【PRESS RELEASE】 Exponential Acceleration of Nonlinear Differential Equation Solving on Quantum Computers - World’s First Practical-Level Readout Method Developed by Quemix and Sumitomo Rubber
- 田中拓哉
- 16 時間前
- 読了時間: 3分
Quemix Inc.
Tokyo, Japan — [November 27, 2025] — Quemix Inc. (President and CEO: Yu-ichiro Matsushita; Headquarters: Nihonbashi, Chuo-ku, Tokyo; hereinafter “Quemix”), a pioneering quantum technology startup specializing in quantum computing algorithms and software, and Sumitomo Rubber Industries, Ltd. (Head Office: Kobe; President: Satoru Yamamoto; “Sumitomo Rubber”) have jointly achieved a major breakthrough: an exponential speedup in computing nonlinear equations using a quantum computer(Note1).
This milestone was made possible by a new technique co-developed by the two companies that enables fast, low-cost readout of quantum computation results—a key challenge in quantum computing. The achievement marks a significant step forward toward the practical use of quantum computation.
Building on this success, the two companies will further advance their joint research to bring this method into real-world applications. Sumitomo Rubber aims to leverage this technology to dramatically accelerate the development of advanced materials such as high-performance rubber, one of its core R&D priorities.
Airflow readout during motorcycle riding
Nonlinear equations are a key benchmark for quantum computing.
Quantum computers are gaining attention in materials science, finance, and fluid analysis for their potential to perform massively parallel calculations far faster than classical computers. However, practical adoption has long been hindered by the readout problem: although quantum calculations are fast, extracting accurate numerical results is costly, often eliminating the expected advantage. This has led both academia and industry to question whether quantum computing can truly outperform classical methods in CAE (Computer-Aided Engineering), where computational demands are especially high.
The issue is particularly critical for nonlinear equations, which capture complex, non-proportional relationships at the core of CAE simulations—such as structural deformation, fluid vortices, and heat transfer. Despite theoretical indications of possible speedups, conventional quantum algorithms have struggled to deliver real benefits in practice because they require repeated measurements to read out each result.
Comparison of conventional vs. new methods (nonlinear equations)
（Computational results for the Burgers’ equation（Note 2）used in this demonstration.The figure shows the progression of computational steps from top to bottom.）
Through their joint research, the two companies have identified a practical solution to this long-standing challenge. By efficiently extracting waveform and periodic features embedded in quantum computations—such as Fourier components—and reconstructing solutions from only a small number of principal components, they achieved a readout method that maintains efficiency even as data size grows, enabling scalability to large-scale simulations. This new approach allows quantum computers to realize their theoretically predicted speedup for nonlinear equations in actual computations, effectively overcoming the readout problem and demonstrating an exponential acceleration.
Both companies view this achievement as a meaningful step toward shifting quantum computing from a “theoretical possibility” to a “practical reality.” It opens the door to a future in which quantum computers can offer a genuine competitive edge in fields that demand massive computational resources, such as materials property analysis, fluid simulation, and financial risk modeling. Going forward, the design of quantum algorithmsNote 3 will increasingly incorporate not only computational techniques but also readout protocolsNote 4, accelerating research that directly links to industrial applications.
The joint research also confirmed an exponential speedup in polymer SCFT (Self-Consistent Field Theory) calculationsNote 5—an essential computational method used for analyzing polymer structures.
These joint research results will be presented in a keynote session by researchers from Quemix and Sumitomo Rubber at Q2B 2025 Silicon Valley, a leading international conference on quantum technology, to be held December 9–11, 2025, at the Santa Clara Convention Center in California.
Q2B 2025 Silicon Valley
About Quemix Inc.
Quemix Inc., a consolidated subsidiary of TerraSky Co., Ltd. (Headquarters: Chuo-ku, Tokyo; CEO: Hideya Sato), conducts R&D in quantum computers, quantum sensors, and materials computation. Guided by its vision, “Realizing the Future Envisioned through Quantum Technology,” Quemix’s mission is to enable breakthrough innovations for companies leading the quantum era.
Since its establishment in 2019, Quemix has specialized in developing algorithms for Fault-Tolerant Quantum Computers (FTQC). The company developed and patented the Probabilistic Imaginary-Time Evolution (PITE®) algorithm, which has been mathematically proven to achieve quantum speedup in quantum chemistry calculations. As Japan’s pioneer in FTQC algorithms, Quemix aims to bring practical quantum computing applications to materials computation and simulation by 2028.
Note 1: As the problem size grows, the new method achieves exponential speedup over conventional approaches.
Note 2: A representative nonlinear partial differential equation in fluid dynamics.
Note 3: A quantum-mechanical computing method offering faster processing than classical techniques.
Note 4: A procedure for measuring quantum states to obtain computational results.
Note 5: A theoretical method used to predict polymer structures and properties, including self-assembly and nanoscale patterns.
For more information:
- Business Inquiries: Quemix Inc.
- Media Inquiries: TerraSky Co., Ltd. – PR Contact: pr@terrasky.co.jp
