SiFive to Integrate Nvidia NVLink Technology in RISC-V Chip Designs
GeokHub
San Francisco | Jan 15 — GeokHub Semiconductor design firm SiFive said on Thursday it will integrate Nvidia’s NVLink technology into its RISC-V chip designs, marking a first for the open-standard processor architecture and strengthening its position in the fast-growing AI data center market.
The move will allow customers using SiFive’s RISC-V blueprints to connect central processing units (CPUs) directly to Nvidia’s advanced artificial intelligence chips using the same high-speed interconnects relied on by Intel- and Arm-based systems.
RISC-V Gains Ground in AI Infrastructure
SiFive, like Arm Holdings, does not manufacture chips itself. Instead, it supplies design blueprints that customers use to build custom processors. RISC-V, an open-standard alternative to Arm, has gained increasing attention from major technology companies including Google and Meta, as concerns grow over Arm’s ambitions to become a full chip designer rather than a neutral licensor.
By adopting Nvidia’s NVLink — a technology critical for linking thousands of chips together inside AI data centers — SiFive aims to make RISC-V a more viable option for large-scale artificial intelligence workloads.
High-Speed Connectivity Is Critical for AI
NVLink enables ultra-fast communication between CPUs and Nvidia’s AI accelerators, allowing massive volumes of data to move efficiently across interconnected chips. Such capability is essential in modern AI systems, where performance depends on seamless coordination across entire clusters of processors.
SiFive CEO Patrick Little said chip designs incorporating NVLink are unlikely to reach the market until 2027 or later, but stressed the partnership represents a long-term commitment.
“This is a multi-generational effort,” Little said, describing the collaboration as an ongoing partnership to deliver NVLink compatibility across future chip generations.
Strategic Implications
Once available, SiFive’s NVLink-enabled RISC-V designs would allow customers to pair open-standard CPUs with Nvidia’s leading AI chips without sacrificing interconnect performance — a capability previously limited to Intel- or Arm-based architectures.
Financial details of the agreement were not disclosed.
Impact & Outlook
The partnership highlights Nvidia’s growing influence across the semiconductor ecosystem and signals rising confidence in RISC-V as a serious contender in AI infrastructure. As data centers expand and demand for flexible chip architectures grows, open-standard designs paired with high-performance interconnects could play a larger role in future AI systems.
