The published story overstates SambaNova's position. The verifiable record points to Intel acquisition talks and a sharply lower implied valuation, not a fresh $1 billion round at $11 billion.
SambaNova is still an important AI chip company, but this article can't stand as written. I couldn't verify the reported July 8 Series F, the $11 billion post money valuation, or a JPMorganChase deployment of SN40 and SN50 systems from reliable public reporting. Worse, the available reporting points the other way.
Wired reported in December 2025 that Intel had signed a nonbinding term sheet to acquire SambaNova Systems, after Bloomberg had earlier reported Intel's interest. The Economic Times, citing Bloomberg, put the possible deal value at about $1.6 billion including debt. That's not a rounding error. It's a completely different story from a private company raising $1 billion at an $11 billion valuation.
That's the correction you have to make before you publish anything here.
The Real Story Is Intel, Not JPMorgan
SambaNova was founded in 2017 by Rodrigo Liang, Kunle Olukotun and Christopher Ré, and it has long sold itself as a full stack AI platform built around its Reconfigurable Dataflow Unit architecture. That part is real. So is Intel's long connection to the company. Wired noted that Intel CEO Lip-Bu Tan is SambaNova's chairman, and Intel Capital has been an investor in the startup.
That relationship now matters more than any unsupported funding claim. If Intel is trying to buy SambaNova, it says something blunt about both companies. Intel still needs credible AI accelerator technology. SambaNova, once one of the better funded challengers to Nvidia, has struggled to match the market power its earlier valuation implied.
Look at the numbers. According to PitchBook data cited by Wired, SambaNova had raised about $1.14 billion by early 2025. It reached a valuation above $5 billion after a $676 million Series D round led by SoftBank Vision Fund 2 in 2021. Wired also reported that BlackRock had marked down the value of its SambaNova shares by 17 percent over the prior year, citing The Information.
Those details don't fit the draft's thesis. They point to a company whose technology may still be valuable, but whose investor pricing has moved in the wrong direction.
Why The Original Framing Fails
The draft says JPMorganChase chose SambaNova over Nvidia for on premises inference infrastructure. That would be a meaningful proof point if it were verified. It isn't verified in the research I could find. There was no reliable public confirmation of a multi-year JPMorgan agreement, no clear announcement from JPMorganChase, and no current public source backing the claim that SN40 and SN50 systems are being deployed inside the bank's data centers.
You can't build an article around that. Not for this publication.
The safer and stronger piece is about valuation pressure in AI hardware. Nvidia still dominates training and inference infrastructure, while challengers have to prove they can win production workloads without burning through venture funding faster than revenue arrives. SambaNova has an interesting architecture. Interesting architecture doesn't automatically become market share.
That is where Intel comes in. A deal near $1.6 billion, if completed at the level reported by Bloomberg and picked up by The Economic Times, would sit far below SambaNova's 2021 valuation. It would also give Intel a way to fold SambaNova's inference technology into a broader AI push at a time when Intel has badly trailed Nvidia in accelerators.
There's still no clean ending here. Wired described the term sheet as nonbinding, which means the deal could still fall apart after due diligence, regulatory review or final price negotiations. SambaNova and Intel also declined or did not provide comment in that reporting. Those are dry details, but they matter. They keep the story honest.
The piece to run isn't that SambaNova just became an $11 billion AI chip winner. The piece to run is that one of Nvidia's better known private challengers may be headed into Intel's arms at a much lower valuation than its last boom era price. That's less flattering. It's also the story the facts support.
Also read: Helsing is closing in on an $18 billion valuation as Europe's defense tech boom accelerates • Anthropic's Billing System Just Failed Two Different Ways in One Week • TSMC's June Sales Jump 68 Percent as AI Chip Demand Keeps Climbing
