Amid fierce competition for AI dominance, Intel and SambaNova join forces, promising cheaper, faster inference at scale with new chips and cloud-powered enterprise solutions.
Intel and artificial intelligence (AI) provider SambaNova have entered a multiyear partnership to reduce costs for enterprises deploying large-scale inference and agent-based AI systems.
The agreement, announced on Tuesday, February 24, 2026, will see Intel’s infrastructure underpin SambaNova’s AI cloud offering, enabling support for large language and multimodal models.
SambaNova confirmed it has raised $350 million in Series E funding, with Intel Capital among the investors. The financing will be used to expand manufacturing and cloud capacity, including the rollout of the new SN50 AI chip. The company claims the chip delivers agentic AI at 1/3 the cost of traditional GPUs and 5x the performance of rival processors.
SoftBank Corp. will be the first to deploy the SN50 in its Asia-Pacific data centres later this year.
Intel’s involvement is designed to accelerate the launch of an Intel-powered AI cloud built on Xeon-based infrastructure to deliver scalable inference for applications such as reasoning, code generation, and multimodal workflows.
According to a report by CIO Dive, the partnership complements Intel’s existing GPU commitments and forms part of its wider strategy to develop heterogeneous AI data centres that integrate processors, GPUs, networking, and storage.
Industry analysts suggest the move reflects Intel’s efforts to regain ground in the competitive AI compute market. They noted that while model training is dominated by NVIDIA, inference remains “up for grabs,” with enterprises seeking integrated solutions.
An analyst also described the partnership as a lower-risk alternative to acquisition, allowing both firms to prove the technology before considering deeper consolidation.
Meanwhile, Intel CEO Lip-Bu Tan, who also chairs SambaNova’s board, is overseeing the company’s turnaround strategy following supply chain challenges last year. Executives at Intel emphasised that combining its compute and memory expertise with SambaNova’s full-stack AI systems offers enterprises a viable alternative to GPU-based deployments.
The collaboration positions both companies to compete more directly in the evolving inference market, where efficiency and scalability are increasingly critical for enterprise adoption.
