Products
Technology
Resources
Solutions
Community
About

We Are SambaNova Systems

We empower organizations to innovate through a reimagined computing infrastructure free from the constraints of today’s conventional solutions.

Born in the Era of AI, Built for the Future of Computing

Hear from our co-founder and CEO, Rodrigo Liang, as we debut the groundbreaking SambaNova Systems DataScale®. DataScale helps to advance AI innovations across a broad range of use cases, including natural language processing (NLP), high-resolution computer vision, recommender systems, AI for science, and more.

SambaNova Systems DataScale

DataScale is the core infrastructure to develop and deploy next-generation AI algorithms and applications at scale with unmatched performance, state-of-the-art accuracy, scale and ease of use.

DataScale Datasheet

SambaFlow Datasheet

Reconfigurable Dataflow Architecture Whitepaper

Dataflow-as-a-Service

Daas-image

Accelerate workloads deployed in your own data center with a subscription service designed to jump-start your AI initiatives.

SambaNova AI Cloud Platform

DevCloud-image

A new AI systems platform in the cloud for university researchers. Compelling research proposals are being accepted now.

World Record-Breaking Results in Performace, Accuracy, Scale, and Ease of Use

World Record-Breaking Results in Performace, Accuracy, Scale, and Ease of Use
benchmark_performance
PERFORMANCE
  • World record BERT-large training 1.4 times faster than NVIDIA DGX A100*
  • World record DLRM inference 20 times better throughput and latency than NVIDIA DGX A100*
benchmark_accuracy
ACCURACY
  • World record accuracy of 90.23% out of the box for high-resolution computer vision compared to NVIDIA DGX A100*
  • World record accuracy of 80.46% for DLRM recommendation engines compared to NVIDIA A100 GPUs*
benchmark_scale
SCALE
  • World record BERT- large training and state-of-the-art accuracy at multi-rack scale
  • Easily integrates into any existing infrastructure running customer workloads in 45 minutes to enable rapid deployments at scale*
benchmark_easy
EASE OF USE
  • Integrates with open source and open standards tools and interfaces for effortless developer productivity
  • Run pre-trained Hugging Face transformer models on DataScale with state-of-the-art accuracy and without code changes

What our customers say

Lawrence Livermore National Laboratory

“The support for running multiple models built into the hardware was a significant aspect of our selection of SambaNova Systems. … We’re seeing roughly a 5X improvement compared to a comparable GPU running the same models.”

Bronis de Supinski, Chief Technology Officer

Argonne National Laboratory

“We’re interested in technology … that can train models faster than individual GPUs, that has more scalability and has more performance, and so the SambaNova DataScale system met that criteria.”

Rick Stevens, Associate Laboratory Director