Fujitsu Limited and the Center for Quantum Information and Quantum Biology at The University of Osaka have announced a significant advancement in quantum computing technology, aimed at enhancing the industrial application of quantum computers. This development focuses on the early fault-tolerant quantum computing (early-FTQC) era.
The breakthrough combines the third version of the STAR architecture, a highly efficient phase rotation gate quantum computing framework, with a new molecular model optimization technique. This combination has substantially reduced computational resource requirements, enabling energy calculations for chemical material design, such as catalyst molecules, within practical timeframes using early-FTQC quantum computers.
These calculations, previously impractical with current computers, would have taken millennia even with earlier versions of the STAR architecture. The new technologies are poised to address societal challenges by accelerating drug discovery, enhancing ammonia synthesis processes, and advancing carbon recycling technologies.
The STAR architecture, initially established in March 2023 and followed by a second version in August 2024, has evolved to significantly expand computational scale. The third version improves computational accuracy by integrating phase rotation gates with logical-T gates, facilitating more complex molecular calculations with the same qubit count and lowering error rate requirements for qubits.
The molecular model optimization technology, designed for use with STAR architecture ver. 3, refines existing methods by decomposing molecular models into numerous terms and selectively applying techniques like time evolution and random sampling. This approach minimizes the number of gates in quantum circuits, substantially reducing computation time compared to conventional methods.
Validation of these technologies involved energy calculations for three molecules: Cytochrome P450, iron-sulfur clusters, and ruthenium catalysts. These calculations, infeasible with classical computers due to memory limitations, were demonstrated to be feasible on early-FTQC quantum computers with reduced qubit requirements and lowered physical error rate requirements.
Fujitsu and The University of Osaka confirmed that computation times could be reduced to approximately 35 days with a qubit error rate of 0.10% and 10 days with 0.01%. Further reductions are anticipated with future improvements in quantum computing error rates and the use of parallel computing.
Looking ahead, Fujitsu and The University of Osaka plan to continue advancing the STAR architecture and molecular model optimization technology, aiming to expand the practical application range of quantum computing in various industrial fields, including drug discovery, new material development, and finance.