Lab for Unconventional Computing using Emerging Nano-Technologies
Empowering the Future of Computing
We design sustainable, energy-efficient computing systems using emerging nano-technologies — silicon photonics, memristive devices, and AI-driven hardware co-design. Based at the University of Texas at San Antonio.
Department of Computer Engineering · College of AI, Cyber and Computing
Why this matters
AI compute is doubling every five months. Data center electricity will double by 2030. Training a frontier model now burns the electricity of a small city and the water of an Olympic pool — and we're still early. To keep AI sustainable, we need to rethink the hardware substrate itself. At LUCENT, we explore what comes after the transistor: light, memristors, and computer architectures designed from first principles for the workloads of the next decade — not the last one.
Designing the chip and the system that runs on light — from micro-ring devices to instruction sets.
Learn more → Area 02Wiring up thousands of accelerators with light — topologies for distributed AI training and inference.
Learn more → Area 03Letting algorithms design our devices — accelerated by GPUs and machine-learned simulation surrogates.
Learn more → Area 04Bringing photonics and memristors to reconfigurable hardware — deployable accelerators, today.
Learn more →A scalable photonic MVM engine with phase-error compensation for analog AI accelerators.
Carbon-aware exascale AI training over analog photonic fabrics.
A fully-analog photonic softmax block enabling end-to-end optical inference.
Architecture / interconnect co-design for edge transformer inference on 2.5D photonic systems.
A photonic tensor-core architecture targeting real-time AI workloads.
A photonic CNN accelerator drawing on bio-inspired analog computation.
[date]
[date]
[date]
[date]
[date]
[date]
[date]
[date]
