
SambaNova Systems: A Palo Alto-based software firm offering dataflow-as-a-service for diverse sectors.
SambaNova designs AI inference and training accelerators built on its Reconfigurable Dataflow Unit (RDU) architecture, including its SN40L RDU and the newer SN50 chip. The company packages its silicon into DataScale rack-scale systems sold to data-center customers for large-language-model workloads.
The company's current focus is AI inference and on-premise deployments, selling complete server and rack systems rather than discrete chips. Customers such as JPMorgan use SambaNova as an on-premise inference provider, and the company positions its systems as an alternative to GPU-based stacks for enterprise AI.
SambaNova operates in the rapidly expanding AI semiconductor market, where demand for inference accelerators is growing as enterprises deploy large language models in production. The company is part of a wave of startups seeking to challenge Nvidia's dominance in AI chips.
Following its $1 billion Series F at an $11 billion valuation in July 2026, SambaNova said it is considering a U.S. IPO in 2027. Co-founder and CEO Rodrigo Liang has emphasized the company's focus on remaining a standalone, independent business as it scales its inference and on-premise offerings.
SambaNova's primary competitive advantage is its integrated full-stack approach, combining custom RDU silicon, DataScale systems, and software into a single offering tuned for large-model inference. This vertical integration lets the company optimize performance and total cost of ownership for enterprise AI workloads.
The company also differentiates by emphasizing on-premise, inference-first deployments that appeal to regulated enterprises wanting data control. Its ability to attract investors such as General Atlantic, T. Rowe Price, and Intel at an $11 billion valuation signals strong confidence in its position as a Nvidia alternative.