The US government has broadened its restrictions on the sales of Nvidia's AI chips, extending beyond China to include certain Middle Eastern countries. The affected chips are Nvidia's H100 and A100 models, which are in high demand due to their capabilities in developing and running generative AI programs, such as OpenAI's ChatGPT.
In a statement to Reuters, Nvidia said, “During the second quarter of fiscal year 2024, the USG (U.S. government) informed us of an additional licensing requirement for a subset of A100 and H100 products destined to certain customers and other regions, including some countries in the Middle East.”
The expansion of restrictions on Nvidia's AI chip sales underscores the US government's ongoing concerns about the potential misuse of advanced technology. While the immediate financial impact on Nvidia and other companies like Advanced Micro Devices (AMD) may not be significant, the broader implications for global trade and technological development remain to be seen.
Reasons Behind the Restrictions
The US restriction expansion of exports of Nvidia and AMD AI chips beyond China to other regions, including some Middle Eastern countries comes amid concerns from the Biden administration about the potential use of these chips by the Chinese government to further its AI ambitions. Previously, the White House had imposed export restrictions to prevent the chips from being accessed by Chinese companies and the military. The new restrictions might be aimed at preventing a Chinese office located in the Middle East from acquiring Nvidia chips and subsequently transferring the technology to China.
Nvidia mentioned in a regulatory filing that the curbs on its A100 and H100 chips, designed for machine-learning tasks, wouldn't have a significant immediate impact on its results. AMD also received a letter with similar restrictions, which reportedly won't materially affect its revenue. The US typically imposes export controls for national security reasons. Last year, such a move indicated an intensification of the US's efforts to limit China's technological capabilities. However, the risks posed by exports to the Middle East remain unclear. Nvidia did not specify which Middle Eastern countries were affected by the restrictions.
Impact on Nvidia and Other Companies
While the restrictions target Nvidia's H100 and A100 GPUs, the company has stated in an SEC filing that these curbs won't have an immediate significant impact on its financial results.
The H100 AI GPU is the latest and most powerful data center GPU from Nvidia, based on the new Hopper architecture. It is designed to handle large-scale AI and HPC workloads, such as deep learning, natural language processing, and recommender systems. The Nvidia A100 AI GPU is the previous generation of data center GPU from Nvidia, based on the Ampere architecture. It is also optimized for AI and HPC applications, such as image recognition, speech recognition, and genomics.
Nvidia has expressed concerns about the potential implications of shifting export controls, which might prompt customers outside China and other affected regions to exclude US semiconductors from their products, thereby reducing compliance burdens and risks. The company has already faced sales restrictions for these chips in China and Russia, leading the company to develop alternative models like the H800 and A800 with reduced performance for the Chinese market.