Eleven billion dollars. That is the new price tag for SambaNova Systems, a valuation that has more than tripled in just five months. The Palo Alto-based chipmaker announced a $1 billion Series F round today, led by General Atlantic, marking a massive vote of confidence in the company’s bet on "premium inference" hardware.

This capital injection arrives with unusual speed. Only five months ago, the company closed a $350 million Series E round. The rapid-fire fundraising underscores the frantic, capital-intensive race to build the infrastructure that will actually run the massive AI models currently being developed by frontier labs.

The Pivot to Private Infrastructure

While much of the AI industry’s attention remains fixed on the training of massive models in public clouds, SambaNova is carving out a different niche: on-premises, secure inference. The company’s latest win, a partnership with JPMorgan Chase to power the bank's internal AI systems, serves as a proof point for this strategy.

"Having JPMorgan Chase decide they’re going to use SambaNova for their inference solution is a big deal," CEO and co-founder Rodrigo Liang said. "It sends a message to the banking industry that it’s time not to completely depend on cloud services."

For banks and government entities, the reliance on public cloud providers for sensitive data is becoming a liability. SambaNova’s hardware, specifically its SN40L and upcoming SN50 chips, is designed to fit multi-trillion-parameter models onto a single rack. This density allows organizations to maintain control over their data while achieving the speed required for real-time AI applications.

A Complex Relationship with Intel

SambaNova’s trajectory has been anything but conventional. Just last December, reports surfaced that Intel was in talks to acquire the startup for roughly $1.6 billion. That deal never materialized, yet the two companies have only grown closer. Intel is not only a long-term backer but also a strategic partner, with the two firms co-developing products that leverage Intel’s Xeon processors alongside SambaNova’s specialized silicon.

When asked if the recent funding rounds signal a permanent commitment to independence, Liang remained measured. While he noted that the company is "always being approached" by potential acquirers, the current momentum is pushing the firm toward an eventual public offering. For now, the priority is scaling production to meet what Liang describes as an "incredible wave of demand."

What This Means for Enterprise AI

SambaNova is positioning itself to capture the "huge amount of revenue" that remains on the table as enterprises move beyond the experimental phase. The company categorizes its target market into three buckets: sovereign clouds, neoclouds, and large-scale enterprises.

With the new capital, the company plans to shore up its supply chain. In the current market, hardware availability is the primary bottleneck for any company attempting to deploy AI at scale. By securing materials now, SambaNova is attempting to insulate itself from the volatility that has plagued other hardware startups.

Key Takeaways

  • Rapid Valuation Growth: SambaNova’s valuation has jumped to $11 billion, a significant increase from its $350 million Series E round just five months ago.
  • Strategic Focus: The company is doubling down on on-premises, secure inference for high-stakes industries like banking, moving away from a pure reliance on public cloud providers.
  • Supply Chain Security: The $1 billion in fresh capital is earmarked primarily for scaling production and securing the supply chain to meet the surging demand for its SN50 chips.

As the company prepares to ship its next-generation SN50 hardware in the second half of 2026, the focus will shift from fundraising to execution. The partnership with SoftBank, which is slated to be the first deployment partner for the new chip, will be the next major milestone. For the broader market, the success of SambaNova’s private infrastructure model will determine whether the future of enterprise AI remains in the cloud or moves back into the data center.