Navigating GPU Export Controls and AI Use Restrictions in Data Center Operations

There have been a number of notable changes in the AI and data center regulatory landscape in 2026. As AI has dominated business decisions around the world, the race for AI leadership has reshaped the global geopolitical landscape. The broad consequences of AI have confounded — and at times, captivated — policymakers at all levels, and now, leaders in Washington, D.C. must grapple with harnessing AI’s capabilities while defending against malign use of AI to harm America’s strategic interests.

While there remains a general bipartisan interest in controlling exports of strategic technologies, especially regarding China, the Trump administration’s actions to date reflect the balance between national security controls on exports and the need to maintain a dominant global market share in AI technologies. Since its announcement that the January 2025 AI Diffusion Rule would not be enforced, the Trump administration has taken some actions to increase the accessibility and availability of graphics processing units (GPUs) to U.S. partners. Perhaps the earliest harbinger of this policy shift came in late May 2025, when the United States inked deals with the UAE and Saudi Arabia to allow for significant exports of advanced American semiconductors to each country. Given years of security concerns related to semiconductor exports to the region, the deals included strict security and reporting requirements for a select group of end users.

Two months later, the White House released America’s AI Action Plan, a comprehensive strategy that combined both offensive and defensive policies to win the global AI race. Alongside the AI Action Plan came an executive order that eventually led to the American AI Exports Program, which aimed to boost the export of the full American AI stack to global customers.

Then in December 2025, the Trump administration departed from previous policy expectations on semiconductor exports and technology security by approving exports of certain advanced U.S.-origin chips to customers in mainland China.

Amid reports that diverted and smuggled American semiconductors may be aiding China’s rapid AI advancements, broad bipartisan support has emerged in Congress for strict export controls on American AI technology. In 2026, the House Foreign Affairs Committee passed with overwhelming support measures that would codify export controls on advanced American semiconductors (AI Overwatch Act), direct BIS to require location tracking mechanisms for exported advanced semiconductors (Chip Security Act), and ensure that our allies match U.S. controls on semiconductor manufacturing equipment or face the imposition of extraterritorial controls on such technology (MATCH Act). Companion versions for each of these bills were included in the Senate’s annual defense authorization bill through a manager’s amendment.

This GT Advisory considers how data center operators that host GPU servers might navigate this evolving regulatory environment, providing an overview of current U.S. export and use restrictions and identifying compliance considerations for participants in this sector.

As detailed below, a strong foundation of export control compliance, supply chain relationships, and contract terms may help companies position themselves to absorb and adapt to potential changes in the U.S. regulatory regime.

Why GPU Hosting Creates Export Control Risks for Data Centers

Within data centers, GPUs have emerged as key components, transforming how complex computations are handled. GPUs are employed for their exceptional ability to perform parallel data processing, making them ideal for a range of tasks, including scientific computations, machine learning algorithms, and processing large-scale data. As demand for infrastructure capable of supporting AI model training and inference has grown, the ability to host GPU servers has become increasingly important for data centers.

The massive increase in processing power provided by GPUs as compared to central processing units (or CPUs) has, however, given rise to considerable disquiet amongst Western governments. In particular, the United States — where the biggest producers of GPUs are based — has expressed concern over their potential application for military and malign uses, and the Biden administration in January 2025 introduced comprehensive restrictions on the export and use of GPUs (the January 2025 AI Diffusion Rule). The Trump administration has also emphasized, as a policy imperative, the continuation (and even tightening) of these restrictions and has revoked the Biden-era restrictions and indicated that it would be replacing them with new restrictions, which, as of the date of this briefing, have not yet been issued. This regulatory uncertainty leaves industry in an interim phase questioning how best to manage current and possible future restrictions on GPU exports and use.

Historically, data center operators that merely hosted the GPU servers of their tenants (rather than exporting or providing GPU as a service) may have assumed these U.S. export controls were not a material compliance concern. That assumption, however, may no longer be appropriate. U.S. export controls apply to the items themselves — meaning that even non-U.S. operators may face liability under the EAR if restricted GPUs, controlled technology, or sanctioned end users are present in their facilities, even indirectly through tenants or sub-tenants. As regulators focus increasingly on the downstream use and custody of advanced computing hardware, data center operators must be prepared to demonstrate robust compliance measures and control frameworks. This includes knowing what GPUs are being hosted, who owns and accesses them, and for what purposes they are used.

