Four months ago, a quantum computer simulated a protein with 303 atoms for the first time. This month, the same team simulated one with 12,635 atoms. That is a 40-fold increase, in four months.
Simulating how a drug molecule interacts with a protein at the quantum level is one of the most computationally demanding problems in chemistry. Classical computers can approximate it for small molecules, but the calculations required for biologically meaningful proteins, the kind actually involved in drug design, have been considered far beyond the reach of quantum hardware.
In late 2025, a collaboration between Cleveland Clinic, RIKEN, and IBM used quantum computing to simulate the electronic structure of Trp-cage, a 303-atom miniprotein, for the first time. It was described at the time as a milestone.
On May 5, 2026, the same team announced they had simulated two much larger protein-ligand complexes: T4-Lysozyme, at 11,608 atoms, and Trypsin, at 12,635 atoms, both modelled with bound drug-like molecules and immersed in water to replicate real biological conditions.
That is a 40-fold increase in the size of the system simulated, and a 210-fold improvement in accuracy on a key step in the calculation, achieved in four months.
The method is called quantum-centric supercomputing. Classical computers first break the massive protein-ligand complex down into smaller, computable fragments. IBM's 156-qubit Quantum Heron processors, running at facilities at both Cleveland Clinic and RIKEN, then calculate the quantum-mechanical behaviour of those fragments, using up to 94 qubits and nearly 6,000 quantum operations in certain parts of the simulation. The results are reassembled using two of the world's most powerful classical supercomputers, RIKEN's Fugaku and the University of Tokyo and University of Tsukuba's Miyabi-G.
A key technical refinement made this possible: the team restricted the most computationally expensive quantum calculations to a local sphere of 7 to 10 angstroms around specific atoms of interest, because the quantum entanglement relevant to the calculation diminishes sharply at greater distances. This let them avoid the impossible task of simulating the entire 12,635-atom system at full quantum detail.
Lead author Dr. Kenneth Merz of Cleveland Clinic described the moment: this is one of those things you dream about.
The researchers are clear about the current limitation. This method does not yet outperform the best classical computational chemistry methods for proteins. What it establishes is the trajectory: a 40-fold increase in scale and a 210-fold increase in accuracy, in four months, using a technique that is improving faster than classical methods are.
Source: IBM Quantum Blog and official press release, May 5, 2026. Cleveland Clinic, RIKEN, and IBM. Lead author: Dr. Kenneth Merz, Cleveland Clinic.