Six Japanese companies, led by SoftBank, have launched a joint research programme to build a materials development platform that combines quantum computing with artificial intelligence, after their proposal was selected by Japan’s New Energy and Industrial Technology Development Organization, known as NEDO, the companies announced on October 28, 2026. The group — TOPPAN Holdings, AISIN, Asahi Kasei, SoftBank, AGC and DIC — will work together from fiscal 2026 through fiscal 2028 to test whether quantum calculations, paired with machine learning, can discover new materials faster and more cheaply than either approach alone. Rather than chasing a single breakthrough material, the project aims to build a reusable platform that spans batteries, displays, sensors and semiconductors.
Why Materials Discovery Needs A New Method
For Japanese chemical and materials manufacturers, finding new materials quickly has become an urgent competitive priority, the companies said. Two existing approaches each hit a wall. Data-driven methods, where AI predicts and explores new materials from historical data, struggle to secure enough high-quality, high-precision training data. Computational science, which simulates materials from physical laws, faces surging costs as accuracy and scale increase. The result is a familiar trade-off: the more precisely you want to model a material, the more expensive the calculation becomes.
Quantum computers offer a possible way through, because they can in principle simulate molecules and electrons, which are governed by quantum mechanics, more efficiently than conventional computers. But today’s machines are limited by qubit counts and error rates, so large-scale calculations using quantum computers alone are still not a reality, the companies acknowledged. The project’s answer is a hybrid: use quantum calculations where they help most, let AI fill gaps and cut costs, and check every prediction against real experiments.
What The Six Companies Will Each Do
The work is divided by each company’s strengths, and the division shows how broad the ambition is. TOPPAN Holdings will run hybrid methods combining quantum and classical algorithms for additive materials, develop combined quantum and AI methods, and feed experiment results back into the models. AISIN will verify quantum computing for superconducting materials, compare the results with conventional development methods, and forecast practical implementation timelines.
Asahi Kasei will test the platform on organic compounds such as electrolytes, polymer materials and inorganic materials such as electrodes, again with experiment feedback. SoftBank will use a computing platform connecting quantum computers with supercomputers to generate high-level data for material development, and will run demonstrations across domains including energy-related materials such as next-generation batteries. AGC will develop and evaluate methods combining quantum computing with existing computational science and AI for inorganic materials such as optical glass and organic materials such as electronic materials, aiming to derive design guidelines for faster exploration.
DIC completes the group by integrating quantum computing, simulation and data science to predict key properties of colorants and semiconductor materials before synthesis. Together, the six cover much of the industrial materials map, from battery chemicals to display glass.
Three Steps, With Experiments In The Loop
The project has three linked stages. First, the partners will verify where quantum computing actually applies, across major materials with high industrial demand, including materials for next-generation batteries, optical functional materials for new displays and sensors, and superconducting materials for next-generation devices. Second, they will develop the technologies that integrate quantum calculations with AI, expanding the scope of computational methods and testing solutions to the scale and cost limits of conventional approaches.
Third, they will synthesise materials based on simulation results, evaluate their performance, and compare experiment results against simulation data to improve predictive accuracy. That feedback loop separates a platform from a demonstration: a model that gets measurably better after each round of lab work is useful, and the companies aim to build a continuously evolving platform rather than a one-off study.
The project was chosen under NEDO’s solicitation for enhanced post-5G infrastructures, a reminder that materials research is increasingly treated as digital infrastructure. The same logic drives today’s AI hardware race, where chip and battery materials set limits on models, as in our report on Infineon’s €1.1 billion gallium nitride fab and our report on IBM and Rutgers’ self-checking quantum processor.
Why It Matters
This matters because materials discovery is one of the slowest steps in building better technology. A better battery chemistry, a more efficient display material or a new superconductor can unlock products that no amount of software tuning can, but finding those materials by trial and error takes years. If a quantum-plus-AI platform can narrow the search reliably, it could shorten development cycles across industries that have little in common except their dependence on new materials.
It also matters as a test of quantum computing’s near-term value. Rather than promising one machine that solves everything, the six companies bet that quantum’s first practical payoff comes from hybrid work: small quantum calculations, amplified by AI, checked by experiments, and repeated — a more modest, more believable claim worth watching through fiscal 2028.
Background
The six companies entered into a joint agreement following NEDO’s selection of their proposal, and will conduct the research from fiscal 2026, the year ending March 31, 2027, through fiscal 2028, working with specialised quantum computing institutions to verify applicability across diverse materials and build practical industrial value.
Source: SoftBank announcement
0 Comments