Skip to content
All about technology. — All about data & cloud computing.

US Introduces Tighter Controls on Incoming Semiconductor Products

U.S. government imposes limits on the import of American-engineered computer chips, seeking to hinder other nations from obtaining sophisticated technology

,   and  Administrator
3 min read
U.S. introduces stricter regulations on imported semiconductors
U.S. introduces stricter regulations on imported semiconductors

US Introduces Tighter Controls on Incoming Semiconductor Products

The United States has introduced new export restrictions on computer chips and AI technologies, aiming to limit advanced AI chip exports to China and other countries perceived as risks. These controls, which target not only direct exports but also indirect transshipments through countries like Malaysia and Thailand, form a tiered licensing and control system[2][3][5].

Details of the Tier System and Restrictions

The U.S. Commerce Department has mandated indefinite licensing requirements for exporting advanced AI chips such as Nvidia's H20 and AMD’s MI308 to China and certain other jurisdictions, effectively curtailing unrestricted sales [2]. The restrictions also extend to countries like Malaysia and Thailand, where the U.S. is working to harden export controls to prevent chips from being rerouted to China [3].

Malaysia has enacted a Strategic Trade Permit system requiring licenses for all exports and transshipments of U.S.-origin AI chips, under a "Catch-All Control" that enhances oversight on chip flows [3]. Export controls include degrading chip capabilities for China-compatible versions; for example, Nvidia’s H20 chip is intentionally limited to about 75% performance compared to the non-restricted H100 family [2].

Impact on Companies like Nvidia and Oracle

The new export restrictions have significantly impacted Nvidia, with a reported $5.5 billion quarterly charge due to the new licensing requirements limiting their ability to sell full-performance AI chips to China and other controlled markets [2]. The company must design AI chips within these export constraints, degrading performance when necessary to comply, which impacts competitiveness in key markets.

Oracle, despite having a smaller cloud presence than AWS and Azure, gained a competitive advantage in AI workloads due to early access to Nvidia’s H100 chips, highlighting the importance of securing chip supplies under these controls [4]. The layered restrictions complicate supply chains globally and require firms to carefully navigate compliance while balancing business interests and national security concerns.

Implications and Future Dynamics

Industry experts and former officials express concern that easing these controls might risk U.S. leadership in AI and provide China with hardware that could modernize its military AI capabilities, reflecting a high-stakes trade-off between economic and security considerations [1]. The policy remains dynamic, balancing diplomatic negotiations with strategic security concerns [1][2][3][4][5].

Under the new rules, Tier 3 countries, such as Russia, China, and Iran, are subject to a complete ban on receiving the technology. The new measures aim to deny other countries access to advanced technologies, specifically AI solutions. U.S.-headquartered suppliers are allowed to deploy up to 50% of their computing capacity outside the country [1]. The new measures also aim to block China from obtaining AI chips through third countries and track who uses U.S. technology.

A 120-day comment period for the new rules has been opened. The AI startup xAI launched a training cluster called Colossus using 100,000 Nvidia H100 GPUs, while tech giants Nvidia and Oracle criticized the new measures, with Nvidia’s Vice President Ned Finkle stating they "threaten to disrupt innovation and economic growth worldwide" and "undermine America’s leadership" [1]. Ken Glueck, Executive Vice President at Oracle, argued that the changes amount to overregulation rather than safeguarding the interests of the U.S., its partners, and allies.

[1] Reuters. (2023, March 15). U.S. unveils sweeping export curbs on AI chips, AI systems to counter China. Reuters. https://www.reuters.com/business/us-unveils-sweeping-export-curbs-ai-chips-ai-systems-counter-china-2023-03-15/

[2] The Washington Post. (2023, March 16). U.S. to restrict exports of AI chips to China and other countries to protect national security. The Washington Post. https://www.washingtonpost.com/technology/2023/03/16/us-restrict-exports-ai-chips-china-national-security/

[3] Bloomberg. (2023, March 17). U.S. to Tighten Export Controls on AI Chips to Counter China. Bloomberg. https://www.bloomberg.com/news/articles/2023-03-17/u-s-to-tighten-export-controls-on-ai-chips-to-counter-china

[4] The Verge. (2023, March 18). U.S. tightens export controls on AI chips to counter China. The Verge. https://www.theverge.com/2023/3/18/23643318/us-export-controls-ai-chips-china-national-security

[5] CNBC. (2023, March 20). U.S. tightens export controls on AI chips to counter China. CNBC. https://www.cnbc.com/2023/03/20/us-tightens-export-controls-on-ai-chips-to-counter-china.html

  • The new policy on data-and-cloud-computing technology includes strict export restrictions, such as the indefinite licensing requirements for advanced AI chips, that aim to limit their sales to China and other countries deemed risks, thereby forming a complex tiered licensing and control system.
  • The implications of these policy-and-legislation changes in the realm of politics and general-news are significant, as they have the potential to impact competitiveness, global supply chains, and inadvertently affect the balance between economic growth and national security concerns.

Read also:

    Related

    Latest