Skip to content
Back to publications
Nature Nanotechnology Journal Review

How nanophotonics can drive optical computing toward practical applications

Yitong Chen, Guoqiang Yang, Tao Yan, Chunyang Tang, Jiamin Wu, Qionghai Dai

A review of photonic computing platforms and the materials, fabrication, and integration requirements for practical AI and scientific computing.

Read paper (opens in a new tab)

My role

I drafted the “Large-scale AI models” subsection within the “Nanotechnology-enabled optical computing” section.

Coverage: People's Daily · 人民日报客户端 (opens in a new tab) DeepTech深科技 (opens in a new tab) All coverage

Nanotechnology-enabled optical computing in large-scale AI models, with six panels illustrating spatial encoding, large-scale metasurfaces, polarization and wavelength multiplexing, fast reconfiguration, and nonlinear operations.
Nanotechnology-enabled optical computing in large-scale AI models. Approaches to scaling, multiplexing, reconfiguration, and nonlinear operations. Chen et al., Nature Nanotechnology (2026).

Overview

Optical computing uses the physical properties of light to perform computation. This review examines photonic platforms with demonstrated programmability, integration density, and stability, and considers their suitability for large-scale AI models, edge computing, and logic and scientific computing. It connects these application demands to requirements for nanomaterials, architecture, fabrication, and system integration.

From devices to practical systems

The review considers how computing scale, precision, energy efficiency, and programmability interact at the system level. It brings together two perspectives: what existing materials and manufacturing processes make possible, and what particular applications require from a usable computing platform. Representative commercial systems provide further context for assessing progress toward deployment.

The article synthesizes existing work into an assessment of engineering requirements and directions for future development.

Media coverage

DOI: 10.1038/s41565-026-02264-4 (opens in a new tab)