SambaNova Systems has raised $1 billion at an $11 billion valuation in the first close of its Series F, led by General Atlantic, five months after a $350 million Series E in February 2026. The Palo Alto chip maker also announced JPMorgan Chase as an inference-infrastructure partner, with its SN40L and SN50 systems set to power secure, on-premises AI inference at the bank. Founded in 2017, the company had been in acquisition talks with Intel in December at a roughly $1.6 billion valuation — a mark the new round has now cleared roughly seven times over.
CEO Rodrigo Liang told TechCrunch the round is not yet closed. More investors are expected to join in the second close, alongside a lineup that already includes Seligman Ventures, T. Rowe Price Associates, Capital Group, BlackRock, Qatar Investment Authority, Vista Equity Partners, Battery Ventures, and Volantis. Intel, a backer since SambaNova's Series C, also participated.
“In the next few weeks, a few more investors will be coming in, and the second close is likely to finish up.”— Rodrigo Liang, CEO and co-founder of SambaNova
Liang was noncommittal on whether the back-to-back Series E and F rounds mean SambaNova has settled on staying independent, saying the company keeps fielding interest. He pointed to an eventual public listing as the more likely outcome if growth continues, without ruling out an acquisition in what he described as a dynamic AI market.
Key facts
- 01SambaNova raised $1B at an $11B valuation in a Series F first close led by General Atlantic.
- 02The round lands five months after a $350M Series E in February 2026 that unveiled the SN50 chip.
- 03JPMorgan Chase selected SambaNova as an inference-infrastructure partner, deploying SN40L and SN50 systems on-premises.
- 04Intel, a backer since Series C, deepened ties after December reports of $1.6B acquisition talks.
- 05SN50 ships in the second half of 2026 with SoftBank as first deployment partner.
The Intel relationship has moved well past passive investment. Five months ago, SambaNova announced a multi-year partnership with Intel to support AI inference development based on Intel's Xeon chip, and the two now co-develop products and take them to market together. That arrangement gives a nine-year-old startup access to Intel's manufacturing and enterprise sales reach at a moment when securing supply is the binding constraint on the AI chip market.
The JPMorgan Chase win matters as a signal more than as a single contract. SambaNova is pitching its systems as premium inference hardware built for models spanning trillions of parameters, fitting multi-trillion-parameter models onto a single rack. Landing one of the largest US banks as an on-premises inference customer sets a template for regulated industries that have been reluctant to run their most sensitive workloads on public cloud infrastructure.
Liang argued the deal marks a broader shift in how large financial institutions want to run AI. Banks of JPMorgan's scale, he said, are now building their own private, secure infrastructure to run inference on their most sensitive models — a move he expects to resonate beyond banking. That framing positions SambaNova against hyperscaler-hosted inference from AWS, Azure, and Google Cloud, and against Nvidia-based on-prem stacks that dominate the enterprise buildout.
The customer mix breaks into three segments, per Liang: sovereign clouds where governments fund local partners to build private clouds, neoclouds selling capacity to third parties, and enterprises building for their own use. Named customers now include JPMorgan Chase, Saudi Aramco, Intel, and Japanese firms Liang did not identify. SoftBank will be the first deployment partner for the SN50 when it begins shipping in the second half of 2026.
The $1 billion is earmarked largely for supply. Liang said the capital will go toward securing components and manufacturing capacity over the next 12 months, describing demand as an incredible wave that the company is still working to catch. SambaNova's SN40L launched in September 2023 in the cloud and became available on-premises from November 2023; the SN50 was unveiled in February 2026.
The obvious risk is that SambaNova is selling into a market where Nvidia's incumbency in inference remains overwhelming, and where AMD, Groq, Cerebras, and hyperscaler in-house silicon are all chasing the same enterprise budgets. Premium inference is a defensible niche only as long as the largest models remain hard to serve efficiently on general-purpose accelerators — a moat that narrows every time Nvidia ships a new generation. Whether JPMorgan's endorsement translates into a repeatable enterprise pipeline is the open question the next four quarters will answer.
An $11 billion valuation for a specialist inference vendor lands somewhere between vindication and pressure. It puts SambaNova in the tier of AI infrastructure companies that need enterprise revenue growth on a public-market trajectory, not just design wins. The JPMorgan announcement was timed to make exactly that case — that regulated buyers with real budgets are willing to bet against a cloud-only future, and that specialized chip vendors have a durable seat at the table when trillion-parameter models meet compliance requirements.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
