SambaNova Systems
Reconfigurable dataflow (RDU) systems for enterprise AI training and inference
SambaNova designs the Reconfigurable Dataflow Unit (RDU), sold as full SN40L systems and a cloud service aimed at enterprises and governments wanting to run large models on-prem. Its architecture uses a three-tier memory design (including large DRAM) to hold many large models on few chips, and in 2025 it emphasized fast inference for open frontier models. It competes with Nvidia in the enterprise/sovereign-AI niche and is fabbed at TSMC.
Product
RDU / SN40L systems + SambaNova Cloud
Edge
3-tier memory, many models per node
Market
Enterprise + sovereign/on-prem AI
How it fits the stack
SambaNova Systems with what it depends on (above) and what it feeds (below). The figure renders as a crawlable diagram and upgrades to an interactive 3D graph as it scrolls into view.
SambaNova Systems in the AI stack. SambaNova Systems with its immediate upstream dependencies (top) and downstream dependents (bottom) in the AI value chain. Hover a node in 3D, or read the full relationships below.
Graph data (text) — 3 entities, 2 relationships
- SambaNova Systems —depends on→ TSMC (Taiwan Semiconductor Manufacturing Company)
- SambaNova Systems —competes with→ Nvidia
Context — capital, rivals, policy · · 1