IBM, AMD and Oak Ridge National Laboratory launched a quantum error correction research platform on November 5, 2026. According to IBM, it brings quantum chips and standard computers together at its Poughkeepsie, New York, site.
IBM says the shared setup will help teams test ways to cut errors in quantum computers. It lets teams write and test programs without first having to build links between the two types of machine.
How quantum error correction needs help from other computers
IBM explains that quantum computers store data in units called qubits. These can make errors. To help catch those errors, teams spread data across many qubits and run checks for signs of trouble.
A standard computer reads those checks and works out which errors took place, according to IBM. The software doing that job is called a decoder. It must keep pace with the quantum chip.
IBM says teams often use quantum machines through the cloud, linked to standard computers at other sites. Those links add delays and make the work more complex. As a result, putting the machines together cuts the time needed to act on checks made while a program runs.
Nighthawk processors work alongside Umbra
The partners’ announcement says the platform includes IBM’s 120-qubit Nighthawk processor and Oak Ridge’s Umbra high-performance computing system. Umbra uses AMD Instinct MI355X graphics processors and fifth-generation AMD EPYC central processors.
According to IBM, teams can measure qubits while a program runs and act on the results in real time. They can also set the timing of signals and see the raw data from those checks.
The partners say teams will compare methods for quantum error correction. They will test how much computing power, time and how many qubits each method needs. The platform will also connect with the Oak Ridge Leadership Computing Facility in Tennessee.
A research platform for future systems
IBM says the work supports the Department of Energy’s Quantum Genesis Mission. The department wants quantum computers that can carry out science tasks despite errors. Its Q Competition targets work in chemistry, materials, physics and applied mathematics.
The department has planned up to $215 million for that contest. Funds for future years depend on approval from Congress. That sum covers the contest as a whole. It is not a stated budget for this platform.
IBM says lessons from the platform will guide work on its planned Starling fault-tolerant quantum computer. The company aims to deliver that system by 2029.
Sources: IBM announcement, IBM technical explanation, Department of Energy
