Markets Open
Global Markets
S&P 500 7,446.61 ▼ -0.8% DOW 52,227.07 ▼ -1.3% NASDAQ 25,653.41 ▼ -0.6% RUSSELL 2K 2,956 ▼ -0.9% VIX 17.62 ▲ +9.2% GOLD 4,063.3 ▼ -2.0% CRUDE OIL 75.6 ▲ +7.3% EUR/USD 1.14 ▼ -0.3% BTC 61,860 ▼ -1.7% ETH 1,728.92 ▼ -2.2%
Markets

SambaNova reaches $11 billion valuation after $1 billion financing

The AI chip startup’s latest round, led by General Atlantic, underscores investor demand for inference hardware alternatives to Nvidia.

Amanda Ross

By Amanda Ross · Deals Correspondent

· 3 min read

SambaNova reaches $11 billion valuation after $1 billion financing
Photo: CNBC

SambaNova Systems has raised $1 billion in new financing, valuing the AI semiconductor startup at $11 billion, according to CNBC. The round extends private-market funding for companies seeking to compete with Nvidia in chips used to run artificial intelligence models.

General Atlantic led the financing, with Seligman Ventures, T. Rowe Price and Capital Group also participating, CNBC reported. The announcement follows a separate raise of more than $350 million earlier this year from backers including Intel, which also entered into a partnership with SambaNova.

Focus on inference workloads

SambaNova is targeting the market for inference chips, the processors used after an AI model has been trained to produce answers, generate content or perform tasks. Inference has drawn more attention as companies deploy larger and more complex AI agents and seek to cut the cost and time required to run them.

The company sells its SN50 chip inside a server system designed for use in data centers, CNBC reported. That approach differs from the graphics processing unit, or GPU, architecture that has made Nvidia central to the training of large AI models.

Inference hardware works by processing requests against an existing model, rather than repeatedly adjusting the model’s parameters during training. For large enterprises, performance, power use and deployment location can affect both operating costs and control over sensitive data.

SambaNova is also emphasizing on-premises deployments, where its systems are installed in a data center controlled by the customer. The company has said that running inference on premises can improve speed and security because the user controls the deployment instead of relying on a third-party cloud provider or AI lab.

JPMorgan Chase said Wednesday that it would deploy SambaNova systems for on-premises inference in demanding enterprise AI workloads, according to CNBC. The bank’s planned use highlights the kind of regulated, data-intensive customer SambaNova is pursuing.

Capital follows the AI chip buildout

The latest financing comes as semiconductor companies tied to AI infrastructure have drawn strong public-market interest. CNBC reported that the PHLX semiconductor index, which tracks a basket of chip stocks, is up around 80% this year.

Private markets have also been active in AI chip companies that aim to challenge established suppliers. South Korean startup Rebellions is preparing for an initial public offering on the Kospi in the first or second quarter of 2027, its chief executive Sunghyun Park told CNBC.

Nvidia has also engaged with the inference chip segment. Last year, the company signed a deal to license technology from Groq, another startup focused on inference processors, CNBC reported.

SambaNova’s new valuation shows that investors continue to fund alternative approaches to AI computation, particularly where enterprises want dedicated systems for running models inside their own infrastructure. Nvidia remains the dominant reference point for the sector, but the latest round indicates that capital is flowing to companies building specialized chips and systems around inference demand.

This story draws on original reporting from CNBC.

More from Markets

All Markets →