Vishal Chandrasekar, Director of Project Management at Ayar Labs, demonstrates the company’s comprehensive optical scale-up technology solution for next-generation GPU connectivity. In this detailed walkthrough, Chandrasekar reveals how Ayar Labs is preparing for the industry’s shift from copper to optical interconnects in high-performance computing environments, showcasing everything from silicon photonics wafers to fully integrated rack-scale systems.
**What You Will Learn:**
– How silicon photonics technology enables optical connectivity for GPUs and switches at scale
– The complete supply chain approach from TSMC-manufactured optical engine wafers to packaged solutions
– Integration strategies for multi-chip packages combining custom hyperscaler silicon with optical engines
– Rack-scale system architecture featuring liquid-cooled components and fiber optic connectivity
– The role of external pluggable laser technology in achieving high-bandwidth optical interconnects
– How Ayar Labs’ solutions align with emerging industry standards like the OCI MSA for scale-up networks
Chandrasekar provides insights into critical partnerships across the ecosystem, from fiber connector suppliers to system integrators, and explains the technical requirements for deploying optical interconnects in data centers targeting thousands of GPUs.
📚 CHAPTERS:
0:00:00 – Introduction to Optical Scale-Up Technology for GPU Connectivity
0:00:22 – Silicon Photonics: From Optical Engine Wafers to Single Die Packages
0:00:49 – Fiber Optic Connectors and Supply Chain Partnerships
0:01:07 – Multi-Chip Packaging Integration with Hyperscaler Custom Silicon
0:02:02 – Rack-Scale System Architecture and VWin Partnership
0:02:54 – External Pluggable Laser Technology and Liquid Cooling Solutions
0:03:45 – High-Power Laser Demonstration and OCI MSA Compliance
Want to be involved our video series? Contact [email protected]
https://ngi.fyi/dcnet26-ayarlabs-vc
Check out full showcase at: https://ngi.fyi/26DCNetworkAIyt to learn more about data center networking for AI and cloud workloads
