will define the next generation of quantum computers and accelerated classical computing.” Krysta Svore, vice president of applied research for quantum computing at NVIDIA, said the collaboration would give scientists “more powerful tools to investigate quantum designs and develop applications across these systems.”

A separate NVIDIA decoder claim, and a Slurm interface

Separately, NVIDIA said a new decoding algorithm, TCBP-JOBS, now in the CUDA-Q QEC 0.9.0 library, “delivers up to a 13,000x speedup for quantum error-correction research” and cuts “the computation time for researching new quantum computing designs from 12 weeks to just 45 minutes.” Those figures are NVIDIA’s claims about a research decoder, not measured results from the planned codesign platform.

NVIDIA also said it is expanding the open-source Slurm workload manager with a quantum interface, and that IBM is contributing integration through its Quantum Resource Management Interface, so that centres can schedule quantum processors alongside CPUs and GPUs. IBM said the interface is vendor-agnostic and was developed with NVIDIA, Pasqal and the open-source community.

NVIDIA also pointed to IBM and RIKEN work that, it said, modelled protein-ligand complexes of up to 12,635 atoms and has been named a finalist for the 2026 ACM Gordon Bell Prize. That chemistry project is a neighbouring result, not a demonstration of the new Fermilab platform.

The announcement sits beside earlier official work on quantum error correction, including the ORNL, IBM and AMD codesign platform, which Fermilab said the new effort would complement, IBM Quantum’s directed-execution tools, and NVIDIA’s earlier simulation of 104 qubits on the JUPITER supercomputer. Those are related research efforts, not this planned facility.

A planned platform is not a standing quantum facility

The three organisations have published a plan. The verbs in their own releases are “plans,” “intend,” “would” and “proposed.” Grassellino’s reference to a future DOE quantum computing user facility is an ambition, not a decision that such a facility has been authorised or opened. Nothing in the announcements names a site completion date, a user-application process, or a first scientific result from the colocated DGX B200 and Nighthawk systems.

The “up to 13,000x” decoder figure, and the comparison of 12 weeks with 45 minutes, should also be kept in their proper box. NVIDIA presents them as speed-ups for error-correction research workloads in CUDA-Q QEC 0.9.0, against “existing leading algorithms.” They are not a claim that the planned IBM–Fermilab platform has already decoded live quantum experiments 13,000 times faster. Treating the software announcement as proof that the platform is already running would collapse two different disclosures.

What to watch next is whether the partners publish a schedule, a location, or first results from the colocated systems — the steps that would move this from a planned codesign platform to a working one.

Sources

Related stories

Follow @tsnmediahub on X for breaking AI news as it happens.

Follow on X

Share this story

Latest stories

  1. TensorWave and OpenAI Agree a 457MW AMD MI455X Deal
    AI Chips & Compute

More in this category

Latest stories

  1. TensorWave and OpenAI Agree a 457MW AMD MI455X Deal
    AI Chips & Compute

Explore the TSN Atlas: Every topic we cover, mapped. Click a bubble to dive in.