US authorities have granted antitrust clearance for Intel’s recent investment in AI chip company SambaNova, as indicated in a regulatory disclosure on 1 May (Friday), reported Reuters.
This development follows Intel’s $35m investment in SambaNova in February 2026, part of a wider Series E fundraising round amounting to $350m. Intel’s stake in the AI chip start-up increased to 8.2% from 6.8% as a result.
Go deeper with GlobalData
Access deeper industry intelligence
Experience unmatched clarity with a single platform that combines unique data, AI, and human expertise.
Vista Equity Partners and Cambium Capital led the funding, with Intel Capital participating significantly. SambaNova stated it will use the funds to augment production of its SN50 chip, advance its SambaCloud offering, and further integrate its enterprise software.
Based in San Jose, California, SambaNova said in February that it had released its SN50, a new chip aimed at large-scale AI deployments. The company said shipments would commence later this year. SambaNova claims the SN50 builds upon its Reconfigurable Data Unit (RDU) architecture, offering fivefold performance per accelerator and quadrupled network bandwidth over its predecessor.
According to the company, the chip’s architecture supports linking up to 256 accelerators, allowing the deployment of AI models with more than ten trillion parameters and context lengths exceeding ten million tokens.
The same month, Intel and SambaNova announced plans for a multi-year strategic partnership to deliver AI inference solutions using SambaNova technology and Intel’s infrastructure. This collaboration will focus on combining SambaNova’s systems with Intel’s CPUs, accelerators, and networking products.
The alliance aims to offer an alternative to GPU-based offerings, with a focus on providing consistent throughput and total cost of ownership for key open-source AI models.
Intel intends to support this initiative through continued investment, while the companies will also engage in joint marketing and sales. Both aim to advance solutions for data centres reliant on varied AI infrastructure.
In April 2026, SambaNova announced further progress in its collaboration with Intel.
The companies plan to deliver a hardware system integrating GPUs, Intel’s Xeon 6 processors, and SambaNova’s RDUs to meet the requirements of demanding agentic AI applications. This design is scheduled for release in the second half of 2026 and will target enterprise, cloud providers, and public sector clients aiming to implement large-scale agentic workloads.
SambaNova described the system as incorporating Xeon 6 as a host processor with the SN50 RDU for decoding and handling large AI models. The combination is intended to deliver higher throughput and lower latency for decoding, with Xeon 6 contributing memory bandwidth, PCIe lane density, and on-die acceleration.
Performance figures provided by the company indicate that Xeon 6 achieves over 50% faster LLVM compilation than ARM-based server CPUs, and up to 70% improved vector database performance compared with other x86 processors. SambaNova says this boosts the pace at which coding agents can be developed and moved to production.
Also in April, SambaNova signed a distribution deal with TEPCO Systems, a digital solutions subsidiary of Tokyo Electric Power Company (TEPCO). Under this agreement, TEPCO Systems will deploy SambaNova’s AI infrastructure across its own operations and distribute it to external enterprises in Japan.
SambaNova’s systems will support TEPCO’s forthcoming AI platform, intended for performance-focused and secure enterprise applications. The collaboration also aims to help Japanese organisations reduce energy use while running advanced AI workloads.
