Nvidia is collaborating with Samsung to develop custom non-x86 CPUs and XPUs aimed at countering competition in the AI sector. This partnership focuses on embedding custom chips to strengthen Nvidia's position in global data centers. The custom chips under NVLink Fusion will remain exclusive to Nvidia products.
Nvidia has announced a partnership with Samsung to build custom non-x86 CPU and XPU technologies. The initiative is designed to stave off competition from key players including OpenAI, Google, AWS, Broadcom, Meta, and other partners.
This collaboration aims to embed custom chips within Nvidia's ecosystem, enhancing its dominance in next-generation data center landscapes globally. According to the partnership details, these custom chips operate under NVLink Fusion and are tied exclusively to Nvidia products, ensuring integration remains within Nvidia's control.
The move strengthens Nvidia's AI empire by leveraging Samsung's expertise in chip manufacturing to create tailored solutions for high-performance computing needs. No specific timelines or additional technical details were provided in the announcement, but it underscores the intensifying race among tech giants to innovate in AI hardware.