SambaNova raises $1B at $11B valuation amid surging AI infrastructure demand
The Palo Alto-based AI chip startup's latest mega-round signals that investor appetite for Nvidia alternatives shows no signs of cooling
SambaNova Systems just closed a $1 billion funding round that values the company at $11 billion, a staggering leap for a startup that was worth roughly half that just a few years ago. The raise lands the AI chip maker squarely in the conversation alongside the handful of companies trying to challenge Nvidia’s dominance in AI hardware.
Inside the round and what it signals
The fundraise reportedly targeted between $800 million and $1 billion, ultimately landing at the top end.
SambaNova completed a roughly $350 million Series E just months earlier, in February 2026. That round was led by Vista Equity Partners and Cambium Capital, with Intel contributing approximately $100 million. Intel’s involvement is worth noting given that the two companies previously explored acquisition talks that ultimately stalled.
The new round pushes SambaNova’s total funding well past the $1.4 billion mark it had already accumulated. For context, the company raised a $676 million Series D back in 2021 at a $5.1 billion valuation. Going from $5.1 billion to $11 billion in a few years is the kind of trajectory that makes venture capitalists very happy and competitors very nervous.
What SambaNova actually does
Founded in 2017 by a group of Stanford engineers, SambaNova builds integrated AI systems that combine custom-designed hardware with purpose-built software. Think of it as an alternative computing stack, one designed from the ground up for AI workloads rather than adapted from general-purpose chips.
The company’s flagship technology centers on its Reconfigurable Dataflow Units, or RDUs. These are chips specifically architected to handle AI inference tasks with an emphasis on energy efficiency and throughput. While Nvidia’s GPUs remain the default choice for training massive models, inference—the process of actually running those models in production—is a different computational problem where specialized hardware can offer meaningful advantages in cost and power consumption.
SambaNova targets enterprise and government customers, organizations that need to run AI models at scale but are increasingly concerned about energy costs and hardware lock-in.
The competitive landscape and why this matters for investors
Nvidia controls the lion’s share of the market with its GPU ecosystem. AMD has been making steady progress. Google has its TPUs. Amazon has Trainium and Inferentia. And a growing roster of startups, including Cerebras, Groq, and now a significantly better-funded SambaNova, are all vying for the portion of the market that wants alternatives.
The stalled Intel acquisition discussions add an interesting wrinkle. Intel’s decision to invest $100 million in the Series E rather than acquire the company outright suggests it sees value in SambaNova as an independent partner rather than an internal division.
The risk, as always with pre-IPO AI hardware companies, is execution. Designing competitive chips is hard. Manufacturing them at scale is harder. And convincing enterprise customers to adopt a new computing platform when they already have Nvidia relationships is perhaps the hardest challenge of all. But with $1 billion in fresh capital and a valuation that reflects genuine market confidence, SambaNova has bought itself the runway to find out whether its technology can match its ambitions.