AMD CPUs, GPUs will power PsiQuantum’s quantum computing platform

AMD and PsiQuantum Join Forces on Classical‑Quantum Infrastructure

PsiQuantum, a leader in photonic fault‑tolerant quantum computing, announced a multi‑layer collaboration with AMD. The partnership expands an existing relationship that already uses AMD Versal FPGAs for error‑correction hardware, adding AMD EPYC CPUs, Instinct GPUs, and the ROCm software stack to the quantum‑classical stack. The goal is to build the compute backbone for utility‑scale quantum machines and to enable hybrid quantum‑classical applications, starting with drug‑discovery chemistry.

Why Classical Power Matters for Photonic Quantum Computing

Photonic quantum processors excel at low‑error operations but still need massive classical resources for tasks such as state preparation, read‑out, and especially real‑time error correction. By leveraging AMD’s high‑performance CPUs and GPUs, PsiQuantum can run sophisticated decoding algorithms—like the PRISM machine‑learning decoder—at the sub‑microsecond latencies required for fault‑tolerant operation. The inclusion of Versal adaptive compute accelerators also allows the team to prototype custom error‑correction pipelines on programmable hardware before moving to ASICs.

Early Demonstrations: Chemistry on AMD Hardware

The announcement cites a recent computational‑chemistry study that used AMD GPUs to simulate covalent inhibitors relevant to oncology drug discovery. While the results are still preliminary, they illustrate how existing AMD hardware can already accelerate quantum‑inspired workloads, providing a testbed for the larger hybrid workflow the partnership envisions.

Strategic Fit and Future Plans

Both companies frame the collaboration as a response to the U.S. Quantum Genesis Priority Applications, which highlights drug discovery as a high‑impact early use case. AMD’s CTO, Mark Papermaster, emphasized that leadership in classical computing is essential for scaling quantum systems—an argument that aligns with the broader industry view that quantum advantage will emerge from tightly coupled quantum‑classical pipelines.

PsiQuantum and AMD plan to reveal more technical details at the SC26 supercomputing conference in Chicago later this year. The partnership also includes an equity investment from AMD into PsiQuantum, signaling confidence in the long‑term commercial potential of photonic quantum computers.

What to Watch Next

• Error‑Correction Performance: How effectively the PRISM decoder runs on AMD GPUs and FPGAs will be a key metric.
• Hybrid Application Development: Early use cases in drug discovery could set a template for other domains.
• Scaling Roadmap: The joint effort will need to demonstrate that classical infrastructure can keep pace with the projected growth of photonic qubit counts.

The collaboration underscores a growing trend: quantum startups are increasingly dependent on established HPC vendors to bridge the gap between experimental qubits and practical, large‑scale quantum applications.

--- This article is based on the press release from PsiQuantum and AMD. No independent verification of technical claims has been reported.