[Teas] COAC 2026 | Call for Papers — Extended Submission Deadline: November 15, 2026

[Zhiyao Liu <[e-mail]>]

|
|

Call for papers

2026 International Conference on Optics for AI Clusters (COAC 2026)

Hefei, China | December 18-20, 2026

Held in conjunction with AI3T 2026

Theme: Weaving Light into the Fabric of AI Clusters

|
|

Introduction

The 2026 International Conference on Optics for AI Clusters (COAC 2026) will convene in Hefei, China, held in conjunction with the 2026 International Conference on AI Infrastructure and Interconnect Technologies (AI3T 2026). As AI clusters grow to tens of thousands of accelerators, the bandwidth, reach, power, and packaging limits of electrical interconnects make optical technologies essential to further scaling. COAC 2026 will bring together researchers and industry leaders across optical devices, photonic integration, networking, and AI systems to explore advances from materials and packaging to transceivers, switching fabrics, and cluster architectures. Join this global forum to share research and shape the optical infrastructure powering the next generation of AI!

|

Scope and Topics

|
|

We solicit original, high-quality research papers and practical experience reports on all aspects of optics for AI clusters. Topics of interest include, but are not limited to:

Optical Interconnect Architectures for AI Clusters

• Optical interconnect for scale-up domains and scale-out fabrics in AI clusters

• Co-packaged optics, near-packaged optics, and on-board optical engines

• Optical input/output and chip-to-chip photonic links for accelerators and memory

• Hybrid electrical-optical topologies, network design, and cluster-level case studies

Optical Switching and Network Fabrics

• Optical circuit switching, wavelength routing, and reconfigurable optical fabrics

• Optical packet and burst switching for bursty AI training traffic

• All-to-all, collective communication, and in-network computing over optical fabrics

• Control planes, scheduling, and fast reconfiguration for optical AI networks

Optical Transceivers, Links, and Physical Layer Technologies

• High-speed optical transceivers, intensity-modulation and coherent-lite architectures

• Linear-drive, low-power optics and the removal of retimers and DSP in short-reach links

• Modulation formats, forward error correction, and digital signal processing for AI links

• Link budget, signal integrity, and physical-layer measurement of optical AI interconnects

Optical Devices, Photonic Integration, and Packaging

• Silicon photonics, silicon nitride, thin-film lithium niobate, and heterogeneous integration

• Lasers, modulators, photodetectors, and wavelength-division multiplexing components

• Optical packaging, fiber coupling, co-design with electronic dies, and thermal packaging

• Optical I/O chiplets, photonic interposers, and advanced interconnect packaging

Photonic Computing and Optical Acceleration

• Photonic tensor cores, optical matrix multiplication, and analog optical computing

• Optical neural networks and photonic inference accelerators

• Non-von Neumann and neuromorphic photonic architectures for AI workloads

• Integration of photonic accelerators with electronic AI systems and software stacks

Energy Efficiency, Thermal Management, and Reliability

• Power and energy modeling of optical interconnects versus electrical alternatives

• Thermal management, wavelength drift control, and athermal photonic design

• Reliability, yield, field failures, and lifetime assessment of optical components

• Operational experience from large-scale optical AI cluster deployments

Modeling, Simulation, and Design Automation

• Device-to-system and electronic-photonic co-simulation methodologies

• Design automation, layout, and verification tools for photonic integrated circuits

• Digital twins and system-level simulation of optical AI clusters

• Modeling of collective communication performance over optical fabrics

AI-Driven Design and Operation of Optical Systems

• Machine learning for photonic device design, inverse design, and layout optimization

• AI-assisted optical network planning, fault localization, and performance tuning

• Data-driven modeling, calibration, and control of optical transceivers and fabrics

• Autonomous operation, predictive maintenance, and self-optimizing optical AI clusters



|
|

Important dates

|
|

Full Paper Submission Deadline:

November 15, 2026(Extended)
Author Notification:
November 22, 2026
Camera Ready Submission:
November 30, 2026
Registration Deadline:
November 30, 2026
|
|

Organizing committee

|
|

General Chairs:

Trung Q. Duong, Memorial University, Canada
Yihua Hu, National University of Defense Technology, China


Program Chairs:
Bin Chen, Hefei University of Technology, China
Xuwei Xue, Beijing University of Posts and Telecommunications, China
|
|

Paper submission

|
|

Ø  Submitted papers must be original, written in English, and not under consideration for publication elsewhere. All submissions will undergo double-blind peer review.

Ø  Full Papers: up to 8 pages, with up to 4 additional pages. Short Papers: up to 4 pages, with up to 4 additional pages. Posters: up to 2 pages, with up to 2 additional pages. Additional pages are subject to extra-page fees.

Ø  Manuscript format: IEEE Computer Society Proceedings format (two columns, single-spaced, 10-point font).

Ø  Accepted papers can be presented in oral, poster, or online format.

Ø  Registration fees must be paid before uploading the final, publication-ready version of the paper. The final PDF must pass the IEEE PDF eXpress compliance check before final submission.

|
|

Publication

|
|

All papers will be reviewed by three experts. After a careful reviewing process, all accepted papers with proper registration and presentation will be published in IEEE Xplore, and will be submitted to EI Compendex and Scopus for indexing.

|
|

Website

|
|

Ø  COAC 2026 Homepage: http://coac.ieee-aiconf.cn/

Ø  Submission Website: http://coac.hotcrp.ieee-aiconf.cn/

|
|

Contact

|
|

Ø  For any inquiries regarding the conference, please contact us at: [e-mail]

|