SambaNova

SambaNova builds full-stack AI platforms with specialized chips for fast inference, training, and agentic AI in data centers and clouds.

SambaNova develops AI infrastructure for running large AI models. Its technology combines Reconfigurable Dataflow Units (RDUs), SambaRack systems, and SambaStack software to deliver high-speed AI inference in data centers. The platform is designed for large language models, AI agents, and other demanding workloads, with a focus on faster processing, efficient data movement, and lower power use.

Fractile

Fractile builds chips, systems, and software that run frontier AI models faster and at lower cost by removing inference bottlenecks.

Fractile develops specialized AI chips, systems, and software that speed up large language model inference and cut the cost of running generative AI in data centers. It fuses memory and computation on custom hardware to remove bandwidth bottlenecks, enabling higher throughput, lower latency, and improved energy efficiency for advanced AI workloads.