Key Takeaways

- SambaNova raised $1B at $11B valuation, up from a reported $1.6B Intel acquisition offer just 7 months ago
- JPMorgan Chase selected SambaNova as its inference infrastructure partner for on-premises AI
- The company remains open to acquisition despite back-to-back mega rounds, with IPO as the likely path
SambaNova Systems closed a $1 billion Series F at an $11 billion valuation, led by General Atlantic. The first close landed just five months after the AI chip maker raised $350 million in Series E. More investors are expected within weeks.
The speed here is notable. In December, Bloomberg reported Intel was in acquisition talks valuing SambaNova at roughly $1.6 billion. Seven months later, the company is worth nearly seven times that figure.
Why the valuation jumped so fast
Two factors drove the repricing. First, demand for AI inference hardware has outstripped supply. Second, SambaNova landed a signature enterprise customer: JPMorgan Chase.
The bank selected SambaNova as its "inference-infrastructure partner," deploying its SN40L and SN50 systems for on-premises AI. CEO Rodrigo Liang framed the deal as a market signal. "It sends a message to the banking industry that it's time not to completely depend on cloud services," he told TechCrunch. "These banks want heterogeneous infrastructure."
Banks handle sensitive data, regulatory constraints, and latency requirements that make public cloud inference impractical for certain workloads. JPMorgan's decision to build private AI infrastructure validates the on-premises inference market that SambaNova has bet its business on.
The Intel relationship deepens
Intel has backed SambaNova since Series C and participated in this round. The relationship goes beyond capital. In February, the two announced a multi-year partnership for AI inference development based on Intel's Xeon chips, with joint product development and go-to-market.
"That gives us a great relationship with them that lets us leverage the scale of Intel with the technology we have," Liang said.
Asked whether back-to-back mega rounds meant SambaNova had committed to independence, Liang was noncommittal. "We're always being approached," he said. The door stays open to acquisition, but momentum points toward "being public at some point."
Where the money goes
SambaNova will use the capital to secure its supply chain. "We're using that capital to secure the supply chain," Liang said, describing it as essential for fulfilling orders and sourcing materials over the next 12 months.
The company's SN50 chip, unveiled in February, ships in the second half of 2026. SoftBank is the first deployment partner. SambaNova positions itself as "premium inference," optimized for the largest models at the fastest speeds. The company claims it can fit multi-trillion-parameter models onto a single rack.
Three customer segments driving growth
Liang outlined three target markets. Sovereign clouds top the list, where governments fund local partners to build private AI infrastructure. Neoclouds come second. Third is enterprises building for internal use.
Beyond JPMorgan, SambaNova counts Saudi Aramco, Intel, and several Japanese firms as customers. Liang argued that most enterprise and government AI deployment remains ahead. "Enterprises and governments are just starting their AI journey," he said, with growth so far concentrated among model makers and frontier labs.
The investor syndicate
General Atlantic led the round. Other participants include Seligman Ventures, T. Rowe Price Associates, Capital Group, BlackRock, Battery Ventures, Vista Equity Partners, and Qatar Investment Authority. The breadth of the syndicate, spanning sovereign wealth, growth equity, and public market investors, suggests confidence in a near-term liquidity event.
Logicity's Take
The real story is the valuation gap between December and now. Intel's reported $1.6B offer would have been a fire sale. SambaNova's board clearly believed enterprise demand for on-prem inference would catch up to the hype, and the JPMorgan deal proved them right. For enterprises evaluating AI infrastructure, this raises a strategic question: does your inference workload belong in the cloud, on specialized silicon like SambaNova's, or on a hybrid stack? The answer depends on data sensitivity, latency requirements, and total cost of ownership. SambaNova's bet is that more enterprises will answer "not cloud" than current deployment patterns suggest.
What this signals for the AI chip market
SambaNova competes with NVIDIA, AMD, and a growing field of startups for AI inference workloads. Its pitch centers on purpose-built architecture versus repurposed GPUs. Founded in 2017 by Stanford professors and former Oracle executives, the company has spent nine years building silicon specifically for large model inference.
The $11 billion valuation puts it among the most valuable AI chip startups globally. Whether that holds depends on enterprise adoption and the competitive response from NVIDIA and hyperscalers building their own silicon.
Frequently Asked Questions
What does SambaNova do?
SambaNova builds AI chips and systems optimized for running large language models and generative AI workloads. Its hardware is designed specifically for AI inference, unlike GPUs originally built for graphics.
Why did SambaNova's valuation jump from $1.6B to $11B?
The company landed JPMorgan Chase as an inference partner and demonstrated enterprise demand for on-premises AI infrastructure. Rising AI chip demand and supply constraints also pushed valuations across the sector.
Is SambaNova going public?
CEO Rodrigo Liang said an IPO is the likely path, though the company remains open to acquisition offers.
Who are SambaNova's main customers?
JPMorgan Chase, Saudi Aramco, Intel, and several Japanese firms. The company targets sovereign clouds, neoclouds, and enterprises building AI for internal use.
When does the SN50 chip ship?
SambaNova's SN50 chip begins shipping in the second half of 2026, with SoftBank as the first deployment partner.
Another AI startup commanding a double-digit-billion valuation amid the 2026 funding surge
Need Help Implementing This?
Evaluating AI infrastructure options for your enterprise? Reach out to Logicity's consulting network for vendor-neutral guidance on cloud, on-prem, and hybrid AI deployment strategies.
Source: TechCrunch / Kate Park
Huma Shazia
Senior AI & Tech Writer
Produced with AI assistance and reviewed by the Logicity editorial team. Learn more in our Editorial Policy.
Related Articles
More in Trending Tech
Humanity Just Went Farther Into Space Than Ever Before — And Made It Back Alive
Four astronauts splashed down in the Pacific Ocean on April 10, 2026, after traveling farther from Earth than any human beings in history. The Artemis II crew shattered a 56-year-old distance record set by Apollo 13, journeying nearly 253,000 miles from our planet during their 10-day lunar flyby mission. This marks the first time humans have ventured beyond low Earth orbit since 1972.

Amflow's Electric Bikes Are Blowing The Competition Away
Amflow, the e-bike brand spun out of DJI, has just released two impressive new electric mountain bikes that are breaking the mold with unprecedented power, range, and lightness. The flagship bikes are powered by the innovative Avinox motors and come with features like onboard navigation and heart rate control.

Canva Just Made a Power Play: Here's What It Means for the Future of Design and Marketing
Canva has made a bold move by acquiring two companies, Simtheory and Ortto, to boost its AI and marketing automation capabilities. This strategic move is set to revolutionize the way teams work on design and marketing projects. With these acquisitions, Canva is poised to become an all-in-one platform for businesses and individuals alike.


