Description: The rapid growth of AI infrastructures is reshaping the requirements for optical communication and networking. Emerging AI data centers, high-performance computing platforms, and distributed cloud-edge environments are driving unprecedented demand for bandwidth, scalability, energy efficiency, and operational intelligence. These evolving demands are pushing optical systems toward wider spectral utilization, higher capacity, and broader deployment scope, while also driving optical networks toward more programmable, adaptive, and application-aware architectures. At the same time, AI itself is becoming an important tool for optical systems and networks, enabling tasks such as transmission optimization, telemetry-driven control, and autonomous operation.
As a result, the design space of optical systems and networks is rapidly expanding, encompassing a wide range of transmission technologies, device innovations, and architectural approaches, including ultra-wideband systems as one promising direction. However, as optical technologies diversify, fair and meaningful evaluation becomes increasingly difficult. Performance comparisons are often based on inconsistent assumptions regarding traffic models, workload scenarios, performance metrics, energy efficiency, reliability, reconfiguration behavior, and baseline architectures.
This raises a timely need for common benchmarking methodologies for optical systems and networks designed for and enhanced with AI.
Themes:
Theme: Ultra-wideband Systems and AI-driven Optical Infrastructure
This theme explores how AI-driven infrastructure demands are pushing optical systems toward wider spectral utilization, higher capacity, and broader deployment scope. It also considers technological directions such as ultra-wideband transmission, device and system innovations, and the emerging role of AI in transmission and system optimization.
Invited Speakers:
Theme: Benchmarking Methodologies for AI-driven Optical Networks
This theme addresses how emerging optical systems and architectures should be benchmarked. The focus is on AI-oriented workloads, evaluation metrics, reproducibility, and fair comparison across diverse technological approaches.
Invited Speakers:
Format:
The workshop will consist of two 90-minute sessions, each featuring invited presentations followed by a panel discussion involving speakers and participants.
The detailed schedule, session order, and final list of invited speakers, including any additional invited speaker(s), will be announced in the final program.
The workshop aims to foster discussion on future directions for optical infrastructures supporting AI, as well as benchmarking methodologies for next-generation optical networks.
Abstract:
Next-generation AI computing will be shaped by advances in processors, accelerators, and communication infrastructures that interconnect them. As AI systems scale, performance, scalability, and energy efficiency of computing and switching fabrics become key constraints, exposing the limits of conventional electrical technologies in bandwidth, latency, and power. Emerging photonic solutions, such as optical circuit switching, coherent and co-packaged optics, and linear photonic computing, offer promising alternatives by reducing data movement costs and enabling high-capacity, scalable architectures. This workshop examines these developments, focusing on trade-offs between electrical and optical approaches, energy efficiency (pJ/bit), novel network topologies, and geo-distributed strategies for sustainable AI infrastructure.
Abstract:
Can integrated photonics truly enable quantum‑enhanced data centers?
As quantum and classical systems move toward tighter integration, data‑center‑scale environments may represent one of the first realistic settings for deploying quantum communication technologies, avoiding many of the challenges of long‑distance networks.
This workshop explores whether integrated photonics can effectively bridge these domains at scale, addressing key challenges in switching, control, coexistence, and system integration. It covers technologies such as QKD and hybrid photonic platforms, and examines their role in emerging data‑center‑scale architectures.
Bringing together leading experts across devices, networks, and systems, the workshop will assess both the potential and the limitations, and identify realistic pathways beyond today’s experimental demonstrations.
Speakers:
| Yasutaka Morioka | NTT DOCOMO, Japan | A Proposal for Hybrid Use of PQC and QKD |
| Antonio Melgar | Telefonica, Spain | Operator Perspective on Integrating Photonic and Quantum Security Technologies for Next‑Generation Data‑Center Networks with Built‑In Security |
| Elisa Redolfi | FiberCop, Italy | Protecting Critical Infrastructure Data with Quantum-Safe Communication Networks |
| Rikizo Ikuta | The University of Osaka, Japan | Frequency-conversion quantum switches for heterogeneous quantum networks |
| Elham Heidari | University of Florida, USA | Kerr Optical Frequency Combs in both classical and quantum PIC applications |
| António Teixeira | PICadvanced, Portugal | Packaging of integrated photonics in classical and transition to quantum communications |
| Marco Avesani | University of Padova, Italy | Photonic integration of QKD: roadmap and challenges |
Format:
The workshop will consist of two 90-minute sessions, each featuring invited presentations followed by a final panel discussion involving speakers and participants.
The detailed schedule, session order, and final list of invited speakers, including any additional invited speaker(s), will be announced in the final program.
