Nvidia Alumnus Explains Shift From Copper to Optical Interconnects
In an interview shared by Molly O'Shea, former Nvidia engineer Yannick De Koninck says AI models outgrow single GPUs and copper is running out of bandwidth, pushing data centers toward optical interconnects. He now works at Thema, which is building photonics manufacturing in Europe.
Original post · 1 min read
Yannick De Koninck, who helped develop silicon photonics at Nvidia, explains why AI models & agents are pushing data centers from copper to light:
“The models have gotten so large that they no longer fit on a single GPU. So what we need to do is interconnect multiple GPUs together to run these models.”
“As these GPUs get faster, they need more data, they need more data at a faster rate, and copper is running out of steam.”
“That's why we're transitioning to optical interconnects, which bring much higher data transfer bandwidths.”
Molly O’Shea @MollySOSheaNEW: Why This NVIDIA Engineer Left to Build @ThemaFoundry
Thema is building photonics manufacturing capacity in Europe
“If you look at an AI factory today, 50% of that is GPUs, but the other 50% is the technology to interconnect these GPUs.”
I sat down with Thema's CEO Herwig Van Hove & CTO Yannick De Koninck in Monaco to go deep on the photonics supply chain powering the next generation of AI infrastructure.
Yannick spent 5 years at NVIDIA, where he helped build its silicon photonics technology from scratch to product-ready maturity. He left what he calls the “golden palace” to join Thema…



