Kopin Analysis Finds Optical Interconnect Architecture Could Break the Data Center “Copper Wall”
Business Wire•19/10/2026•08:31 ET
0
Key Highlights
- ➤Kopin (KOPN) models 1.3–1.6x greater useful training throughput per GPU
- ➤10,000-GPU systems could cut equivalent training time 23%–38%
- ➤Neural I/o extends high-bandwidth connectivity from one meter to tens of meters
- ➤Initial transmission-link testing underway; demonstration targeted January 2027
- ➤First-generation Neural I/o production silicon targeted for 2028
Expert Statements
Michael Murray, Chief Executive Officer of Kopin Corporation
“AI infrastructure’s ability to connect compute efficiently is becoming as important as the compute itself”
Michael Murray, Chief Executive Officer of Kopin Corporation
“The industry has invested heavily in increasing the performance and scale of AI processors, but those processors ultimately have to operate as a system. When the physical limitations of today’s interconnect no longer dictate that system’s architecture, our analysis shows the potential for better utilization of existing compute and much greater freedom in how future AI infrastructure is designed.”
Kopin Analysis Finds Optical Interconnect Architecture Could Break the Data Center “Copper Wall”
New white paper shows how moving beyond copper to massively parallel MicroLED-based optical channels could improve GPU utilization and reshape future AI data centers. Join Kopin’s live “Ask Me Anything” discussion on October 20, 2026, for insights into the findings and the future of AI.
Get started
Create a free account to read the full story.