Implementation of Photonic AI accelerator system

Al applications require massive computing resources that consume immense energy. Al compute is expected to account for a significant percentage of the global power consumption. As current digital Al chips cannot keep up with demand, more energy-efficient Al chips are required.
This project addresses major milestones toward implementing an energy-efficient photonic Al accelerator system. A successful implementation is expected to improve energy efficiency of Al computations and have significant sustainability benefits.
This project aligns with the interests of the industrial partner 3E8, a Montreal start-up. 3E8 plans to commercialize the PAAS technology for both datacenter and edge applications. 3E8 expects significant gains from a successful technology implementation. The project benefits Canada by strengthening its leadership in the semiconductors and Al domains. The project will lead to ~100 well-paying jobs during R&D and 200+ jobs during production. Additionally, there are massive environmental benefits from
deploying the technology at-scale.

Faculty Supervisor:

Odile Liboiron-Ladouceur;Glenn Cowan;Yves-Alain Peter

Student:

Partner:

3E8 Inc

Discipline:

Engineering

Sector:

Manufacturing

University:

Concordia University; McGill University; Polytechnique Montréal

Program: