News | RPI’s IBM Quantum System One gets Nighthawk r1 chip - poly.rpi.edu

RPI’s IBM Quantum System One was unveiled at the Voorhees Computing Center in April 2024 as the first IBM Quantum System One installed at a university, which originally ran on IBM's 127-qubit Eagle processor. This system has now been upgraded from the Eagle processor to the 120-qubit Nighthawk r1, as of August 15, 2026.In an interview with The Polytechnic, RPI Future of Computing Institute Research Scientist Dr. Cameron Cogburn described running a calculation that modeled the interaction between a kink and an anti-kink. Kink-anti-kink potentials are a simplified system used to study quantum chromodynamics, which is the theory of the strong nuclear force. Circuit noise on the Eagle chip had caused the result to diverge from the expected physical interactions; whereas on Nighthawk r1 the output closely matched the numerical solution. The difference between the two chips is how their qubits are physically connected. IBM's previous heavy hex architecture, on the Eagle and Heron, limited most qubits to two or three direct neighbors. Nighthawk arranges qubits in a two-dimensional grid, giving each qubit three to four neighbors. This reduces the need for SWAP operations, where a qubit that is not adjacent to the one it needs to interact with is moved into the correct position through a series of operations, each of which introduces additional noise into calculations. In his kink-and-anti-kink calculation, Dr. Cogburn was able to arrange data and measurement qubits so that few or no swaps were required with the new chip, which he said was a primary factor in the improved result. The qubits on Nighthawk also have longer coherence times than the previous generation, independent of the layout change. The tradeoff here is runtime. Nighthawk also has a longer default "repetition time," the interval a qubit needs to reset between operations before the next one can run without introducing additional noise. This is expected to improve with the r2 chip revision IBM has released on its cloud platform, which could reach RPI within a year if approved, although Dr. Cogburn did not confirm a timeline.The quantum computer is connected to the RPI supercomputer at the Rensselaer Technology Park through a direct fiber optic link, separate from the cloud connection used for most jobs. Dr. Cogburn said he avoids the term "quantum advantage" for most current applications, while noting that a mathematically proven speedup over classical computing exists for a small number of algorithms, such as Shor's algorithm for factoring large numbers. For most other problems, he considers the more accurate framing to be "computational advantage," in which quantum processors are used alongside classical computers rather than as a replacement for them, comparable to how GPUs are used alongside CPUs. Classical tensor networks can solve many of the same problems quantum computers target, particularly for systems that can be modeled in one dimension, and that tensor networks become computationally unfeasible only for two and three-dimensional systems past a certain size. Dr. Cogburn said his group is developing methods that use those classical calculations to partially process, or “warm start,” a problem before handing the remainder to the quantum computer. The Poly asked a question about reports circulating among faculty in the Department of Computer Science and other academic departments regarding a possible second RPI quantum computer. Dr. Cogburn did not confirm a specific system, funding source, or timeline, but said he expects RPI to have access to a next generation quantum system within the next several years. He added that funding structure, potential cost-sharing with New York State or other State University of New York institutions, and the physical location of any new system relative to the existing supercomputer are undetermined.Dr. Cogburn described RPI's collaboration with IBM as one of several industry and national partnerships. He also cited a funded collaboration with Western Digital, in which a student researcher works on cryptography-related problems using both quantum and classical high performance computing resources. referenced experimental quantum hardware research at RPI led by Department of Physics, Applied Physics, and Astrophysics Professor Xiangyi Meng, and noted that outside foundries are able to fabricate qubit chips for research groups that design them; the challenge there lies in building support systems that are robust enough to host an operational quantum system.For those interested in learning more about quantum computing, Dr. Cogburn suggests Qiskit, the IBM software stack for quantum computing and algorithms. Qiskit has a Youtube channel and textbooks for students to access. Those who have an RPI email can send in a request and receive access to the computer, allowing the user to run small programs on the computer along with simulations. RPI also offers a Quantum Computing Minor, structured around higher level classes that allow you to work with RPI’s quantum computer. The Department of Computer Science offers CSCI 4620 Quantum Computer Organization, the most robust course to learn about the quantum computer through hands-on experience.