Nvidia experiencing advancement, IBM on the brink of quantum decimalization, D-Wave contributing significantly.
The JUPITER supercomputer, located at Forschungszentrum Jülich, is making waves in the world of high-performance computing (HPC). Positioned as one of the world's most powerful supercomputers, JUPITER is a modular exascale system excelling at AI, scientific simulations, and climate modeling [1][3][5].
While the D-Wave quantum annealer's role in the JUPITER project is not explicitly detailed, the supercomputer is designed to support hybrid quantum-classical workflows. This includes the use of powerful tools like NVIDIA CUDA-Q and cuQuantum SDK, which advance quantum algorithm research and are compatible with gate-model quantum simulators and hardware development [3].
JUPITER, with more than 1 exaFlop/s in double precision and up to 80 exaFlop/s in AI-optimized modes, is a classical exascale HPC system. It does not appear to directly incorporate D-Wave’s quantum annealer hardware [1][3][5].
In contrast, IBM is focusing on drug development and materials research with its quantum computing efforts. The tech giant has announced plans to perform 100 million quantum operations with 200 logical qubits by 2029, aiming for a further tenfold performance increase by 2033 [2].
The difference between D-Wave's quantum annealer and IBM's universal gate-model quantum computers is significant. D-Wave's annealer specializes in quantum annealing, optimized for solving certain combinatorial optimization problems and sampling tasks, while IBM's systems pursue broad quantum algorithm applications with the goal of fault-tolerant universal quantum computing [6].
In large-scale HPC like JUPITER, quantum annealers like D-Wave's are envisioned as specialized accelerators complementing classical supercomputers for niche optimization tasks, while IBM's gate-model systems pursue a wider range of quantum-enhanced computing tasks.
An example use case for this combination is climate simulations, where the supercomputer performs the central modeling, and the quantum computer handles partial tasks such as complex optimization issues, potentially involving atmospheric chemistry [4].
Dr. Thomas Eickermann from the Juelich Supercomputing Centre stated that although annealers are not universal quantum computers, they have a high level of maturity and will continue to play a role in Jülich's user infrastructure [4].
The German supercomputer, Jupiter, ranks 5th among the world's most powerful supercomputers, performing over 60 billion calculations per watt. This combination of classical and quantum computing could lead to breakthroughs in artificial intelligence and optimization, particularly in logistics.
It is important to note that the publication discusses financial instruments of Nvidia and IBM, which could benefit from the potential price development resulting from the publication. IBM has already saved billions in quantum computing and AI, according to a new statement from an IBM manager [7].
Moreover, IBM has invested in a platform that could make all new robo-cars available worldwide and potentially multiply its profits by 2040. Additionally, IBM has purchased a Xiaomi beneficiary that could benefit from the massive rollout of 800-volt fast-charging technology by 2030 [8].
Jülich has also purchased an upgrade of D-Wave's quantum annealer, and IBM researcher Heike Riel has given an interview about the company's quantum computing progress in the latest HSR issue 23/2025 [9][10].
Dr. Eickermann confirmed that the Jupiter supercomputer is "on track" [11]. This duo of classical and quantum computing could revolutionize various fields, from climate modeling to artificial intelligence, and logistics.
The JUPITER supercomputer, with its capabilities in AI, is designed to support hybrid quantum-classical workflows, incorporating tools like NVIDIA CUDA-Q and cuQuantum SDK for quantum algorithm research and development.
In contrast, the role of D-Wave's quantum annealer in JUPITER is not explicit, but it is envisioned as a specialized accelerator for niche optimization tasks, potentially complementing AI-optimized modes within the supercomputer.
