"CamGraPhIC's technology addresses a problem that has become increasingly acute as AI systems scale: compute power is no longer the primary bottleneck. Instead, performance is being constrained by the speed and efficiency with which data moves between chips, accelerators, and memory."
"Existing electrical interconnects and conventional silicon photonic links struggle to keep pace with the bandwidth demands of large AI models, consuming disproportionate power and generating heat."
"CamGraPhIC's approach replaces these links with graphene-based optical input/output technology, which the company says delivers substantially higher bandwidth density, lower latency, and significantly lower energy consumption than the best available silicon photonics."
CamGraPhIC, a subsidiary of 2D Photonics, has secured €211 million in funding from the European Commission to develop its graphene photonic interconnect technology. This funding is one of the largest public investments in an Italian deep-tech startup. The investment will facilitate the construction of a pilot manufacturing facility near Milan, set to operate by 2028. The technology aims to overcome limitations in data transfer speed and efficiency in AI systems, offering higher bandwidth density and lower energy consumption compared to traditional silicon photonics.
Read at TNW | Deep-Tech
Unable to calculate read time
Collection
[
|
...
]