“AI is changing the world and at SambaNova, we believe that you don’t need unlimited resources to take advantage of the most advanced, valuable AI capabilities – capabilities that are helping organizations explore the universe, find cures for cancer, and giving companies access to insights that provide a competitive edge.”

‍

SambaNova, co-founded by Kunle Olukotun, Rodrigo Liang and Chris Re in 2017 in Palo Alto, California, delivers advanced AI infrastructure by combining custom-designed hardware, integrated software, and pre-trained models. This comprehensive approach enables enterprises and government agencies to deploy large-scale AI workloads efficiently, whether on-premise or through dedicated cloud solutions. The company’s mission is to simplify AI adoption while ensuring robust performance and data security across industries

‍

Business Model

SambaNova Systems is a Silicon Valley-based AI computing company offering a full-stack AI platform that combines custom hardware, software, and pre-trained models to support advanced AI (such as large language models) in enterprise environments. SambaNova’s mission is to deliver enterprise-grade AI infrastructure that organizations can deploy on-premises or via cloud services for high-performance AI workloads​.

‍

‍

<row>

<col>

SambaNova DataScale

DataScale is a fully integrated system that combines SambaNova’s custom Reconfigurable Dataflow Unit (RDU) chips, large memory architecture, and software stack. By using a dataflow architecture rather than traditional GPU-style processing, DataScale can execute AI workloads with higher parallelism and efficiency.

</col>

<col>

‍

</col>

</row>

<row>

<col>

‍

</col>

<col>

SambaNova Suite

The SambaNova Suite refers to the comprehensive platform that includes the DataScale hardware plus integrated software and models for generative AI and can be deployed on-premises (with physical DataScale units) or in the cloud, giving customers flexibility​.

</col>

</row>

‍

SambaNova's Growth Timeline

‍

‍

SambaNova and Competition

‍

SambaNova Systems is a leading innovator in AI infrastructure, competing with industry giants like NVIDIA, Cerebras, Graphcore, and Groq. Unlike NVIDIA, which dominates the AI market with its powerful GPUs and extensive software ecosystem, SambaNova differentiates itself with its full-stack AI platform, offering integrated hardware, software, and pre-trained models optimized for enterprise deployment. Cerebras, known for its massive wafer-scale chip, excels in training ultra-large AI models, while Graphcore’s Intelligence Processing Unit (IPU) focuses on parallel AI processing for high-speed computation. Groq, on the other hand, specializes in ultra-low-latency AI inference, targeting real-time AI applications. SambaNova's Dataflow architecture and memory-rich design allow it to efficiently handle large-scale AI workloads, offering an alternative to traditional GPU-based solutions. By combining cutting-edge hardware with an AI-as-a-Service model, SambaNova aims to simplify AI adoption for enterprises, competing not just on raw performance but on ease of use, efficiency, and scalability in the rapidly evolving AI market.

The table below is a replica of the table that was published by Karl Freund, Founder and Principal Analyst at Cambrian-AI Research LLC, to show the competitive landscape of AI chip vendors. To read the full Forbes article titled "AI Chip Vendors: A Look At Who's Who In The Zoo In 2024," click here.

‍

‍

Prominent News

On SambaNova Cloud, Tülu 3 405B, A New Model Better than DeepSeek V3, February 4, 2025. - SambaNova announces the addition of Tülu 3 405B, an Ai2-developed Llama 405B fine-tune that outperforms DeepSeek V3 on SambaNova Cloud and achieves over 90 tokens/second through their RDU technology. Click here to read more.‍

Hugging Face Partners with SambaNova to Supercharge its Inference API Capabilities, January 28, 2025 - SambaNova's newly launched inference API enables Hugging Face developers to access fast inference speeds on RDU Hardware and utilize their full model lineup, including the Llama 3 series with Llama Guard and various Qwen models. Click here to read more.‍

Oak Ridge National Laboratory Selects SambaNova to Expand Its Research in Secure and Energy-Efficient AI, November 18, 2024 - SambaNova announces that the U.S. Department of Energy’s Oak Ridge National Laboratory has deployed SambaNova Suite for secure and energy-efficient AI research, marking another national laboratory adoption of their AI platform. Click here to read more.

‍

Return on Investment (ROI)

ROI computed as of February 10, 2025.

‍

‍

Funding Round in USD

‍

<row>

<card>

Series D-1 Valuation

<span class="date">2021-04-09</span>
<span class="value">$5.0B</span>

</card>

<card>

Implied Valuation

<span class="date">2025-02-10</span>
<span class="value">$2.13B<span>

</card>

<card>

Change Since Last Round

<span class="date">2021-04-09 → 2025-02-10</span>

<span class="value red">-57.48%</span>

</card>

</row>

‍

‍

Quarterly Bid:Ask Volume Ratio

PM Insights highlights the quarterly bid:ask volume ratio where there were at least three institutionally-sized broker contributions within that period.

SambaNova received a high volume of bids and offers during Q3 of 2024 and has seen volumes trickle down since then. Between January 1, 2025, and February 10, 2025, SambaNova had $3M in total bid and ask volume, with bid and ask volumes being evenly distributed.

‍

‍

90D Activity

Activity at each date comprises the total unique broker contributions (bids/asks/trades) within a 90-day rolling window ending on that date. SambaNova's activity began increasing in Q2 of 2024 and peaked at 19 contributions by the end of 2024. As of February 10, 2025, the activity level stands at 12 contributions.

‍

‍