News
DealsCorvex agrees USD 103 million take-or-pay deal to perform AI inference with Mid-Atlantic GPUsSovereignNeysa scales AI infrastructure in India with AMD Instinct MI350X GPUs on TytanInferenceMicrosoft Foundry adds Baseten, DeepSeek AI inference (MSFT:NASDAQ)ClustersDataOne and Rowan University sign engineering co-op program agreement for DataOne's Vineland AI factoryClustersLook inside Cal Poly's $3 million AI Factory (Video)DealsOpenAI Taps AMD Chips for 457 MW of AI Capacity in Six-Year TensorWave DealClustersCal Poly launches an AI data center on campus. Take a look insideInferenceArista takes aim at the next AI network battleground: distributed inferenceDealsOpenAI Signs 457 MW AI Data Center Deal as AMD Expands Its Challenge to NvidiaClustersBehind Dell AI Factory’s NVIDIA & AMD Enterprise ExpansionDealsCorvex agrees USD 103 million take-or-pay deal to perform AI inference with Mid-Atlantic GPUsSovereignNeysa scales AI infrastructure in India with AMD Instinct MI350X GPUs on TytanInferenceMicrosoft Foundry adds Baseten, DeepSeek AI inference (MSFT:NASDAQ)ClustersDataOne and Rowan University sign engineering co-op program agreement for DataOne's Vineland AI factoryClustersLook inside Cal Poly's $3 million AI Factory (Video)DealsOpenAI Taps AMD Chips for 457 MW of AI Capacity in Six-Year TensorWave DealClustersCal Poly launches an AI data center on campus. Take a look insideInferenceArista takes aim at the next AI network battleground: distributed inferenceDealsOpenAI Signs 457 MW AI Data Center Deal as AMD Expands Its Challenge to NvidiaClustersBehind Dell AI Factory’s NVIDIA & AMD Enterprise Expansion
EveryNeocloud
SambaNova Systems

SambaNova Systems

American provider of custom AI silicon and cloud inference services based on Reconfigurable Dataflow Units.

Founded in late 2017, SambaNova Systems designs proprietary hardware accelerators called RDUs for deep learning. The company offers both physical chips and a cloud platform named SambaCloud to handle large language model inference.

Share on WhatsApp

  • Founded in November 2017 by three researchers.
  • Develops custom Reconfigurable Dataflow Units instead of GPUs.
  • Valued at $11 billion as of July 2026.
  • Partnered with JPMorgan Chase for inference services.
  • SN50 chip reached 763 tokens per second in testing.

What it is

SambaNova Systems is an American technology firm that creates specialized hardware and software for artificial intelligence. It was established in November 2017 by Kunle Olukotun, Christopher Ré, and Rodrigo Liang. The company focuses on building systems using its own Reconfigurable Dataflow Units, or RDUs, which are processors designed specifically for machine learning tasks. Beyond selling physical chips, SambaNova operates SambaCloud, a service that allows users to run large language models without managing the underlying infrastructure. This approach combines custom semiconductor design with accessible cloud computing capabilities.

How it came about

The company originated from research into microprocessing arrays funded by the Defense Advanced Research Projects Agency. After its founding, it secured significant capital, including $150 million from Intel and Alphabet in 2019. By 2021, a funding round led by SoftBank pushed its valuation to $5 billion, bringing total funds raised to $1 billion. The business expanded into cloud services during the early 2020s. In 2024, it announced the Samba-1 bundle of generative models. Later that year, Time magazine recognized its suite as a top invention. A major funding event occurred in February 2026 when Vista Equity Partners and Cambium Capital invested $350 million, followed by a $1 billion Series F round in July 2026 led by General Atlantic, raising the valuation to $11 billion.

