Jensen Huang, co-founder and CEO of Nvidia, during the Computex conference in Taipei, Taiwan, on May 20, 2025.
Annabelle Chih | Bloomberg | Getty Images
Nvidia shares jumped on Thursday after posting a positive set of earnings, sparking a rally in global semiconductor stocks.
Shares of Nvidia closed more than 3% higher after the company posted better-than-expected earnings and revenue on Wednesday, even as it took a hit from U.S. semiconductor export restrictions to China.
Nvidia has been seen by investors as a bellwether for the broader semiconductor industry and artificial intelligence-related stocks, with its latest strong numbers sparking a rally among global semiconductor names.
Taiwan Semiconductor, AMD and Qualcomm posted slight gains on the day as tariff uncertainty weighed on the broader market.
In Japan, Tokyo Electron closed more than 4% higher, while SK Hynix, which is a supplier of high bandwidth memory to Nvidia, was nearly 2% up at the close of markets in South Korea.
In Europe, ASM International, BE Semiconductor Industries and ASML closed largely unchanged, pulling back from earlier gains.
The semiconductor industry has faced a number of headwinds from uncertainty around tariff policy in the U.S. and chip export restrictions to China.
Companies such as ASML, which makes machines that are critical for manufacturing the most advanced chips, have seen billions wiped off their value as a result.
Nvidia on Wednesday said it wrote off $4.5 billion of H20 chip inventory that it couldn’t ship to China because of export curbs, saying it also calculated $2.5 billion of lost revenue as well.
The restrictions on China do not seem to be going away.
The U.S. has ordered a number of companies, including those producing chemicals and design software for semiconductors, to stop shipping goods to China without a license, according to a Reuters report on Thursday.
Despite this, Nvidia still managed to post financial results for the April quarter that beat market expectations, allaying fears that demand for its graphics processing units, which have become key for training huge AI models, is dwindling.