U.S. Restrictions on GPU Exports and AI Model Use

Overview of U.S. Export Controls on GPUs

• S. export controls currently regulate the physical export of GPUs from the United States and re-export from third countries. The January 2025 AI Diffusion Rule, and possible replacement regulations, may also seek to control post-delivery use of the GPUs for certain AI models.
• Under the Export Administration Regulations (EAR) that BIS administers, some U.S. export controls are based on the export control classification number (ECCN) of particular goods, software, and technology (to country destination, to nationality of recipient, to nationality of temporary custodians or lessees).
• In addition, the EAR also has restrictions and license requirements that are based on the particular end user, such as end users identified on the BIS Entity List.
• Currently, most GPUs are classified under ECCNs 3A090 and 3A991, which carry varying degrees of export license requirements. Under the EAR, various license exceptions are available for GPUs exports. Related computers and equipment may be controlled under ECCNs 4A090 and 5A002. These require a fact specific, case-by-case analysis and may involve notifications or certifications to BIS.
• If a license is required and no exceptions are applicable, the U.S. exporter (or third country reexporter) may apply to BIS for a license. BIS’s review typically looks at the parties involved, and, in the case of GPU exports or other computing technology, the potential for the exported goods to be diverted to China and/or used for the development of advanced AI or military capabilities.
• It is a violation of the EAR (even for non-U.S. persons or entities) to proceed with a transaction knowing that U.S. export controls have been, or will be, violated. For example, it would be a violation for a data center operator to knowingly purchase GPUs that were exported without an appropriate license, even if the GPUs were already outside the United States. Thus, in procuring GPUs, data center operators may wish to conduct diligence on all parties in the chain of custody and to verify that appropriate export licenses have been sought and obtained.
• Liability and penalties under U.S. export controls and economic sanctions:

- Most export controls and economic sanctions are subject to strict liability, meaning it is irrelevant to a finding of violation whether the person knew or intended to violate the law. However, whether the specific offense includes a knowledge or intent element, enforcement agencies typically view intent or knowledge of a violation as an aggravating factor when assessing heightened penalties. In the most egregious cases, BIS and OFAC may make criminal referrals to the U.S. Department of Justice.

- Current potential penalties are civil monetary penalties for a U.S. economic sanctions or export controls violation, which can reach approximately USD $370,000 per violation or twice the value of each violation, whichever is greater. If U.S. sanctions or export control violations are prosecuted as criminal offenses, the criminal penalty can reach up to USD $1 million and for individuals, imprisonment for up to 20 years, though criminal referrals generally occur only in the most egregious cases (circumvention, evasion, willful pattern of conduct with high-risk technologies and end users). Additionally, egregious violations may result in denial of export privileges, or designation to the Entity List as a prohibited recipient of items subject to the EAR.

- The responsibility for seeking export licenses (or determining the applicability of exceptions) falls on the exporter of record from the United States (and/or reexporter from a third country). Recipients or intermediate consignees have a responsibility to prevent diversion of the items to unauthorized destinations or end users.

For operators, understanding who owns and uses the GPUs in their facilities, and taking reasonable steps to verify their compliance measures, may be a beneficial risk management component.

Overview of Restrictions on AI Model Use of GPUs

• The January 2025 AI Diffusion Rule placed new restrictions on the export of GPUs and related equipment (under ECCNs 3A090, 4A090, and 5A002), but also on certain services and the use of AI technology itself.
• Specifically, the January 2025 AI Diffusion Rule applied significant export restrictions to closed-weight AI models trained with 10^26 computational operations or more. Additionally, the rule created a new foreign direct product rule that applied these controls to certain model weights produced abroad using advanced computing chips made with U.S. technology or equipment. Accordingly, even AI models developed outside of the United States with controlled U.S. technology were subject to the EAR end-use controls (which, for example, prohibit exports for the use of certain weapons development).
• The rule established a framework of license exceptions and allocations that depended on the destination country of the above-referenced chips and services. It separated countries into three groups: 1) U.S.-allied countries eligible for broad license exceptions (specifically, Australia, Belgium, Canada, Denmark, Finland, France, Germany, Ireland, Italy, Japan, Netherlands, New Zealand, Norway, Republic of Korea, Spain, Sweden, Taiwan, United Kingdom, and the United States); 2) restricted, arms-embargoed countries that are subject to stringent licensing requirements with a presumption of denial (China, Macau, Afghanistan, Belarus, Burma, Cambodia, Central African Republic, DRC, Cuba, Cyprus, Eritrea, Haiti, Iran, Iraq, North Korea, Lebanon, Libya, Nicaragua, Russia, Somalia, South Sudan, Sudan, Syria, Venezuela, and Zimbabwe); and 3) all other countries, a broad group that includes many countries in Southeast Asia.
• Violations of the model weight export controls would carry the same penalties and risks as violative GPU exports noted above. In practice, compliance with technology export controls may be more difficult to monitor for, as technology like model weights may be easily transmitted across borders.
• The Trump administration rescinded the January 2025 AI Diffusion Rule in a press release and we await a replacement rule (which might be based on similar policy concerns and be more stringent in some regards).
• Given the strict restrictions on transfers to China (including in some cases, Chinese nationals or Chinese-controlled companies in third countries), any company that exports, reexports, receives, or utilizes advanced GPUs must be mindful of the EAR controls and diversion risks. For example, if a data center operator provided access and/or GPUs as a service to a Chinese military end user, that may run afoul of U.S. export controls.

