AI & Machine Learning

SambaNova Systems

4.31

SambaNova builds custom AI processors and a full-stack AI platform for enterprise generative AI deployment.

Visit Website

SambaNova Systems was founded in 2017 by Kunle Olukotun, Christopher Re, and Rodrigo Liang in Palo Alto, California. Olukotun and Re are Stanford professors with deep expertise in computer architecture and machine learning systems.

SambaNova raised over $1.1 billion in total funding, including a massive $676 million Series D in 2021 at a $5 billion valuation. Investors include SoftBank Vision Fund, BlackRock, Intel Capital, and GV.

The company’s hardware platform is built on the SN40L Reconfigurable Dataflow Unit (RDU), a custom processor designed for AI workloads. Unlike GPUs, which are adapted from graphics processing, SambaNova’s chips are built from the ground up for the dataflow patterns of neural networks.

SambaNova offers its technology as SambaNova Cloud and on-premise Dataflow-as-a-Service solutions, where customers get a complete AI system — hardware, software, and pre-trained models — delivered as a managed service. This approach appeals to enterprises and government agencies that want generative AI capabilities without building ML expertise in-house.

The company serves customers in healthcare, financial services, government, and energy. SambaNova has also launched fast inference services for open-source models, competing on speed with Groq and Cerebras. The company employs around 500 people across offices in Palo Alto and international locations.