Nvidia Corp (NASDAQ: NVDA) stock is soaring after the company again confirmed its intention to dominate the Chinese chip market, albeit with increasingly severe export restrictions from the U.S. government.
Shares of the company are up 2.9% on Thursday and 9% on the week at the time of writing, as several reports unveiled a new series of AI chips that are being released to accommodate new restrictions announced by the federal government in mid-October.
The rumor was confirmed by several independent sources, including Chinastarmarket, a state-affiliated Chinese news outlet and The Financial Times, which was able to review documents distributed by Nvidia to prospective customers of the upcoming chips.
According to the Chinese source, Nvidia could share an official announcement on the new products by Thursday, Nov. 16, one week from now.
Nvidia's New Chips, Explained
Recently, Nvidia successfully capitalized on its strategic position to become a key provider of the infrastructure needed to train large language models like OpenAI's ChatGPT and Alphabet Inc.'s (NASDAQ: GOOG) (NASDAQ: GOOGL) Google Bard.
With product series such as the H100 or the A100 chips, Nvidia currently dominates the market for semiconductors needed to train AI systems, with as much as 80% of the total market share, according to some reports.
Shares in the company are up 234% since January, boosting Nvidia above the $1 trillion market cap for the first time in its market history.
Yet geopolitical tensions between the Chinese Communist Party and the U.S. government risk putting a ceiling on Nvidia's growth potential in China, the largest market for semiconductors in the globe.
The Biden administration has cataloged AI dominance as a crucial asset in current efforts to retain military, economic and commercial supremacy over China. The government released a number of programs aimed at boosting domestic chip capacity and slowing China's AI development.
China's domestic semiconductor industry is currently not advanced enough to produce some of the world's most powerful chips and the country is forced to rely on imports to access leading edge technology.
Taking advantage of China's limitations, the Biden administration launched a measure last year that restricts the export of top-of-the-line AI chips from U.S. companies to China.
As a result, Nvidia announced the launch of a series of chips, such as the A800 and the H800, made specifically to comply with U.S. sanctions, in an effort to retain market share in China.
But late last month, the government rolled out new limitations on exports citing national security concerns, which effectively bar the sale of A800 and H800 chips to Chinese companies, causing Nvidia shares to drop over 10% throughout the week starting Oct. 16.
Nvidia's latest plot twist is the new series of chips that would comply with the latest restrictions, called H20, L20 and L2. While the chips will be below the performance of Nvidia's flagship H100 as well as the China-tailored A800 and H800, they could still remain competitive within the limited Chinese market.
While investors have reacted positively to Nvidia's grit in the face of restrictions, these are starting to hurt the company's bottom line.
This week, Chinese internet giant Baidu Inc (NASDAQ: BIDU) dropped Nvidia as its preferred chip supplier in favor of Huawei, in a move attributed to increasing U.S. restrictions on chip exports.
A recent report estimated the company could lose up to $5 billion in the opportunity cost of not being able to provide China with cutting-edge chips.
While the U.S. is looking to curb the output of high-tech semiconductors to China, the government walks a fine line between imposing limitations and causing U.S. companies to completely lose access to the Chinese market.
On Thursday, shares of Nvidia competitors Micron Technology, Inc (NASDAQ: MU), Qualcomm (NASDAQ: QCOM), and Advanced Micro Devices, Inc. (NASDAQ: AMD) were up on news that the companies formed part of the China International Import Expo, of which Nvidia was absent for the fifth consecutive year.