The tougher US controls in recent years have prevented the California-based firm from selling certain AI chips -- widely regarded as the most advanced in the world -- to China.

As a result, it is now facing tougher competition from local players in the crucial market, including Huawei.

Nvidia boss Jensen Huang discussed plans to set up a research and development centre in Shanghai with its mayor during a visit to the city last month, the FT reported, citing two unnamed people familiar with the matter.

The site would "research the specific demands of Chinese customers and the complex technical requirements needed to satisfy Washington's curbs", said the report.

It added that "actual core design and production" would remain outside of China in order to comply with intellectual property transferral regulations.

Nvidia did not immediately respond to an AFP request for comment, nor did Shanghai authorities.

During a visit to Beijing in April, Huang met with Vice Premier He Lifeng, telling him that he "looked favourably upon the potential of the Chinese economy", according to state news agency Xinhua.

Huang said he was "willing to continue to plough deeply into the Chinese market and play a positive role in promoting US-China trade cooperation", Xinhua said.

The tightened US export curbs come as China's economy wavers, with domestic consumers reluctant to spend and a prolonged property sector crisis weighing on growth.

President Xi Jinping has called for the country to become more self-reliant as uncertainty in the external environment increases.

Xi said last month that China should "strengthen basic research, focusing our efforts on overcoming challenges in key technologies such as advanced chips and core software, and building an autonomous AI system", according to Xinhua.

Washington has expanded its efforts in recent years to curb exports of state-of-the-art chips to China, concerned that these can be used to advance Beijing's military systems and otherwise undermine US dominance in artificial intelligence.

pfc/reb/sco

Nvidia