Nvidia and AMD, two leading semiconductor companies, are facing new restrictions in some Middle Eastern countries, according to a report by Reuters. This development expands the limitations on their chip exports beyond China and Russia. Nvidia has stated that the licensing requirement now applies to “a subset” of its high-end chips.
In response to these restrictions, the Biden administration has clarified that it has not blocked chip sales to the Middle East, as reported by Reuters. However, the Commerce Department has declined to comment on whether it has imposed new requirements on specific U.S. companies. As of now, the Commerce Department, Nvidia, and AMD have not provided any official comments on the matter.
The scope of licensing requirements has also been extended to other products, including networking products, intended for specific end users and end uses in China. In the second quarter of fiscal year 2024, the U.S. government informed Nvidia of an additional licensing requirement for certain A100 and H100 products destined for selected customers and other regions, including certain countries in the Middle East, as stated in a regulatory filing.
Although Nvidia claimed that it does not anticipate the constraints to have an immediate material impact, AMD also received a letter with similar restrictions. The Reuters report did not disclose the specific countries affected, nor did it explain the reasons behind these limitations in Nvidia’s regulatory filing.
Notably, Nvidia’s A100 and H100 chips have already been subject to export restrictions imposed by the Biden administration, primarily due to concerns over their utilization in military applications in China. The U.S. government previously informed Nvidia of its intentions to mitigate the risk of these chips being employed in military end uses or by military end users in China and Russia. Nvidia’s recent filing highlights the potential harm to results and competitive positioning if further changes occur in export controls, potentially leading to exclusion from all or part of the Chinese market.
The H100 chips, in particular, hold significant value, to the extent that startups can use them as collateral for loans. However, availability of H100 chips is predicted to become extremely limited in the foreseeable future. Industry sources, including AI labs, cloud providers, and small startups, presume that obtaining H100 chips in the first half of next year will be nearly impossible. The lead time for new orders, even if they can be secured, is estimated to be approximately six months, which is a considerable duration within the rapidly evolving AI sector.
In conclusion, Nvidia and AMD find themselves confronting new restrictions on chip exports in several Middle Eastern countries. These limitations have extended beyond China and Russia, posing potential challenges to their operations and market access. The implications of these restrictions, along with the already existing export controls, create uncertainties for both companies in terms of their competitive positioning and ability to meet market demands. Furthermore, the scarcity of H100 chips, which serve as valuable collateral, introduces additional obstacles for startups and companies relying on these high-performance chips for their AI applications. The evolving landscape of export controls and licensing requirements in the semiconductor industry continues to shape the strategic decisions and prospects of companies like Nvidia and AMD.