Gimlet Labs Raises $300M Series B at $3B Valuation to Build Multi-Silicon Inference Cloud
Gimlet Labs, an AI inference optimization startup, announced a $300M Series B at a $3B valuation led by a16z, with participation reported across posts including Sapphire, ARM, Microsoft M12, Samsung, and Hudson River Trading. The company dispatches AI inference workloads across heterogeneous chips to cut cost and improve efficiency, and the round signals strong investor conviction in inference infrastructure.
Confirmed
- $300M Series B led by a16z at a $3B valuation
- Backers vary by post: @dinabass cites ARM and Microsoft M12; @a16z cites SapphireVC; @AccBalanced cites Samsung and Hudson River Trading
- Business model: dispatching AI workloads across many chip types to optimize inference cost and efficiency
Why it matters
- Gimlet Labs calls itself the first "multi-silicon inference cloud," betting that inference becomes the dominant AI workload requiring infrastructure rebuilt from scratch
- @AccBalanced notes Hudson River Trading's participation is a meaningful signal: high-frequency trading is latency-sensitive, implying customers willing to pay for low-latency inference
- A chipstrat interview (via @AccBalanced's repost) reveals additional business model details, including fine-grained model-splitting and scheduling for inference
2026-09-04 ~ 2026-09-05 · 5 related posts
Primary sources
- [source] Gimlet Labs raises $300M at $3B valuation to split AI workloads across chip types — dinabass · 2026-09-04
- [source] Gimlet Labs Raises $300M Series B Led by a16z at $3B Valuation for Multi-Silicon Inference Cloud — a16z · 2026-09-04
- Gimlet Labs raises $300M Series B at $3B valuation for multi-silicon inference cloud — stuffyokodraws · 2026-09-05
- [source] Gimlet raises $300M at $3B valuation to end idle compute in AI inference — AccBalanced · 2026-09-05
1 near-duplicate retellings: AccBalanced