CAREER: Cross-Layer Electronic-Photonic Design Automation and Co-Design Toolflow for Large-Scale Photonics-Empowered AI Systems

Artificial intelligence (AI) is accelerating demand for faster computing and data communication, but meeting this demand with electronics alone is becoming increasingly expensive and power-hungry. As AI workloads grow, achieving higher performance often requires disproportionately more energy and hardware cost. Integrated photonics uses light on a chip to move and process information at light speed and low energy, creating a promising path to more efficient hybrid electronic-photonic systems. However, designing and manufacturing photonic chips remains slow and costly, in part because the supporting design software, models, and workflows are not yet mature or widely accessible. This project will deliver an open-source, end-to-end workflow that enables scalable, rapid, and high-quality design and simulation of electronic-photonic chips, while improving manufacturability and reducing design iterations. The resulting productivity gains and lower barriers to entry will broaden access to this technology and accelerate its adoption in computing, communication, and sensing systems. The project will also expand education and workforce development by integrating hands-on modules into undergraduate and graduate courses, offering online materials and courses, and growing a public seminar series that connects students with researchers and industry. The technical goal of this project is to establish a full-stack electronic-photonic design automation framework for large-scale electronic-photonic integrated chips for AI computing, optical interconnects, and sensing. The research will develop (1) fast, physics-guided simulation and co-simulation methods, including a learned electromagnetic solver for photonic device modeling and a circuit-level co-simulator for mixed electronic-photonic systems; (2) fabrication-robust inverse design methods that model process variations to improve yield and reduce costly redesign cycles; and (3) scalable physical design automation that synthesizes large-scale photonic circuit layout under manufacturing rules and performance constraints. These capabilities will be integrated into a system-level modeling and design-space exploration engine that connects device behavior and layout effects to end-to-end system metrics such as speed, energy, area, and robustness. The expected outcome is an open, reproducible ecosystem with benchmarks and reference flows that accelerate design iteration and enable closed-loop co-design of next-generation hybrid electronic-photonic AI systems. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria. NSF Award ID: 2544675 | Program: 01002627DB NSF RESEARCH & RELATED ACTIVIT,01003031DB NSF RESEARCH & RELATED ACTIVIT,01002930DB NSF RESEARCH & RELATED ACTIVIT | Principal Investigator: Jiaqi Gu | Institution: Arizona State University, SCOTTSDALE, AZ | Award Amount: $360,002 View on NSF Award Search: https://www.nsf.gov/awardsearch/show-award/?AWD_ID=2544675 View on Research.gov: https://www.research.gov/awardapi-service/v1/awards/2544675.html