SambaNova SN40L
Reconfigurable dataflow units with three-tier memory for serving many models per node.
Full SambaNova SN40L record →Head to head
SambaNova SN40L vs Google TPU Ironwood
SambaNova SN40L (SambaNova) and Google TPU Ironwood (Google) both sit in Chips & Compute. SambaNova SN40L is the cheaper entry point at Free tier available; pay-as-you-go developer and subscription-based enterprise plans offered.. Google TPU Ironwood shipped more tracked changes in the last 30 days (11 vs 2). Every value below comes from the latest crawl of the vendors' own pages.
| Attribute | SambaNova SN40L SambaNova | Google TPU Ironwood |
|---|---|---|
| Segment | Chips & Compute | Chips & Compute |
| Version | SN40L | 7th-generation (Ironwood) |
| Entry cost | Free tier available; pay-as-you-go developer and subscription-based enterprise plans offered. | from $5.4/unit |
| Pricing tiers | Free free · Developer Pay as you go free | On Demand (us-central1) $12 · DWS Flex-start (us-central1) $6 · DWS Calendar Mode (us-central1) $8.4 · 1-year Commitment (us-central1) $8.4 · 3-year Commitment (us-central1) $5.4 · On Demand (europe-west2) $13.2 · DWS Flex-start (europe-west2) $6 · DWS Calendar Mode (europe-west2) $8.4 |
| Model stack | DeepSeek-V3.1 · gpt-oss-120b · Llama / DeepSeek open weights · Llama 3 · Llama 3.1 · Llama 3.1 (8B, 70B, 405B) · Llama 3.1 405B · Llama 3.1 70B · Llama 3.1 8B · MiniMax M2.7 · MiniMax M3.1 Flash · Production & Preview models · RDU · RDU dataflow · SambaNova RDU · SambaNova SN40L RDUs · SambaRack SN40-16 · SN40L RDU | Gemini · Google TPU Ironwood · Ironwood · Ironwood (TPU v7) · Ironwood TPU · Ironwood TPU (7th Generation) · TPU v7 · TPU7x · TPU7x (Ironwood) · XLA compiler |
| Context window | — | — |
| Public API | ||
| Routes models | ||
| Changes / 30d | 2 | 11 |
| Origin | Global | Global |
| Capabilities | three-tier memory, many models per node, fast open-model serving, Reconfigurable Dataflow Unit (RDU), Three-tier memory architecture, Dataflow processing, Low power inference, Multi-model concurrency | 9,216-chip pods, inference optimised, liquid cooled, JAX + PyTorch/XLA, Large-scale training, Reasoning and inference, 9,216 liquid-cooled chips per pod, 42.5 ExaFlops performance |
Google TPU Ironwood
v7 TPU pods scaling to 9,216 chips, available through Google Cloud and used for Gemini.
Full Google TPU Ironwood record →