SambaNova Series F funding at $11B valuation accelerates enterprise AI inference hardware

SambaNova Systems, the AI inference hardware company, has raised $1 billion at an $11 billion post-money valuation in the first close of its Series F funding round. The round was led by General Atlantic, with more investors expected to join in a second close. This is the company's second major round in 2026, following a $350 million Series E in February when it also unveiled its SN50 chip. The funding signals continued investor confidence in non-Nvidia AI hardware for enterprise and sovereign cloud deployments.

What happened

The Series F first close was announced on July 8, 2026, by SambaNova CEO and co-founder Rodrigo Liang. The company also deepened its strategic relationships with two major partners:

• Intel: The chip giant, a backer since SambaNova's Series C, has extended its partnership to a multi-year agreement focused on AI inference development on Intel Xeon hardware. The two now co-develop products and bring them to market together, a relationship that CEO Liang characterized as allowing SambaNova to "leverage the scale of Intel."
• JPMorgan Chase: The bank has named SambaNova an "inference-infrastructure partner," adopting its SN40L and SN50 systems for secure, on-premises AI inference. Liang described the JPMorgan win as a signal to the banking industry that it is time to build private infrastructure rather than depend entirely on cloud services.

SambaNova's SN50 chip, unveiled in February 2026, is its next-generation inference accelerator designed for multi-trillion-parameter models. It is due to begin shipping to customers in the second half of 2026, with SoftBank named as its first deployment partner. The existing SN40L has been shipping since late 2023 in both cloud and on-premises versions.

Liang confirmed that the proceeds will be used to "secure the supply chain" and scale the business to meet what he called "an incredible wave of demand" over the next 12 months.

Why AI builders should care

For teams building AI products that depend on inference at scale, SambaNova's funding and partnerships signal a growing alternative to Nvidia-dominated cloud infrastructure. The company's focus on on-premises AI inference directly addresses a pain point for regulated industries: banks, governments, and healthcare providers that need to run large models on their own hardware for data security and latency reasons.

SambaNova targets three customer archetypes: sovereign clouds (government-funded local private clouds), neoclouds (alternative cloud providers), and enterprises building for their own use. Publicly named customers include JPMorgan Chase, Saudi Aramco, Intel, and other Japanese firms.

For developers, the practical implication is that deploying large models in private or hybrid environments may become more viable without relying on Nvidia's HGX or DGX systems. If SN50 delivers on its claim of running multi-trillion-parameter models on a single rack, that reduces the hardware cost and complexity of on-premises inference clusters.

Practical implications

The funding round provides SambaNova with the capital to scale production and reduce supply chain risk, which has been a bottleneck for many AI hardware startups. The Intel co-development partnership could also accelerate software optimization for Xeon-based servers, making it easier to run SambaNova hardware in existing data center footprints.

For builders evaluating inference hardware, the key differentiator SambaNova is selling is "premium inference" -- high throughput on the largest models. If your workload involves running GPT-4-class or larger models with low latency requirements, SambaNova's single-rack design could be compelling. However, the company has not published independent benchmarks comparing its hardware against Nvidia's H100/B200 or AMD's MI300X on specific model families or latency curves.

Caveats

Several important unknowns remain. First, the SN50 has not yet shipped to customers; availability is H2 2026, and SoftBank is the only confirmed deployment partner. Second, pricing for SN50 hardware has not been disclosed, making it difficult to compare total cost of ownership against cloud inference services or other on-premises options. Third, while JPMorgan Chase is a strong endorsement, the bank's deployment may be limited in scope; it is too early to call it a broad enterprise trend. Finally, despite the strong funding momentum, SambaNova's CEO left the door open for a potential exit, noting the company is "always being approached." The company's ultimate independence is not guaranteed.

FAQs

What is SambaNova and what do they do in AI hardware?

SambaNova develops AI inference hardware designed to run large models at scale, including the SN40L and next-generation SN50 systems. Its products target high-throughput, low-latency inference on multi-trillion-parameter models, with a focus on on-premises and sovereign/private cloud deployments. Recent funding and partnerships underscore its push into enterprise and regulated-industry use cases.

What does a Series F funding round mean for SambaNova?

A Series F at an $11 billion valuation signals strong investor confidence and capital to expand production, build supply chain resilience, and grow market share. The round includes multiple participants beyond the lead, including Intel, BlackRock, T. Rowe Price, and the Qatar Investment Authority. This round follows a recent $350 million Series E and shows broad backing for SambaNova's growth plan.

Who are SambaNova's key partners and customers?

Key partners include Intel (multi-year AI inference development partnership and co-developed products) and JPMorgan Chase (inference-infrastructure partner for on-premises deployment). Publicly named customers include JPMorgan Chase, Saudi Aramco, Intel, and other Japanese firms. SoftBank is the first deployment partner for the SN50 chip.

What is the SN50 chip and its role in SambaNova's products?

The SN50 is SambaNova's next-generation inference accelerator unveiled in February 2026. It is designed to handle large, multi-trillion-parameter models at high performance for on-premises deployment. Shipping to customers is planned for the second half of 2026, with SoftBank as the first deployment partner.