SambaNova Raises $1 Billion at $11 Billion Valuation, Posing a Strong Challenge in Nvidia’s Inference Track

Unlike giants such as Nvidia ( NVDA) that focus primarily on GPUs for large-scale model training, SambaNova has centered its business on the AI inference chip market.

Inference chips are mainly used to run already-trained AI models, with the core objective of reducing costs while maintaining operational speed. As the deployment of complex AI models accelerates, the importance of the inference stage has become increasingly prominent, driving explosive growth in market demand.

SambaNova's core product is a server unit integrated with its latest SN50 chip, which can be deployed directly in data centers. Its architectural design is fundamentally different from Nvidia's GPU architecture, focusing more on the optimization of inference scenarios.

Beyond the inference chips themselves, SambaNova is also betting heavily on on-premise deployment for enterprises. Customers can install its servers in their own data centers to run AI models internally without relying on third-party cloud service providers. This model offers distinct advantages in data security and processing speed, making it particularly attractive to highly data-sensitive industries such as finance and healthcare.

Recently, JPMorgan Chase ( JPM) announced it would deploy SambaNova's systems to handle demanding enterprise AI workloads and on-premise inference tasks.

Rodrigo Liang stated that for industries like banking, where data is the lifeline, on-premise deployment ensures that AI systems run within the firewall, allowing enterprises to maintain complete control over their models and private data to achieve highly secure and private AI applications. This on-premise deployment strategy has carved out a unique growth path for SambaNova in the highly competitive AI chip market.

The rise of SambaNova is just a microcosm of the intensifying competition in the AI inference chip market. Currently, a large number of startups focusing on inference chips are emerging in the market, attempting to challenge the dominance of giants like Nvidia in this burgeoning sector.

Park Sunghyun, CEO of South Korean AI chip company Rebellions, recently revealed that the company plans to conduct an initial public offering on South Korea's KOSPI market in the first or second quarter of 2027, further demonstrating the heat in the inference chip sector.

In the face of competition in the inference chip market, Nvidia is also actively positioning itself. Last year, Nvidia signed an agreement with inference chip startup Groq to license the latter's relevant technology, in an attempt to consolidate its leading position in the AI chip field.

As AI model applications gradually transition from the training phase to large-scale inference, the scope of competition in the AI chip market is expanding further, with technical architectures, deployment models, and industry solutions all becoming focal points of competition.

This content was translated using AI and reviewed for clarity. It is for informational purposes only.