A dedicated joint working group has been formed, with roughly 10 or more Nvidia personnel spanning Singapore, Hong Kong, and mainland China . The technical work is based on Nvidia's CUDA ecosystem and open-source CUDA systems
.
The core concept is AI-RAN: base stations that are not just communication pipes but also perform AI inference and training at the network edge . Nvidia sees an opportunity to supply AI-enabled RAN gear to global telecom operators as they transition from 5G to 6G
.
Nvidia is reportedly pushing this project on an "urgent" timeline . The stated goal is to enter trial networks in 2027 or 2028
. A Jiaxian insider told Jiemian News that they expect an AI base station prototype by the end of 2026, with a market launch within one to two years after verification
.
Reports indicate that the two companies are exploring the possibility of forming a commercial entity for joint overseas expansion . However, no formal joint venture has been announced.
In a notable development, Tongyu Communications — a company that proposed an equity investment in Jiaxian — issued a clarification on August 6, 2026, stating that Jiaxian has no R&D collaboration agreement with Nvidia at the corporate level and no investment from Nvidia . This suggests that as of that date, the collaboration was operational and technical rather than a formal equity partnership. The reports from Jiemian News, however, indicate direct technical cooperation on the CUDA ecosystem
.
The base stations are explicitly being developed to meet overseas market requirements, not for deployment inside China . Nvidia reportedly does not intend to manufacture or sell complete base stations itself; instead, it participates through its chips and technology ecosystem: base-station companies purchase Nvidia GPUs and develop 6G AI-RAN base stations on the CUDA platform for sale to overseas markets
.
The Jiaxian partnership is one track in a much larger, parallel strategy that Nvidia has been executing since at least 2025:
Nvidia's overarching goal is to make its CUDA ecosystem and GPU-accelerated AI-RAN platform (branded as AI Aerial) the foundational compute layer for all 6G networks worldwide .
Together, these parallel efforts reflect Nvidia's view that "compute is moving from the data center to the edge" — and base stations are the critical last-mile infrastructure for AI delivery to billions of devices .