SambaNova raises $1 billion Series E at $11 billion valuation, securing resources for AI inference chip and systems.

SambaNova has closed a $1 billion Series F funding round, bringing the AI chip startup's valuation to $11 billion. The round was led by General Atlantic, with participation from Seligman Ventures, T. Rowe Price, and Capital Group. This follows over $350 million the company raised in the first half of 2026, including backing from Intel, with whom SambaNova has established a deep partnership.

Rodrigo Liang, SambaNova's co-founder and CEO, announced the milestone at the Raise AI summit in Paris, emphasizing that the shift to inference technology has opened "an entirely new growth space" for the company. As an independent enterprise, he said, SambaNova can respond quickly to market demand and expand across a wider range of industries. The company plans to use the funding to accelerate deployment of rack systems for customers and scale operations further. Liang also revealed that SambaNova is actively considering a U.S. IPO in 2027.

The investment underscores capital markets' aggressive pursuit of AI chip challengers to Nvidia. As the industry transitions from the training phase to large-scale inference applications, inference chips have become the sector's new competitive battleground. While Nvidia dominates GPU-based training for large language models, SambaNova has built its business around inference—running already-trained models at lower cost without sacrificing speed.

SambaNova's core offering is a server unit integrated with its SN50 chip, deployable directly in customer data centers. Unlike Nvidia's GPU architecture, the SN50 is optimized specifically for inference workloads. The company's strategy centers on on-premise deployment, allowing enterprises to run AI models within their own firewalls rather than relying on cloud providers. This model delivers critical advantages in data security and processing latency, particularly valuable for data-sensitive sectors like finance and healthcare.

JPMorgan Chase has emerged as a marquee customer, announcing it will deploy SambaNova systems to handle demanding enterprise AI workloads and on-premise inference tasks. Liang stressed that in banking—where data is "the lifeline"—on-premise deployment ensures models run within the firewall, allowing enterprises to maintain complete control over their models and sensitive data.

The competitive intensity in the inference chip market is evident beyond SambaNova. Rebellions, a South Korean AI chip startup, recently disclosed plans for an IPO on Korea's KOSPI market in the first or second quarter of 2027. Nvidia, meanwhile, is actively defending its position; the company licensed technology from Groq, an inference chip startup, to reinforce its leadership.

As AI applications gradually shift from training to inference at scale, competition in the chip market is expanding across technical architecture, deployment models, and industry-specific solutions.