For data center operators, this means that even if no physical export occurs, AI workloads hosted within their facility may still trigger U.S. export concerns — especially if tenants are training or deploying advanced models involving Chinese entities or nationals.

Overview of Proposed Changes to Export Controls and AI Use Framework

AI regulation and control is an area of bipartisan interest in Congress, and there are several pieces of pending legislation that, if enacted, might impact the compliance obligations and risks for companies in this sector. Some key pending proposals are summarized below, although we stress that these may change.

• R. 8170, MATCH Act: The bill would require the United States to strengthen and enforce export controls on semiconductor manufacturing equipment, components, and related services to designated countries of concern. It would mandate an annual review to identify critical technologies and facilities. The bill also seeks to coordinate these controls with allied countries and, if alignment is not achieved, permits the United States to apply extraterritorial restrictions on foreign-produced items that incorporate U.S. technology to address enforcement gaps and prevent circumvention.
• R. 6322, Stop Stealing our Chips Act: The bill would amend the Export Control Reform Act of 2018 to establish a whistleblower program that provides financial incentives for individuals who report export control violations, with a focus on advanced technologies. It would create a secure reporting system, include robust anti-retaliation and confidentiality protections, and establish a dedicated fund to support enforcement and pay awards.
• R. 6996, Full AI Stack Export Promotion Act: The bill would instruct the U.S. government to advance the global adoption of American artificial intelligence systems, hardware, and standards by lowering export barriers, supporting industry consortia, and implementing a diplomatic strategy to increase market access and shape international AI governance. It would mandate security protocols to prevent access by adversaries, establish ongoing reporting on global AI deployment, and introduce measures to strengthen U.S. leadership in technology and supply chain resilience.



Potential Future Regulatory Action

In July 2026, U.S. Commerce Department’s Bureau of Industry and Security (BIS) further relaxed export controls on the UAE for a range of items, including advanced semiconductors. Despite these permissive export controls on advanced semiconductors to select destinations and policy decisions that have allowed previously controlled AI technology to China, administration officials revealed that export control action on semiconductors and other AI technology is on the horizon.

In his testimony in front of the House Foreign Affairs Committee in mid-July 2026, BIS Under Secretary Jeffrey Kessler stated, “there will be future regulatory action in the area of chips and AI. It’s a quickly evolving area and the rules need to keep up with it.” Over the course of this second Trump administration, rumor and news of an AI Diffusion replacement has consistently surfaced, but not yet materialized. This latest on-the-record testimony from the BIS chief is the best indication yet that action is coming.

Considering Congress’ position on AI and export controls, as well as foreign chip deals this administration has made, companies in this sector may wish to prepare for regulatory action on semiconductors that may include:

• Export controls on semiconductors to known transshipment and smuggling hubs. Currently, there are no strict controls on the export of advanced semiconductors to known hubs for semiconductor diversion and smuggling, such as Southeast Asia. Consistent news of diversion, smuggling, and remote access to American chips by restricted end users has drawn the focus of industry, the administration, and Congress alike.
• Controls on remote access to advanced American semiconductors. Members of Congress have repeatedly cited remote access by restricted end users as a significant loophole in U.S. export controls. While legislation to grant remote access authority to BIS has been introduced in both chambers of Congress, BIS might implement remote access restrictions through licensing conditions. For example, to control remote access, BIS might implement a worldwide license requirement for advanced American semiconductors and then make available license exceptions for approved exporters or end users that agree to control remote access in select destinations.
• Permissive licensing for American companies and select end users abroad. The recent action on semiconductor exports to the UAE suggests a more permissive licensing regime for American companies and select end users. Similarly, the American AI Exports Program demonstrates that the administration still prioritizes exports of American semiconductors and AI technology to a broad—yet trusted—set of customers. Future action on semiconductor export controls may be more permissive for American and trusted end users abroad. For example, strict export controls for semiconductors to Southeast Asia may not apply to American companies’ local operations that agree to potential remote access restrictions.
• Clarified controls on semiconductor export controls for Chinese end users outside of mainland China. On May 31 of this year, BIS released guidance reiterating that a license is required to export advanced computing items to entities headquartered in Country Group D:5, which includes China, among other destinations. Future action on semiconductor export controls may include an official clarification entered into the Federal Register.



