AMD Surpasses NVIDIA in the Quantum Computing Race — Company Runs IBM Algorithm on Conventional Chips

IBM has announced a significant milestone in quantum computing, with AMD playing a central role. Rather than relying on dedicated quantum hardware, the team ran a quantum error correction (QEC) algorithm on conventional FPGA chips — achieving performance roughly ten times faster than anticipated.

According to Reuters, IBM researchers executed their QEC algorithm on AMD FPGAs (the FPGA line brought into AMD through its acquisition of Xilinx). Jay Gambetta, Director of IBM Research, emphasized that the experiment validated the approach’s practical viability: “Our algorithm works not only in theory, but also on inexpensive, widely available AMD chips.”

This outcome matters because it demonstrates that classical components can tackle problems once thought to require quantum-only hardware. In quantum systems, information is stored in qubits — highly fragile units that are susceptible to environmental disturbances. QEC techniques detect and correct errors without collapsing the qubit’s state, which is essential for trustworthy quantum computations.

AMD’s FPGA chips are well-suited for these tasks because of their configurability and low latency, making them ideal for real-time feedback in quantum setups. NVIDIA is also pursuing work in this domain with its DGX Quantum initiative and the CUDA‑Q platform, but its solutions do not rely on FPGAs. That distinction gave AMD an edge: it was the first to run a QEC algorithm on commodity hardware that contains no quantum components.

Another decisive factor was AMD’s purchase of Xilinx, a leader in FPGA manufacturing — a merger that helped enable IBM’s experiment. While much of the industry is focused on AI, IBM and AMD’s work suggests the next technological leap could begin with simple, affordable chips.

Source: iXBT.games