A joint research team from The University of Osaka's Center for Quantum Information and Quantum Biology and Fixstars Corporation has successfully demonstrated one of the world's largest classical simulations of iterative quantum phase estimation circuits for quantum chemistry. The achievement utilized 1,024 graphics processing units to simulate systems far exceeding the previous 40-qubit limit. This breakthrough was presented at NVIDIA GTC 2026 in San Jose, California, from March 16 to 19, 2026. The research addresses the urgent need for advanced quantum chemical calculations to solve complex challenges in drug discovery and climate change materials development. While fault-tolerant quantum computers are anticipated to be the key to these solutions, developing and validating the necessary algorithms requires rigorous testing before such hardware becomes widely available. Quantum phase estimation is a core subroutine in many of these algorithms, yet simulating it on classical hardware has historically been constrained by qubit count and computational time. The team, led by Professor Wataru Mizukami and including researchers Shoma Hiraoka, Sho Nishida, and Yusuke Teranishi, implemented Iterative Quantum Phase Estimation in a specialized simulator known as chemqulacs-gpu. A major hurdle was the massive computational demand. To overcome this, the group developed a novel parallel computing technology optimized for large-scale GPU clusters. They executed the simulation using up to 1,024 NVIDIA H100 GPUs on the AIST's ABCI-Q supercomputer system. This approach allowed them to bypass conventional computational bottlenecks and run simulations significantly deeper than previously possible. Professor Mizukami noted that managing 1,024 GPUs simultaneously within a 48-hour window was technically demanding. The team faced unexpected issues during the process but persevered with support from the ABCI-Q operations staff. The successful completion of this simulation represents a significant milestone in preparing for future quantum computing applications. By expanding the range of molecular systems that can be modeled, this work provides a crucial foundation for validating quantum algorithms on future fault-tolerant machines. The ability to simulate larger quantum circuits means researchers can now test algorithms on more complex and realistic molecular systems. This progress directly supports the development of software that will eventually run on fault-tolerant quantum computers, accelerating the path toward practical industrial applications in chemistry and materials science. The collaboration between academic researchers and industry experts like Fixstars highlights the growing synergy required to bridge the gap between theoretical quantum potential and practical computational reality. This achievement sets a new benchmark for classical simulation capabilities, offering a vital tool for the next generation of quantum algorithm development.