The U.S. government is preparing to establish a new framework that would control who gets access to the world's most powerful AI chips, fundamentally reshaping the global technology landscape. According to reported draft regulations unveiled in March 2026, the Commerce Department's Bureau of Industry and Security (BIS) is designing a tiered licensing system that would make American companies the gatekeepers for advanced artificial intelligence hardware worldwide. What Would These New Chip Export Controls Actually Look Like? The draft framework creates three distinct tiers based on the amount of computing power involved in any export. The system is designed to balance national security concerns with the practical realities of global AI development, though the exact restrictions remain unclear since the full draft regulations have not been made public. - Small Exports (Up to 1,000 GPUs): Shipments of up to 1,000 Nvidia GB300 graphics processing units (GPUs) or equivalent computing power would undergo what officials describe as "fairly simple review" with potential exemptions from full licensing requirements. - Medium Exports (1,000 to 200,000 GPUs): Larger exports intended for companies building extensive computing clusters would require pre-clearance and face a stricter export licensing regime with conditions attached, potentially including disclosure of business models or allowing U.S. government site visits to facilities. - Large Exports (Over 200,000 GPUs): Massive shipments exceeding 200,000 GB300s or equivalent would likely require involvement from the host country's government and would only be permitted for projects located in allied nations, with those countries required to make "matching" investments in American AI infrastructure. This tiered approach represents a significant shift in how the U.S. plans to manage AI chip distribution globally. Rather than outright bans, the framework attempts to create a system where the level of scrutiny increases with the scale of the export, allowing smaller transactions to move relatively quickly while subjecting massive cluster deployments to intensive review. Why Is the U.S. Taking This Step Now? The timing of these draft regulations reflects escalating tensions in the semiconductor industry and ongoing concerns about China's technological advancement. In January 2026, the Commerce Department changed its licensing policy for H200 chip exports from a presumption of denial to a case-by-case review approach, though applicants still needed to provide extensive certifications and documentation. Shortly after, Nvidia announced it was halting production of H200 chips intended for the Chinese market, signaling the mounting regulatory pressure on semiconductor exports. The administration also imposed a narrow 25% tariff on certain semiconductors and related equipment, including H200s, adding another layer of restrictions. These moves suggest the Trump administration is pursuing what would be its most comprehensive effort to date toward a global chip export control strategy. How Does This Compare to Previous U.S. Chip Control Policies? Interestingly, aspects of the reported draft regulations appear to mirror elements of the Biden administration's AI Diffusion Rule, which was issued in 2024 and created a global license requirement for advanced AI chips with tiered frameworks for different destinations. However, the Commerce Department has publicly denied that the new framework will resemble Biden-era policies, even as observers note structural similarities. In May 2025, the Trump administration rescinded the AI Diffusion Rule entirely and announced it would issue a replacement diffusion rule, though no timeline was provided. The new draft framework may represent that promised replacement, though officials have been coy about acknowledging continuities with the previous administration's approach. Steps Companies Should Take to Navigate the New Export Framework - Audit Your Supply Chains: Companies exporting or planning to export advanced AI chips should immediately review their current and projected shipments to determine which tier of the new framework would apply to their transactions and begin preparing necessary documentation. - Prepare Compliance Documentation: For medium and large exports, organizations should start compiling business model disclosures, facility information, and investment plans now, as the licensing process will require extensive certifications and detailed information about end-use applications. - Engage with Government Partners: Companies planning large exports to allied countries should begin conversations with both U.S. officials and host country governments about matching AI investment requirements and potential reciprocal arrangements that may be mandated under the new rules. - Monitor Policy Developments: Since the full regulations have not yet been released and questions remain about how the framework will address model weights and foreign AI investment requirements, companies should stay informed about official announcements from the Commerce Department's Bureau of Industry and Security. The uncertainty surrounding the exact implementation details creates challenges for companies in the semiconductor and AI sectors. Questions remain about how the framework will address model weights, the specific definition of "allied countries," and how reciprocal AI investment from foreign countries will be evaluated and enforced. What seems clear is that the reported new framework will not significantly change the administration's broader direction of limiting China's chip production and AI-related technological development. Whether through tariffs, licensing requirements, or investment mandates, the U.S. is consolidating its control over the global AI chip supply chain in ways that will reshape how companies worldwide access the computing power needed to build advanced AI systems.
