NVIDIA Urgently Seeks Chinese AI Base Station Suppliers, Launches Edge Computing Networking in Coming Years

NVIDIA is urgently seeking AI base station suppliers in China to develop 6G AI-RAN base stations (base stations responsible for both communication connectivity and AI computing).

On August 4, a reporter from Jiemian News’ "Hard Tech Headlines" exclusively learned from core figures in the communications industry chain that NVIDIA is accelerating its entry into the telecom operator market and seeking base station manufacturers in China as partners to develop 6G base stations that meet overseas market requirements.

The source described NVIDIA’s search for the "next Zhongji Xuchuang" in edge computing, as NVIDIA believes that computing power will next need to be transmitted from data centers to massive numbers of AI terminals and users, ultimately relying on wireless transmission via base stations.

The source stated that NVIDIA is pushing this project with "considerable urgency," hoping to enter trial networks in 2027 or 2028.

Jiemian News reporters exclusively learned that Shenzhen Jiaxian Communication is one of the base station partners NVIDIA is currently establishing. An insider confirmed to Jiemian News that NVIDIA has been in contact with the company since 2025, and the two parties are currently cooperating to develop 6G AI-RAN base stations based on NVIDIA’s CUDA ecosystem. The collaboration has been ongoing for more than six months.

"Jiaxian Communication and NVIDIA have established a dedicated working group. More than ten people from NVIDIA are involved, with members distributed across Singapore, Hong Kong, and mainland China. Currently, the working group is conducting research around the open-source ecosystem of NVIDIA’s CUDA system," a relevant representative from Shenzhen Jiaxian told Jiemian News.

Founded in 2011 and headquartered in Shenzhen, Jiaxian Communication is a provider of AI-RAN wireless communication products and solutions. On August 4, Tongyu Communications announced the acquisition of approximately 25% equity in Jiaxian Communication for RMB 300 million.

Jiemian News reporters learned that the cooperation between Jiaxian Communication and NVIDIA has entered the technical R&D phase, and the establishment of a commercial entity for joint overseas expansion is not ruled out.

"First, at the technical level, we have already started cooperating." A representative from Jiaxian Communication stated that in the current collaboration, NVIDIA does not directly produce and sell complete base stations, but participates through chips and technology ecosystems. That is, base station companies purchase NVIDIA GPUs and develop 6G AI-RAN base stations based on the NVIDIA CUDA ecosystem, selling the finished products to overseas markets.

"The AI base station prototype developed with NVIDIA is expected to be ready by the end of this year. After verification, it will be launched to the market within one to two years," the source said.

For NVIDIA, AI base stations represent a business opportunity no less significant than computing power data centers.

In the AI 1.0 phase, computing power was mainly concentrated in data centers, where NVIDIA sold a large volume of GPUs. Entering the AI 2.0 phase of application implementation, the space for edge computing applications is broader. NVIDIA hopes its chips will also be adopted for edge and terminal computing power, so laying out AI base stations is NVIDIA’s preparation for the upcoming era of edge computing.

In other words, while chip companies can provide GPUs and model inference capabilities, deploying computing to the network edge and outputting results to terminals still requires base stations and operator networks to carry the load.

NVIDIA’s ambition in the communications sector is very urgent and evident.

On October 28, 2025, NVIDIA announced it would subscribe to new shares of Nokia for $1 billion, holding approximately 2.9% of the stock. As part of the partnership, Nokia will develop commercial AI-RAN products based on the NVIDIA platform and integrate them into its existing wireless access network product line, enabling operators to deploy AI-native 5G-Advanced and 6G-oriented networks.

In February this year, NVIDIA joined forces with Nokia, Ericsson, T-Mobile, Deutsche Telekom, SoftBank, and other enterprises to propose building 6G on an open, software-defined, and AI-native platform. In March, NVIDIA, T-Mobile, and Nokia announced progress in AI-RAN testing, attempting to enable base stations and mobile switching centers to run edge AI tasks while providing 5G connectivity.

Industry chain analysts believe that AI infrastructure will gradually extend from centralized computing centers to a distributed computing network composed of base stations, edge server rooms, and local nodes. Especially overseas, some markets lack the conditions to build large-scale centralized data centers, making distributed computing an important supplement.

Reporters learned that Jiaxian Communication plans to launch a 6G AI-RAN prototype based on the NVIDIA CUDA ecosystem by the end of this year. If the product enters large-scale deployment, the company’s total sales volume in the 6G era could reach the million-unit level. Calculated at approximately RMB 200,000 per unit, 1 million units correspond to a market size of about RMB 200 billion. Analysis firm Omdia predicts that by 2030, the cumulative market size for AI-RAN is expected to exceed $200 billion.

NVIDIA’s choice to seek AI base station suppliers in China also has practical considerations.

Like the optical module industry chain, the global base station industry chain is highly concentrated in China. Business research firm IndexBox estimates that in 2026, China will account for about 60% to 70% of global 5G base station hardware manufacturing, with main production clusters distributed in Shenzhen, Dongguan, Shanghai, and Chengdu.

While NVIDIA possesses chip and software ecosystems, this does not mean it can independently complete base station R&D. Base stations also involve radio frequency, microwave, antennas, AAU (Active Antenna Unit), whole-machine manufacturing, production testing, and operator network adaptation. These links heavily rely on engineering experience long accumulated in the communications industry.

Reporters learned from communications industry chain insiders that the base station industry is highly specialized. Once industrial clusters are formed, it is difficult for other regions to retrain relevant personnel. Meanwhile, training engineers in this industry takes a long time. Talent in the base station industry, especially hardware talent and production capacity, is mainly concentrated in China.

"Therefore, NVIDIA can only find manufacturers in China to produce the cheapest, most cost-effective, and best base stations," said the aforementioned industry chain insider. Even overseas communication equipment enterprises like Nokia and Ericsson have their production capacity located in China, with contract manufacturing participated in by companies like Luxshare Precision and Foxconn.

In his view, the base station industry has basically disappeared overseas, making it difficult to complete full production, assembly, testing, and hardware R&D abroad, while these capabilities remain concentrated in China. "Although this industry has a large market and high output value, it is deeply rooted in China. Other electronic industries might still find alternatives overseas, but the base station industry is hard to replicate."

A representative from Jiaxian Communication told reporters that since major domestic communication equipment enterprises like Huawei and ZTE have their own chip and technology ecosystems, and influenced by the China-US tech industry environment, NVIDIA has very limited options for independent base station partners in China.

Reporters learned from operator channels that because current 5G fees are relatively low, operators are also accelerating the popularization of AI base stations, aiming to increase prices in the future by charging "Token" fees.

In May 2026, China Telecom is set to launch Token packages, while China Mobile attempts to integrate communication traffic with large model Tokens; the centralized procurement scale of China Telecom Ningxia Company’s "Token Factory" is about RMB 17.4 billion. Operators are attempting to shift further from selling voice and traffic to selling model invocation and inference services. Core figures in the communications industry chain analyze that 6G base stations will restart a round of trillion-yuan-level investment in communications infrastructure.

As various large models trained in data centers gradually mature, AI’s edge computing infrastructure is rapidly being put on the agenda. As Cristiano Amon, President and CEO of Qualcomm, judged earlier this year: "Whoever captures edge AI is more likely to win the entire AI era."

Source: Jiemian News

Risk Warning and Disclaimer

The market carries risks; investment requires caution. This article does not constitute personal investment advice, nor does it take into account individual users' specific investment goals, financial status, or needs. Users should consider whether any opinions, views, or conclusions in this article align with their specific circumstances. Investment based on this content is at the user's own risk.