Odaily, citing Reuters, reported that Nvidia has started pitching Vera, its first standalone central processing unit, to customers in China. People familiar with the matter said the chip is designed for Agentic AI systems and has already entered mass production. The move marks Nvidia’s attempt to use a CPU product, rather than only its better-known GPU lineup, to further expand its presence in the Chinese market.
Chinese Cloud Customer Plans Dual-Vera Server Tests
According to the people cited in the report, some Chinese customers have shown interest in Vera. One large Chinese cloud computing company plans to purchase more than 300 servers equipped with dual Vera CPUs for testing. After the testing process is completed, the company will decide whether to increase the size of its procurement. The report did not name the cloud computing company and did not provide a detailed timetable for the test deployment.
Vera is built on an Arm Holdings architecture and is Nvidia’s first standalone CPU product. Nvidia previously said Vera can deliver 1.8 times the performance of competing products in computing tasks related to AI agents. The company also expects the product to contribute about $20 billion in revenue by the end of the current fiscal year, which runs through the end of next January.
Inference Shift Brings CPUs and Custom Chips Into Focus
The report noted that as the AI industry’s center of activity moves gradually from model training toward inference computing, CPUs and custom chips are receiving more attention. Vera also places Nvidia in direct competition with Intel and Advanced Micro Devices, or AMD, in the server CPU market. Intel and AMD have long dominated that segment, while Nvidia has been more closely associated with GPU products used in AI computing.
People familiar with the matter said that because the United States maintains strict export restrictions on high-end GPUs, CPUs face relatively fewer regulatory obstacles in the Chinese market than GPU products. At present, some Chinese customers plan to test Vera chips first in overseas data centers. At the same time, software ecosystem compatibility and existing domestic AI chip deployment systems remain factors that can affect broader adoption of Vera in the future.