Honda R&D and Quemix Achieve DFT Speedup With Quantum Computing

Quemix Inc. and Honda R&D jointly issued a press release on June 3 detailing the development of a new quantum algorithm capable of calculations based on Density Functional Theory (DFT), a core technology for Materials DX. Accelerating DFT calculations is one of the foundational technologies in computational materials science. This joint research marks the first successful creation of a quantum algorithm designed to speed up DFT, potentially enabling simulations of extremely large-scale systems that would have been impossible using conventional computers. The advance promises to accelerate the development of new materials, an area of significant focus for Honda R&D, by addressing a critical need for faster computation rather than solely focusing on increased accuracy.

DFT as Core to Materials DX and Quantum Computing Expectations

The demand for accelerated materials discovery is driving a convergence of quantum computing and computational materials science, with Density Functional Theory (DFT) calculations now positioned as a critical focal point. Quemix Inc. overcame technical challenges to enable DFT calculations on extremely large-scale systems that would have been impossible using conventional computers. DFT’s central role in the shift from experience-based materials development to computationally driven design underpins the significance of this breakthrough. While recent materials research has increasingly focused on improving the accuracy of quantum simulations, the joint team recognized a more pressing need: computational speed for weakly correlated materials, prevalent in industries like drug discovery and semiconductor manufacturing. Most research on quantum-computing-based materials simulations has focused primarily on improving calculation accuracy.

The core innovation lies in a quantum algorithm that circumvents the traditionally nonlinear Gram, Schmidt orthogonalization process within DFT, and directly calculates total energy using Quantum Phase Estimation (QPE) circuits, eliminating the need for explicit readout of electron density distributions. Demonstration experiments confirmed that as the computational scale increased, the calculation time decreased exponentially, alongside accuracy comparable to conventional DFT methods in determining key material properties. Honda R&D anticipates this will accelerate the development of new materials, and the companies plan to implement the algorithm on actual quantum hardware, aiming to contribute to true Materials DX capable of dramatically shortening development cycles.

Novel Quantum Algorithm Avoids Gram-Schmidt Orthogonalization

Much quantum materials research has prioritized improving calculation accuracy, but a critical bottleneck remained: the sheer computational demand for simulating larger, more industrially relevant systems. Conventional DFT calculations are often sufficient in accuracy for weakly correlated materials, but lack the speed needed for extensive modeling; existing quantum algorithms hadn’t addressed this practical need for faster computation. A significant obstacle was the nonlinear Gram, Schmidt orthogonalization process integral to DFT, proving difficult to integrate with the linear operational framework of quantum computers. Determining total energy traditionally required sequential readout of electron density distributions, adding substantial computational cost. Researchers at Quemix Inc. successfully created an algorithm capable of exponentially accelerating Density Functional Theory (DFT) calculations by introducing a novel quantum approach that avoids the Gram, Schmidt process altogether. Representatives from both companies stated that they anticipate broader impact beyond specialized materials and towards practical applications in semiconductors and battery materials.

Demonstrated Exponential Acceleration and Accuracy of the Algorithm

Quemix Inc. and Honda R&D are pushing the boundaries of materials science with a newly developed quantum algorithm designed to dramatically accelerate Density Functional Theory (DFT) calculations. This isn’t merely about incremental speed improvements; the collaborative team has achieved exponential acceleration in DFT calculations, a feat previously unattainable with conventional computers for extremely large-scale systems. The significance lies in DFT’s central role in Materials DX, a shift towards computationally driven materials development, and the algorithm’s potential to unlock simulations of materials far beyond the reach of current technology. A key hurdle overcome was the inherent incompatibility between DFT’s nonlinear computational processes and the linear operations of quantum computers. Specifically, the traditionally challenging Gram, Schmidt orthogonalization process has been bypassed entirely by this new approach. Demonstration experiments conducted on an emulator confirmed that “as the computational scale increased, the calculation time was shown to decrease exponentially compared with conventional algorithms.”

Beyond structural analysis, the team successfully calculated electronic band structures, providing insights into a material’s electrical properties. This advancement broadens the scope of quantum computing applications, extending beyond specialized materials to encompass semiconductors and battery materials, and promises to accelerate the development of next-generation technologies.

Quemix Inc. DFT enables prediction of material properties at the atomic level, but its computational demands have historically limited the scale of simulations. Crucially, they also established a method for directly calculating total energy from Quantum Phase Estimation circuits without requiring explicit readout of electron density distributions, a process that previously created a major bottleneck.