HPCwire examines how Lumai is redefining AI compute by replacing electrons with photons. By performing matrix operations optically, Lumai’s architecture delivers significant gains in speed and efficiency—reducing energy usage by up to 90% while enabling larger, more efficient AI workloads. Read More...

EE Times examines how Lumai is bringing optical computing closer to production with its lens-based accelerator for AI inference. By computing in three-dimensional space using light, Lumai’s approach achieves scalable, high-throughput matrix operations while overcoming limitations that have constrained earlier photonic solutions. Read more...

In this Unite.AI feature, Lumai’s Iris Nova optical server is introduced as a breakthrough in AI infrastructure—using light-based processing to run large language models in real time. The article highlights how Lumai’s architecture delivers massive parallelism and up to 90% lower energy consumption, addressing the growing power constraints of AI inference. Read more...