SambaNova raises $350M, partners with Intel to deploy SN50 chip claiming 5x speed over Nvidia B200
12 Sources
[1]
Sambanova introduces new AI accelerator, partners with Intel to deploy Xeon CPUs for inferencing and agentic workloads -- Sambanova claims SN50 chip is three times more efficient than Nvidia B200
This week, Intel and SambaNova entered a multi-year strategic collaboration aimed at building large-scale AI inference infrastructure around Intel Xeon platforms and SambaNova AI accelerators. In addition, SambaNova introduced its SN50 AI processor for agentic inference, and inked a deal with
[2]
Intel-Backed SambaNova Raises Cash, Touts SoftBank Chip Contract
Intel will combine SambaNova's offerings with its own chips in products for data center customers as part of a "multiyear collaboration", with the goal of serving customers until its own AI accelerators are ready. Artificial intelligence chip startup SambaNova Systems Inc. raised $350 million in a
[3]
SambaNova raises $350M with Intel backing
Upstart's 5th-gen RDU aims to undercut Nvidia's B200 on speed and cost AI infrastructure company SambaNova has raised $350 million to advance its dataflow architecture, which it pitches as an alternative to GPU-based AI systems. Some of the money came from Intel Capital, scotching rumors
[4]
AI chip startup SambaNova raises $350 million in Vista-led round, signs Intel partnership
Feb 24 (Reuters) - SambaNova Systems said on Tuesday it has raised $350 million in a new funding round and struck a partnership with Intel as it seeks to capitalise on surging demand for inference chips used in artificial intelligence applications. Inference chips, which run AI models and power
[5]
Intel partners with AI chip startup SambaNova after acquisition talks reportedly failed
In addition to running Intel, Lip-Bu Tan is chairman of artificial intelligence chipmaker SambaNova, which he first invested in eight years ago. Now Intel is pumping money into the startup as it tries to take on industry leader Nvidia. SambaNova, a maker of chips for running generative AI models,
[6]
SambaNova bags $350m, unveils deals with Intel, SoftBank
SoftBank will be the first to deploy the new SN50 chips, while Intel is partnering with SambaNova to rollout its Intel-powered AI cloud. Intel-backed Nvidia rival SambaNova has raised $350m in a Series E round led by Vista Equity Partners and Cambium Capital, with strong participation from Intel
[7]
SambaNova steps up its challenge to Nvidia with new chip, $350M funding and a powerful ally in Intel - SiliconANGLE
SambaNova steps up its challenge to Nvidia with new chip, $350M funding and a powerful ally in Intel Chipmaker SambaNova Systems Inc. unveiled its most advanced artificial intelligence processor today as it closed on a bumper $350 million late-stage round of funding from Vista Equity Partners,
[8]
Intel strikes deal with a chip startup its CEO invested in
Intel, which has lagged in making chips for artificial intelligence, is turning for help to a startup that its CEO helps lead as an investor and chair. The ailing Silicon Valley chip company has struck a multiyear technical partnership with SambaNova Systems, a 9-year-old semiconductor company
[9]
Intel Misses Another 'AI Opportunity' With SambaNova as Acquisition Talks Die Down, Settling for Xeon CPU Collaboration Instead
Intel's SambaNova acquisition was seen as a way to spearhead the company's AI strategy in inference, but it appears Team Blue has settled for much less. When we talk about Intel, the company is one of the only compute providers that hasn't managed to capitalize on the AI frenzy at all, and this
[10]
Intel Inks 'Multiyear' AI Inference Deal With SambaNova After Acquisition Talks End
With Intel's venture capital arm backing SambaNova Systems' new $350 million funding round, the AI chip startup plans to tap into Intel's 'global enterprise, cloud and partner channels' to drive sales of joint offerings for 'cloud-scale AI inference.' Intel plans to tap into its "enterprise, cloud
[11]
Intel : SambaNova Planning Multi-Year Collaboration for Xeon-Based AI Inference
As AI workloads become more diverse and complex, organizations are looking for different solutions for different needs. This is driving demand for more heterogeneous infrastructure-built on diverse compute, memory, networking and a consistent software foundation-to support inference at scale across
[12]
AI chip startup SambaNova raises $350 million in Vista-led round, signs Intel partnership
Feb 24 (Reuters) - SambaNova Systems said on Tuesday it has raised $350 million in a new funding round and struck a partnership with Intel as it seeks to capitalise on surging demand for inference chips used in artificial intelligence applications. Inference chips, which run AI models and power
Share
Copy Link
SambaNova secured $350 million in Series E funding led by Vista Equity Partners and announced a multi-year strategic collaboration with Intel to deploy AI inference solutions. The startup unveiled its SN50 AI chip, designed specifically for inference workloads, claiming it delivers five times faster performance and three times better efficiency than Nvidia's B200. SoftBank will be the first customer to deploy the technology in its Japanese data centers.