The compute

SambaNova’s core technology relies on a reconfigurable dataflow architecture rather than traditional GPU structures. Its primary hardware component is the Reconfigurable Dataflow Unit, optimized for moving data efficiently during AI computations. The company released the SN40L chip in 2023 to power its full-stack platform. In February 2026, it unveiled the SN50 RDU, which began deployment later that year. Independent tests showed the SN50 processing 763 tokens per second on specific models. The company provides both on-premises hardware solutions and cloud-based inference platforms to support model deployment.

Where the GPUs are

Headquartered in San Jose, California, the company also has ties to Palo Alto. Its technology has been integrated into major research facilities globally. The United States Department of Energy placed SambaNova systems within the AI Testbed at the Argonne Leadership Computing Facility. In Japan, the Riken Center for Computational Science deployed these systems alongside the Fugaku supercomputer. Additionally, third-party providers like OVHcloud and Vector Core Compute utilize SambaNova technology to power their own AI endpoints.

Customers and deals

SambaNova serves a mix of financial institutions, telecommunications firms, and research organizations. In July 2026, it partnered with JPMorgan Chase to supply inference infrastructure using its SN40L and SN50 chips. SoftBank Corp became an early adopter, deploying the SN50 chip in its Japanese data centers. Other clients include Vector Core Compute and OVHcloud, which integrate the hardware into their respective platforms. Public research labs also use the systems to test alternative acceleration models for scientific calculations.

Money and ownership

The company remains privately held. Initial backing came from DARPA-funded research. Major venture investments included $150 million from Intel and Alphabet in 2019, followed by $250 million in 2020. A $676 million round led by SoftBank Vision Fund 2 in April 2021 brought in Temasek and GIC, reaching a $1 billion total raise. Valuation hit $5 billion that year. Recent capital injections include $350 million in February 2026 and $1 billion in July 2026. These funds supported the development of new chips and the expansion of the SambaCloud platform, pushing the valuation to $11 billion.

Its place in AI

SambaNova occupies a niche in the AI industry by offering an alternative to standard GPU architectures through its custom silicon. Its inclusion in the Forbes AI 50 list in 2026 highlights its standing among prominent AI entities. The company plays a role in national and international research efforts, supporting facilities like Argonne National Laboratory. Its focus on efficient dataflow processing aims to improve the speed and cost-effectiveness of running large language models compared to conventional hardware setups.

Questions and criticism

The provided material does not detail specific concerns regarding debt, energy usage, or export controls for SambaNova Systems.

Questions people ask about SambaNova Systems

What makes SambaNova different from other AI chip makers?

SambaNova uses a unique reconfigurable dataflow architecture with its own Reconfigurable Dataflow Units. This design focuses on optimizing data movement for machine learning, differing from standard graphics processing units used by competitors.

Does SambaNova sell only hardware or do they offer cloud services?

The company provides both physical hardware accelerators and a cloud platform called SambaCloud. Users can deploy large language models either on-premises using SambaNova chips or via the cloud service.

Who are some of the major investors in SambaNova?

Major investors include Intel, Alphabet, SoftBank Vision Fund 2, Temasek, GIC, Vista Equity Partners, Cambium Capital, and General Atlantic. The company raised over $1 billion in total funding by mid-2026.

Where is SambaNova based?

SambaNova Systems is headquartered in San Jose, California, in the United States. The founders also have connections to Palo Alto, where initial research activities took place.

Good reading and sources

Where this page comes from

The facts on this page come from the Wikipedia article "SambaNova", written by its contributors and used under the CC BY-SA 4.0 licence, and from its Wikidata record (CC0). EveryNeocloud rewrote the text in its own words, with AI assistance, and compares every page with the article to make sure no passage is copied. It may be reused under the same CC BY-SA 4.0 licence with a link back.

It may contain mistakes, and figures change over time. Nothing here is investment or buying advice. If something is wrong, tell us. Last refreshed 2026-11-15.

Photo credits on this page

The small pictures above come from Wikimedia Commons under free licences. Some are cropped to fit.

The weekly digest

Every Monday: the provider of the day, new pages on EveryNeocloud, and the conferences and anniversaries of the week ahead.