SambaNova raises $1 billion in Series F to expand custom AI inference chips

SambaNova Systems has completed the first close of a $1 billion Series F round, reaching an $11 billion post-money valuation in what is one of the largest AI chip funding rounds of 2026. General Atlantic led the round, joined by T. Rowe Price, Capital Group, BlackRock, Intel Capital, Vista Equity Partners, Battery Ventures, the Qatar Investment Authority, and several other institutional investors, according to TechCrunch.

SambaNova builds full-stack AI inference infrastructure: custom silicon (the SN40 and SN50 chips), integrated hardware systems, and cloud services designed specifically for running large language models at enterprise scale. That focus on inference — the process of generating responses from deployed AI models — is increasingly where the real commercial demand is concentrated, as companies shift from training frontier models to deploying them in production.

The most significant customer disclosure in the announcement is JPMorganChase, which has selected SambaNova as an inference infrastructure partner and is deploying SN40 and SN50 systems for on-premises AI inference. A Tier 1 bank operating AI workloads entirely on-prem represents a direct response to enterprise concerns about data sovereignty and regulatory compliance that public cloud providers have struggled to fully address.

The round arrives five months after SambaNova's previous major financing, reflecting sustained investor appetite for infrastructure alternatives to Nvidia and the hyperscalers. SambaNova says it will use the proceeds to expand capacity, accelerate chip development, and grow deployments among enterprises, sovereign AI programs, and emerging neo-cloud providers.

The $11 billion valuation is still a fraction of Nvidia's dominance, but SambaNova has carved a credible niche by targeting inference efficiency rather than competing on raw training throughput — and the JPMorganChase win suggests that niche is worth a great deal to the right customers.