According to a report from MoneyDJ, citing ITHome, Chinese AI chipmaker Cambricon announced on the 17th that it has revised its A-share private issuance plan for 2025. Under the updated proposal, the company plans to issue no more than 20.9175 million shares, aiming to raise up to RMB 3.9853273 billion. Cambricon stated that, after deducting issuance expenses, the funds will primarily be used for chip projects tailored to large AI models and chip platforms designed for large-model applications, as the report highlights.
Cambricon states that the fundraising project is intended to meet the growing demands of large AI models. The company plans to develop a software platform specifically for large-model workloads. This platform will include three core modules: a flexible compilation system, a training platform, and an inference platform, all designed to improve the usability and adaptability of its chips, as the report notes.
The company says its AI chips support training and inference for large models such as the DeepSeek series, LLaMA series, GPT series, BLOOM series, and GLM series, as well as applications in computer vision, speech processing, and natural language processing, according to the report.
Founded in 2016 and listed on the Shanghai Stock Exchange’s STAR Market since July 2020, Cambricon is considered one of China’s leading players in domestic GPU development, especially as the U.S. tightens export controls on advanced AI products from firms like NVIDIA, as noted by South China Morning Post. The report adds that Cambricon specializes in 7-nanometer AI chips, with its Cambricon-1A used in products from companies such as Huawei Technologies.
In April 2025, Cambricon reported a net profit of RMB 355 million (USD 49.2 million) for the quarter ending March 31, reversing a net loss of RMB 227 million a year earlier. It was the company’s second straight profitable quarter after a full-year loss, driven largely by rising adoption of its AI products, as noted by Chinese media outlet Yicai Global.
Read more
(Photo credit: Cambricon)
Please note that this article cites information from MoneyDJ, ITHome, Cambricon, South China Morning Post, and Yicai Global.
