The 1st International Conference on AI Infrastructure and Interconnect Technologies

AI Infrastructure & Architecture

• Hyperscale AI training and inference system design
• Heterogeneous computing system integration and management
• Memory-centric, near-memory, and in-memory computing for AI
• Reconfigurable and adaptive AI hardware architectures

High-Performance Interconnects & Optical Networking

• Novel electrical and optical interconnect technologies (e.g., CPO, LPO, OIO, co-packaged optics)
• Intra- and inter-data center optical networking
• Ultra-high bandwidth, ultra-low latency network switching and protocols
• Standards and evolution of interconnect technologies

AI Accelerator & Heterogeneous Computing Networks

• Interconnect topologies and algorithms for GPU/TPU/FPGA/ASIC clusters
• Optimization and innovation of accelerator communication libraries (e.g., NCCL, RCCL)
• On-chip networks and packaging for large-scale AI chips

Data Center & Scalable Network Architectures

• Data center network topologies and traffic engineering for AI workloads
• Lossless networking and congestion control
• Compute-Network Convergence and In-Network Computing

Distributed AI Training & Workload Orchestration

• Network communication optimization for large-scale distributed training
• Co-scheduling and orchestration of training and inference workloads
• AI resource pooling and management across clusters and domains

Network Automation & AI-Native Operations

• AI for network prediction, operation, optimization, and self-healing
• Application of digital twins in network management and planning
• Intent-based networking and autonomous network operations

Edge-to-Cloud Computing for AI

• Network architectures for distributed AI inference and training
• Latency and bandwidth management in edge-cloud pipelines
• Software and infrastructure for federated learning at scale

Embodied AI & Robotic Infrastructure

• Compute and communication architectures for robot fleets and autonomous systems
• Low-latency edge processing for perception and control
• Networking for collaborative embodied intelligence

Secure, Trustworthy & Sustainable AI Systems

• Hardware and network security for AI systems (confidential computing, side-channel mitigation)
• Architectures for verifiable and reliable AI execution
• Energy-efficient and green AI infrastructure (cooling, power delivery, PUE optimization)

Emerging Interconnect Technologies & Standards

• Novel materials, devices, and paradigms for future interconnects
• Co-design of interconnects with AI workloads and programming models
• Industry standards and ecosystem development