SambaNova Raises $1B at $11B Valuation as Investors Bankroll Nvidia's Challengers

SambaNova has completed the first close of a $1 billion Series F round at an $11 billion post-money valuation, the Palo Alto AI chipmaker announced — a raise that comes just five months after its previous megaround and signals investors' growing appetite to fund credible alternatives to Nvidia's data center dominance.

The Round

General Atlantic led the financing, with a sprawling investor list that includes Seligman Ventures, T. Rowe Price, Capital Group, Battery Ventures, BlackRock, Intel Capital, and the Qatar Investment Authority. The company described this as a first close and said additional investors are expected to join a second close in the coming weeks.

SambaNova will use the proceeds to expand capacity, accelerate product development, and scale deployments across enterprises, neo-clouds, sovereign AI customers, and service providers.

The JPMorgan Signal

More telling than the valuation was the customer disclosure: JPMorgan Chase has selected SambaNova as an inference infrastructure partner, deploying the company's SN40 and SN50 systems to power secure, on-premises AI inference. For a chip startup, landing one of the world's largest banks as a named on-prem customer is a validation that benchmark scores cannot buy.

The deal underscores SambaNova's strategic wedge: enterprises and governments that want frontier inference performance without sending sensitive data to a hyperscaler cloud. On-prem and sovereign deployments are precisely the segment where Nvidia's cloud-centric ecosystem is least entrenched — and where a specialized inference architecture can compete on total cost and data control rather than raw ubiquity.

The Inference Land Grab

SambaNova's raise fits a broader pattern. As the AI industry pivots from training frontier models to serving them at scale, inference — not training — is where the durable spending lives, and a cohort of well-funded challengers is racing to claim it. Groq, Cerebras, and a wave of custom-silicon startups are all pitching variations of the same thesis: that inference workloads are different enough from training to justify purpose-built hardware, and large enough to sustain multiple winners.

The $11 billion valuation, up sharply from SambaNova's position just months earlier, reflects investor conviction that the inference market is big enough to support more than one architecture. It also reflects the reality that even richly funded challengers must keep raising: competing with Nvidia is a capital-intensive endurance test, and SambaNova's back-to-back megarounds suggest the company intends to stay in it.