Nvidia Seeks AI Base Station Partners in China, Shenzhen Jiaxian Communication Already On Board
Taylor Wilson
Nvidia is co-developing a 6G AI-RAN base station with Shenzhen-based Jiaxian Communications, targeting a prototype by year-end and trial-network deployment by 2027-2028. This means → Nvidia wants to sell compute beyond the data center, and China's manufacturing base is the path it cannot bypass.
What are Nvidia and Jiaxian building together?
The two companies are co-developing a 6G AI-RAN base station on Nvidia's CUDA ecosystem — Nvidia's GPU software-development platform. The partnership has run for over six months, with a dedicated team of more than ten Nvidia staff across Singapore, Hong Kong, and mainland China.
An AI-RAN base station — a next-generation telecom unit with AI compute built in — is expected to have a working prototype by year-end, with commercial availability one to two years after validation.
The division of labor: Nvidia supplies chips and the technology ecosystem; Jiaxian buys Nvidia GPUs, builds the complete base station, then sells the finished product to overseas markets.
Why does Nvidia want to be in the base-station business?
The core logic is edge-side compute. In the AI 1.0 phase, compute sat in data centers and Nvidia sold GPUs to cloud providers. In the 2.0 phase — AI applications going live — compute needs to reach the network edge and the end user.
This means → between the data center and the device, traffic still flows through base stations and carrier networks. The base station becomes Nvidia's next chip-shipment entry point.
Nvidia has already moved: it subscribed to $1 billion in Nokia shares (roughly 2.9% stake), and joined Nokia, Ericsson, T-Mobile, Deutsche Telekom, and SoftBank in proposing an AI-native, software-defined 6G platform.
How big is this market?
Research firm Omdia forecasts the AI-RAN market will exceed $200 billion cumulatively by 2030.
Jiaxian itself estimates that, at scale, its 6G-era shipments could reach millions of units; at roughly ¥200,000 per unit, that implies a market of about ¥200 billion.
In plain terms = two "200 billion" figures — one in global dollars, one in Jiaxian's own renminbi terms — signal that the base station is shifting from a traditional telecom box to a new carrier for AI compute.
Why must Nvidia look for partners specifically in China?
China accounts for roughly 60–70% of global 5G base-station hardware production (IndexBox, 2026 estimate), clustered in Shenzhen, Dongguan, Shanghai, and Chengdu.
Base stations involve RF, microwave, antenna, system assembly, testing, and carrier-network adaptation — all deeply dependent on engineering experience built over decades. Talent and production capacity sit mainly in China.
This reflects a hard reality: even Nokia and Ericsson manufacture in China, through contract producers such as Luxshare Precision and Foxconn. In plain terms = base-station manufacturing has taken root in China so deeply that, unlike some other electronics, it is nearly impossible to replicate elsewhere.
Why is Jiaxian a "scarce target"?
Huawei, ZTE, and other major Chinese telecom-equipment makers run their own chip and technology ecosystems. Combined with the U.S.–China tech environment, Nvidia's pool of independent base-station partners in China is extremely small.
This means → Jiaxian, operating outside the Huawei and ZTE orbits, is one of the few partners Nvidia can work with in the country.
On August 4, Tongyu Communication announced it would acquire roughly 25% of Jiaxian for ¥300 million — the capital market is already pricing in that scarcity.
What is changing on the carrier side that supports this?
China Telecom launched a Token-based plan in May 2026; China Mobile is experimenting with blending data traffic and large-model tokens into a single offering.
In plain terms = carriers are shifting from selling voice and data to selling model calls and inference services — which requires base stations with built-in AI compute, exactly what AI-RAN delivers.
Industry sources estimate that 6G base stations will trigger a fresh round of telecom-infrastructure investment on the order of trillions of renminbi. Whether Nvidia can find enough qualified partners in China is the key variable for turning this strategy into reality.
Content is for reference only, not financial advice.