Press Releases

SambaNova Accelerates AI for Banking and the Public Sector with a World Record Performance System, New Partnerships, and International Expansion in 2022

Written by SambaNova Systems | December 22, 2022

Foundation Models Became the Cornerstone for 2022 Company Milestones and Enable Generative AI, the Transformative AI Trend of the Year

Palo Alto, December 21, 2022  SambaNova Systems, the company delivering the industry’s only comprehensive enterprise software, hardware, and solutions platform to run Artificial Intelligence (AI), deep learning, and foundation models, today announced several milestones in a breakthrough year. This announcement is significant because SambaNova is largely credited with introducing foundational models and large language models (LLMS) for enterprises wanting to leverage advanced AI capabilities.

“2022 was an amazing year for SambaNova, punctuated by continued growth and innovation. It marked a paradigm shift for AI as foundation models and generative AI took center stage,” said Rodrigo Liang, CEO of SambaNova. “We’ll look back on 2022 as the year when foundation models fundamentally changed the trajectory of AI.”

2022 HIGHLIGHTS AND MILESTONES

First to market with pre-trained foundation models for the enterprise: With deep roots at Stanford, where foundation models were conceived, SambaNova was first to market in 2022 with pre-trained foundation models for Banking. The company launched Dataflow-as-a-Service GPT Banking, offering a large language foundation model as a subscription service to simplify the process of deployment. Banks are leveraging foundation models with SambaNova solutions to improve efficiency while accelerating their digital transformation. SambaNova was first-to-market with foundation models in October 2021.

SambaNova delivers its next generation DataScale® system: SambaNova launched its next generation DataScale® system, an integrated hardware-software system that enables organizations to move away from tactical AI deployments to scalable, enterprise-wide solutions. DataScale enables organizations to achieve ROI faster than with any other AI infrastructure. Key enhancements include:

  • world record GPT training performance that is 6x faster to train than DGX A100 systems (GPT 13B)
  • support for the largest models with 8x more memory capacity than DGX A100 systems

SambaNova accelerates AI for Science: This year, SambaNova accelerated AI for Science when it delivered DataScale® to the U.S. Department of Energy’s Argonne National Laboratory. DataScale is available to the scientific community at the Argonne Computing Leadership Facility (ALCF) AI Testbed, a collection of the world’s most advanced AI accelerators. The partnership showcases Argonne’s adoption of a multi-rack SambaNova system and joint efforts on implementing challenging foundation models and deep learning workloads. For scientific research organizations, this means new experiments and discoveries with the potential to change the world.

Expansion to the Japanese Market: In March 2022, SambaNova announced further international expansion with the appointment of Toshinori Kujiraoka as country sales director for Japan to meet the country’s technology priorities. Toshinori Kujiraoka oversees business and sales development, operations and strategy within the region.

Notable Recognition:

  • SambaNova was recognized as the “Best Big Data Deep Learning/AI Solution” by the 2022 Tech Ascension Awards.
  • SambaNova accepted the HPCwire Readers Choice Award for Best Use of High Performance Data Analytics & Artificial Intelligence from HPC Wire. The award highlights how Argonne National Laboratory is using DataScale to accelerate AI for Science workloads, including large-scale imaging data and large language models.
  • SambaNova was awarded the “2022 Best Tech Startups in Palo Alto” award by The Tech Tribune, a technology news website focused on the world’s best technology startups.
  • Presidential Council of Advisors on Science and Technology (PCAST) Discussion: SambaNova CEO Rodrigo Liang participated in the President’s Council of Advisors on Science and Technology (PCAST). He was invited to discuss the tech ecosystem and AI development, representing the role of startups driving innovation. Rodrigo was joined by Pat Gelsinger, CEO at Intel Corporation; Priyanka Raina, Professor of Electrical Engineering at Stanford University; Aart de Geus, co-founder of Synopsys Inc; John Neuffer, President and CEO of the Semiconductor Industry Association. The President’s Council of Advisors on Science and Technology (PCAST), including PCAST Co-Chairs Frances Arnold, Francis Collins, and Maria Zuber, plays an important role in advising public policy decisions.

 

[1] About the test and Comparison Benchmarks:

Performance and accuracy comparisons are made on the following competitive equivalents:

  • SambaNova Dataflow-as-a-Service GPT running on 1 integrated system with 8 RDU, SambaNova GPT 1.5B parameters, global BS=1,024
  • Commodity server, 1 Nvidia DGX A100 with 8 A100 GPU, GPT 1.5B parameters, Global BS=1,024

Pre-training dataset: SEC EDGAR

Downstream task fine tuning dataset: FiQA+PhraseBank

About SambaNova Systems

Customers turn to SambaNova to quickly deploy state-of-the-art generative AI capabilities within the enterprise. Our purpose-built enterprise-scale AI platform is the technology backbone for the next generation of AI computing.

Headquartered in Palo Alto, California, SambaNova Systems was founded in 2017 by industry luminaries, and hardware and software design experts from Sun/Oracle and Stanford University. Investors include SoftBank Vision Fund 2, funds and accounts managed by BlackRock, Intel Capital, GV, Walden International, Temasek, GIC, Redline Capital, Atlantic Bridge Ventures, Celesta, and several others. Visit us at sambanova.ai or contact us at info@sambanova.ai. Follow SambaNova Systems on Linkedin.