Considerations for Data Center Operators, Users, and Tenants

Some GPUs are export restricted and if exporters do not follow appropriate precautions, the GPUs may carry the “taint” of an EAR violation and create circumstances where future use and transfer of the GPUs might violate the EAR. Failure to perform due diligence or to manage restricted tenants might create liability exposure and reputational risk.

• Exporters, data center operators, and tenants should consider the identity of the owner/landlord or tenant who may have physical access to the equipment (including issues such as installation/repair/servicing, site visits/inspections, and/or government regulators/inspectors). Certain U.S. sanctions lists (including the SDN List and Entity List) include a “50%” rule that sanctions subsidiaries of sanctioned parties, so companies may wish to extend diligence reviews to ultimate beneficial owners/parent entities.
• Data center operators and investors should keep in mind that any involvement by an Entity List party — whether through ownership, access to controlled items, or service relationships — triggers heightened compliance obligations and may preclude the transaction entirely without prior BIS authorization. Companies may wish to maintain robust compliance protocols, including screening counterparties and subcontractors against the Entity List, and consider obtaining legal counsel when assessing risk or applying for BIS licenses. Other restricted party lists that may arise in the context of data center transactions (particularly those with any nexus to China) include: (1) the Chinese Military-Industrial Complex Companies (CMICs) list, which targets investments by U.S. persons in China’s military-industrial complex and its related military, intelligence, surveillance, security research and development programs; and (2) the Department of Defense list of entities owned or controlled by the Chinese military or communist party.
• Operators providing GPUs as a service should be aware that AI model training for malign uses is a significant concern for the U.S. government and may be part of the replacement to the AI Diffusion Rule.
• Investors should note that U.S. regulatory regimes impose a range of restrictions and review mechanisms relevant to cross-border transactions in this space. The U.S. government imposes restrictions on U.S. investor dealings in certain outbound investment transactions in the AI, semiconductor, and quantum computing sectors with parties in, or with a significant nexus to, China, Hong Kong, or Macau. On the flip side, non-U.S. investors looking at opportunities in the United States must be mindful of national security review under the Committee on Foreign Investment in the United States. Separately, the U.S. government also has broad authority to review cross-border transactions in information and communications technology or services involving U.S. entities or property subject to U.S. jurisdiction. Investors should consider all of these regimes when evaluating potential cross-border opportunities and partnerships.



Considerations for Data Center Operators Who Are Hosting Servers with GPUs

• Conduct detailed due diligence. The aspects of the due diligence would depend on the circumstances but might include:

- diligence on all parties in the supply chain and project team and include a review of ownership;

- certification from the exporters of the GPUs and intermediaries confirming their compliance with applicable export controls;

- confirmation from the tenant/exporter of the GPUs as to the applicable ECCN of the GPUs to be located at the data center facility, together with the details of the relevant export license (or the applicable license exception) for the GPUs; and

- review of the Electronic Export Information supplied by the U.S. exporter of the GPUs.

• Any gaps in a transaction or intermediaries who refuse to provide information may be viewed as a potential red flag. Note that a failure by intermediaries to obtain required information typically would not shield an operator from liability, particularly when there are inconsistencies in the supply chain or red flags that the operator fails to follow up on.
• A review of the physical and virtual security protocols of the data center (e.g., visitor logs and badges and authentication and ID requirements for contractors).
• While we are waiting for BIS to issue a replacement rule, key compliance elements derived from the AI Diffusion Rule may serve as a helpful commercial benchmark. For example, companies may wish to use the three-tier country groupings established in the January 2025 AI Diffusion Rule to conduct risk assessments or require additional contractual restrictions.
• Request warranties and indemnities in the lease/master services agreement that commit the tenant to compliance with U.S. laws (including laws and regulations related to GPUs), and include provisions for termination of the lease/master services agreement by the landlord in case of any breach of such laws (preferably with an acceleration of the rental payments for the remainder of the lease) or modification in the event of new laws being introduced.
• Seek and obtain certifications from customers that the services would not be used to develop advanced AI models or AI for use in the military or surveillance sectors (particularly in regions with strong connections to China). Include audit rights, as well as termination provisions for breach of these. Require periodic recertifications.
• Require evidence of customers’ compliance measures to prevent use of services for prohibited activities.
• Review end-user requests and press statements for signs or red flags they are using services in a manner inconsistent with certifications.
• Monitor regulatory changes closely and develop contractual provisions to account for changing requirements.

For operators, compliance with U.S. restrictions on export control and use of GPUs is no longer a back-office legal issue — it is a matter of managing key operational and reputational risk. Investors, lenders, and major tenants increasingly expect to see documented compliance frameworks covering GPU hosting and AI workloads. Proactively integrating these considerations into leases, MSAs, and onboarding procedures may help protect operators from inadvertent violations and position them as trusted partners in the AI supply chain.