The AI inference-focused chip manufacturer SambaNova announced the first close of a Series F funding round of over one billion US dollars on July 8, 2026. According to the announcement from lead investor General Atlantic, the company is valued at eleven billion dollars (post-money). It is one of the largest AI infrastructure financings of the summer – and a barometer for where capital is currently flowing in the industry.
The Key Data of the Round
The round is led by the growth equity firm General Atlantic. According to the announcement, the investor list includes Intel Capital, BlackRock, T. Rowe Price, Capital Group, Vista Equity Partners, the Qatar Investment Authority, as well as Seligman Ventures and Battery Ventures. This is explicitly a first close: As TechCrunch reports, a second close with additional investors is expected in the coming weeks.
The pace of the valuation development is remarkable. SambaNova had only completed a Series E round of 350 million dollars about five months earlier, in February 2026. At the end of 2025, the valuation of the company founded in 2017 in San Jose was still reported to be around 1.6 billion dollars – at a time when Intel was discussing an acquisition. The jump to eleven billion dollars within a few months illustrates how aggressively investors are currently betting on challengers to Nvidia.
JPMorgan as a Customer, Not as an Investor
Alongside the financing, SambaNova announced a prominent customer: JPMorgan Chase is using its in-house SN40L and SN50 systems for local, private AI inference. Important for context: The bank is acting here as a customer, not as a capital provider for the round. As Quartz reports, JPMorgan’s responsible infrastructure CIO justified the choice with high demands for performance, control, and reliability.
For SambaNova, this reference customer is strategically valuable. Large financial institutions process highly sensitive data and often hesitate to fully migrate AI workloads to public clouds. An on-premises operation, where the models run in their own data center, addresses this concern directly – and marks a growing market beyond the major cloud providers.
Why Inference is Becoming a Battleground
SambaNova is consciously positioning itself not in the training of large models, where Nvidia’s GPUs dominate the market, but in inference – the execution of already trained models in ongoing operations. With every user request to an AI service, inference costs arise, which is why this part of the value chain is increasingly coming to the forefront economically. CEO Rodrigo Liang ties the valuation directly to this: It underscores the central role that fast inference now plays in the enterprise AI stack.
The company makes bold claims about its technical advantages – for example, that its systems process the decoding part of inference multiple times faster than competing approaches. Such performance figures are manufacturer claims and have not been independently verified; they should be read with caution. However, the market movement behind it is undisputed: Capital that has long almost reflexively flowed into training hardware is increasingly seeking alternatives and more efficient ways to keep AI affordable in operation.
Context
The round fits into a larger pattern. Investors are backing a number of Nvidia challengers as demand for AI computing power continues and the costs of established providers rise. Whether SambaNova can justify the high valuation in the long term depends on whether reference customers like JPMorgan turn into broad, recurring revenue. The company does not provide specific revenue figures. Liang has described an IPO as a likely path; media reports mention 2027 as a possible target, but a date has not been confirmed.
For industry observers, the deal is primarily a signal: The second major AI infrastructure question – after the one about sufficient training power – is who can deliver inference quickly, securely, and affordably enough. SambaNova has secured fresh capital with this round to answer this question for itself.
