A research collaboration spanning the Cleveland Clinic Foundation, the RIKEN Center for Computational Science, and IBM Quantum has executed the largest heterogeneous quantum-classical (HQC) electronic-structure calculation to date, simulating full solvated protein–ligand complexes exceeding 12,000 atoms. Detailed in a technical preprint on arXiv, the team paired two 156-qubit IBM Heron r2 processors with three leadership supercomputers to collect 3.0 billion measurement shots across 21,006 circuits, achieving a greater than 40× increase in molecular size and up to a 210× improvement in fragment energy accuracy over previous quantum-embedding benchmarks.
| 12,635-Atom HQC Biomolecular Simulation Campaign Profile | ||
|---|---|---|
| System Dimension | Experimental Specification | Technical & Computational Benchmark |
| Target Biomolecules | Trypsin–benzamidine (12,635 atoms) & T4-lysozyme–n-butyl-benzene (11,608 atoms) | Up to 31,795 molecular orbitals in solvated environments; >40× larger than prior 303-atom Trp-cage benchmark. |
| Quantum Hardware | Two 156-qubit IBM Heron r2 QPUs (ibm_cleveland & ibm_kobe) | 21,006 circuits executed using up to 94 active qubits over 239+ QPU hours, gathering 3.0 × 109 samples. |
| Classical HPC Systems | Fugaku (152,064 nodes), Miyabi-G (1,104 GH200 nodes), & ROQUO (126 GB200 nodes) | Distributed GPU-accelerated Selected-Basis Diagonalization (SBD-G) reaching 72.5% strong-scaling parallel efficiency. |
| Fragment Solver Accuracy | TrimSQD & O(1)-scaling Embedded Wavefunction (EWF) | 0.10 mHa deviation vs. DMRG on trypsin’s largest 45-orbital fragment (210× accuracy gain over ExtSQD). |
| Biochemical Deliverable | First HQC protein–ligand binding energy prediction | Mixed-basis (3-21G / STO-3G) workflow resolved T4-lysozyme binding affinity at −21.07 kcal/mol. |
Accurately predicting how small-molecule drugs bind to target proteins requires solving the many-electron Schrödinger equation across thousands of solvated atoms, a regime where standard density functional theory approximations can misrank binding affinities and full-system wavefunction methods exceed classical memory limits. To bridge this gap, the researchers introduced a localized Embedded Wavefunction (EWF) formalism that partitions large macromolecules into spatially bounded atomic neighborhoods, reducing fragment construction costs from cubic O(M3) scaling to constant O(1) scaling per fragment.
For heavy-atom fragments spanning 13 to 45 molecular orbitals, the team developed TrimSQD, an extension of sample-based quantum diagonalization that iteratively filters and merges electronic configurations across distributed GPU subgroups before performing a final subspace diagonalization. Using a Prefect-based automated workflow controller to orchestrate asynchronous job queues between IBM Quantum Runtime and the ROQUO supercomputer, the pipeline processed 608 million measurement samples in 62.4 hours of wall-clock time without manual intervention.
The campaign was supported by the U.S. National Science Foundation, the National Institutes of Health, Japan’s New Energy and Industrial Technology Development Organization (NEDO), and the MEXT Supercomputer Fugaku program.
The full technical manuscript is available on arXiv here. Related QCR reporting on IBM’s managed hybrid workflows and SQD scaling includes IBM’s QDC 2026 platform announcements here, and IBM and Mitsubishi Chemical’s 70-qubit generative quantum eigensolver study here.
In GQI Portal
The players behind the news
The team that writes QCR tracks every company, deal and technology in the GQI Factory, GQI's verified database of the quantum industry. Next up: every story linked to its players, coming to QCR's paid plans.
- Players Companies and institutions across the quantum industry, by segment.
- Scorecards How the players compare on hardware, software, funding and more.
Newsletter
QCR Alerts in your inbox
The latest reporting and analysis from Quantum Computing Report. Free, unsubscribe any time.