SambaNova Secures $350 Million Funding and Unveils SN50 AI Chip
SambaNova Systems has raised $350 million in a Series E funding round led by Vista Equity Partners and Cambium Capital, with significant participation from Intel Capital, marking a rare move by Vista beyond its traditional enterprise software focus
2
4
. The funding round was oversubscribed, with Battery Ventures, Mayfield Capital, and other new investors participating, valuing the company above $2 billion—still below its 2021 valuation of around $5 billion2
. The proceeds will fund expansion of the new SN50 AI chip, scale its SambaCloud platform, and deepen enterprise software integrations4
.Intel Partnership Aims to Challenge Nvidia's Dominance
SambaNova announced a multi-year strategic collaboration with Intel to build large-scale AI inference infrastructure around Intel Xeon platforms and SambaNova AI accelerators
1
. The Intel partnership will combine SambaNova's offerings with Intel's own chips in products for data center customers, serving as an interim solution until Intel's own AI accelerators are ready2
. Intel CEO Lip-Bu Tan, who also serves as chairman of SambaNova after first investing in the startup eight years ago, recused himself from partnership discussions to avoid conflicts of interest2
5
. This collaboration comes after acquisition talks between Intel and SambaNova stalled, with Intel initially discussing acquiring the startup for roughly $1.6 billion, including debt4
.New AI Accelerator Built Specifically for Agentic Inference Workloads
The SN50 AI chip represents a fundamental shift in focus toward inference rather than training, targeting real-time applications such as voice assistants with emphasis on low latency, memory bandwidth, and power consumption
1
. The dual-chiplet processor is based on SambaNova's Reconfigurable Data Unit architecture and features a three-tier memory subsystem—SRAM, HBM, and DDR5—designed to keep multiple models resident for rapid hot-swapping1
. Each RDU features 432 MB of on-chip SRAM, 64 GB of HBM2E memory good for 1.8 TB/s of bandwidth, and between 256 GB and 2 TB of DDR5 memory3
. The new chip delivers 2.5x higher 16-bit floating-point performance and 5x higher performance at FP8, working out to 1.6 and 3.2 petaFLOPS respectively3
.Performance Claims Position SN50 Against Nvidia B200
SambaNova claims the SN50 accelerator offers five times more compute than competing offerings, specifically positioning it against the Nvidia B200, with three times better efficiency and a threefold lower cost of ownership for inference compared to GPU-based systems
1
. According to SemiAnalysis's InferenceX benchmark results, when running with FP8 precision, a Llama 3.3 70B model with 1K input and 1K output tokens achieves 895 tokens per second per user on the SN50, compared to 184 tokens per second per user on Nvidia's B2001
. Rodrigo Liang, co-founder and CEO of SambaNova, stated that "AI is no longer a contest to build the biggest model," emphasizing that "the real race is about who can light up entire data centers with AI agents that answer instantly, never stall, and do it at a cost that turns AI from an experiment into the most profitable engine in the cloud"1
.Related Stories
Rack-Scale AI Solutions Target Enterprise and Government Markets
SambaNova positions the SN50 primarily as part of its SambaRack SN50 rack-scale solution rather than a standalone processor
1
. Each 20 kW SambaRack SN50 packs 16 SN50 RDU processors, and 16 racks can interconnect up to 256 accelerators using a multi-terabyte-per-second fabric, with each RDU equipped with 2.2 TB/s of bidirectional chip-to-chip bandwidth via a switched fabric1
3
. The company emphasizes that at 20 kW per rack, the system operates within existing data center power envelopes and relies on air cooling, eliminating the need for liquid cooling or data center modifications1
. A cluster of 256 SN50 accelerators is designed for extremely large models, including configurations exceeding 10 trillion parameters and context windows of more than 10 million tokens, positioning this capability as essential for reasoning-heavy and multi-model agentic AI workloads1
.SoftBank as First Customer Signals Market Validation
SoftBank will be the first customer to deploy the SN50 platform, using it to power AI computing in Japan within its AI data centers
2
4
. The Japanese company already uses older SambaNova products, demonstrating continued confidence in the technology2
. SambaNova expects to ship its SN50 accelerators later this year, with full SambaRack SN50 systems available in the second half of 2026, though pricing remains undisclosed1
3
. The joint effort between Intel and SambaNova targets AI inference solutions for AI-native companies, model providers, as well as enterprises and government organizations worldwide, leveraging Intel's traditional customer relationships1
. SambaNova's dataflow architecture, which aims to reduce data movement overheads by overlapping computation and communication, already positions it as one of the highest performing inference providers, with its SN40L accelerators able to serve models like the 230 billion parameter MiniMax M2 at up to 378 tokens per second3
. The company counts Hugging Face, Meta, and major AI labs as customers, and through hardware-software co-design with Intel, aims to provide cost-effective alternatives in the AI chip market dominated by GPUs3
5
. Kevork Kechichian, general manager of Intel's data center group, noted that "customers are asking for more choice and more efficient ways to scale AI," highlighting the growing demand for alternatives in generative AI deployments2
.References
Summarized by
Navi
[3]
[4]
Related Stories
SambaNova raises $1B at $11B valuation, quintuples worth in months as AI chip race intensifies
08 Jul 2026•Startups
Vista Equity Partners and Intel lead $350M funding round for AI chip startup SambaNova Systems
07 Feb 2026•Startups
Intel Eyes $5 Billion SambaNova Acquisition to Revitalize AI Strategy
31 Oct 2025•Business and Economy
Recent Highlights
1
AI Hallucination Nearly Sparked US Military Confrontation With Chinese Ship in Middle East
Technology
2
OpenAI AI Agent Hacks Australian Government Healthcare Website in First-of-Its-Kind Breach
Technology
3
AI leaders warn UN Security Council of existential risks to humanity, call for global safeguards
Policy and Regulation
Recent Highlights
Today's Top Stories
Don’t drown in AI news. We cut through the noise - filtering, ranking and summarizing the most important AI news, breakthroughs and research daily. Follow topics that matter to you and stay ahead.
