Jane Street

AI safety company and maker of Claude AI models

Specialized cloud infrastructure provider purpose-built for GPU-accelerated workloads. Operates large-scale data centers optimized for AI model training and inference, serving as a key compute partner for frontier AI labs.

MatX designs custom silicon optimized for large language model training and inference, aiming to deliver 10x more computing power per square millimeter than Nvidia GPUs. Founded by former Google TPU engineers, the company is building the MatX One processor featuring a splittable systolic array architecture with SRAM-first design. Manufacturing is planned through TSMC with initial shipments targeted for 2027.