BIS Lays the Groundwork for Global and Metered Access to Frontier AI Models and the Computing Power to Train Them
Client Alert | January 30, 2025
While the new regulations were issued during the closing days of the Biden Administration, they are the product of a broad interagency process and of policy drivers that will continue to motivate Trump Administration officials. Companies should take steps now to evaluate the impact of the regulations on their plans for AI model training and deployment, and to develop and implement the procedures that will be required to win export licenses or to qualify for licensing exceptions.
On January 13, 2025, the U.S. Department of Commerce’s Bureau of Industry and Security (BIS) issued an interim final rule titled “Framework for Artificial Intelligence Diffusion” (the Framework)[1] that lays the groundwork for expansive new controls targeting frontier Artificial Intelligence (AI) models themselves and the computing power to create them. The Framework, taken together with recent U.S. Government actions targeting AI, seeks to use the current leading role played by companies based in the U.S. and a select group of allied countries in the design and production of computing power as a point of leverage to force companies, research institutes and other organizations develop AI models inside an ecosystem closely monitored by the United States and the handful of other countries that have agreed to impose similar controls. In several ways, the Framework bookends a multiyear effort by Commerce officials and their interagency peers to control access to, and impede the indigenous development of, computing power by China, and other countries perceived to pose a threat to U.S. national security and foreign policy interests.
To this end, the Framework, through a multi-part control structure, aims to reduce the risk that “countries of concern” (i.e., countries listed in Country Group D:5, which includes Hong Kong now treated by export regulations as part of China, and Macau) obtain advanced U.S. and allied closed-weight AI models by broadly (i) expanding licensing requirements for the export of advanced integrated circuits (ICs) (ii) imposing controls on frontier AI models, and (iii) closing a significant loophole that previously allowed persons in countries of concern rent access to computing power outside their countries. At the same time, the Framework creates a three-tiered licensing policy with a more permissive structure allowing exports to and among countries whose export controls are aligned with the U.S., imposing an effective embargo against countries the U.S. perceives as threats, and detailing a conditional policy for countries yet to adopt certain safeguards against the unchecked development of frontier AI models. Thus, the Framework seeks to strike a balance between the goal of keeping advanced AI capabilities out of the hands of strategic competitors while facilitating the diffusion of AI technology and its benefits within a U.S.-structured AI ecosystem.
BIS uses several familiar tools to fashion the new Framework, including a new foreign direct product rule that could reach AI frontier models globally and a powerful new nationality-based license exception that conditions the powerful authorization it provides on building certain amounts of computing power within the countries whose companies are eligible to use it. While the Trump Administration has issued an executive order[2] that authorizes the Department of Commerce to postpone the implementation of the Framework, the Framework’s export controls may meet the criteria set in President Trump’s America First Trade Policy memoranda of deploying export controls that help the United States to “maintain, obtain, and enhance” the United States’s “technological edge” and to “identify and eliminate loopholes.” Because of this, and because the Framework is the product of a broad and sustained interagency efforts focused addressing geopolitical threats that have not changed with the new Administration, we expect that enough of the Framework will be implemented that companies in the AI model development, advanced IC manufacturing and distribution, and data center sectors should plan now for its implementation.
I. Background (Prior Actions)
The Framework in many ways bookends a series of measures aimed at targeting AI development capabilities of China and others. In 2022, the Biden Administration’s National Security Strategy identified China as “the only competitor with both the intent to reshape the international order and increasingly, the economic, diplomatic, military, and technological power to advance that objective” and specifically identified export controls as a key tool to “ensure strategic competitors cannot exploit foundational American and allied technologies, know-how, or data to undermine American and allied security.” In October 2022, BIS put in play a “chokepoint” strategy to target indigenous Chinese semiconductor development. It identified, broadly, within the semiconductor ecosystem, four chokepoints where the U.S. and its allies maintained significant technological advantage and crafted export controls around them: (i) ICs, (ii) semiconductor manufacturing equipment (SME), (iii) SME parts and components, and (iv) design and other software for ICs and SME. A leading rationale for imposing these controls was to address China’s use of AI for military modernization as well as surveillance. In October 2023, BIS updated the October 2022 controls, again specifically noting China’s use of “advanced computing ICs and supercomputing capacity in the development and deployment of [] AI models to further its goal of surpassing the military capabilities of the United States and its allies.” Over the course of 2024, BIS clarified the scope of AI and SME controls and, expanded Authorization Validated End User (VEU) to enable data centers to receive VEU authorizations in order to facilitate the responsible diffusion of advanced AI technology. In November 2024, the Biden Administration issued a National Security Memorandum on AI that called for, among other, ensuring the “safety, security, and trustworthiness of American AI innovation writ large.” And, as BIS notes in the Supplemental Information to the Framework, the Department of Commerce has engaged in an extensive and ongoing policy process with partners across the U.S. Government to consider strategic, tailored, and effective controls on the diffusion of advanced AI technology to entities and destinations around the world.
II. The Framework: New Controls and Jurisdiction-Based Rules
a. “Chokepoint” Strategy for AI Development
The Framework attempts to control access to three elements critical for AI model training:
-
- Advanced ICs: Training advanced AI models requires large clusters of advanced computing ICs capable of handling large quantities of data and models containing large numbers of parameters. The Framework expands current restrictions and imposes a global export licensing requirement for advanced ICs.
- Compute Power: These advanced IC clusters are housed within data centers, which provide processing power to run AI applications, including training and inference applications. The Framework creates multi-year quotas that meter access to computing power in most countries. It also creates a revamped validated end user (VEU) authorization system for data centers which is premised on nationality of those seeking to procure computing power and data center country location and which conditions VEU authorization both on ensuring that the computing power required to train frontier AI models remains in installed in only a handful of countries and on the adoption of significant physical and cyber security controls.
- Model Weights: Model weights are numerical parameters that define the internal logic of an AI model and which are the product of model design and training. The Framework creates a new Export Control Classification Number (ECCN) 4E091 for certain advanced closed-weight AI models and imposes a global licensing requirement for such model weights (subject to license exceptions discussed below). At present, the controls apply to model weights trained with 10^26 computational operations or more, a threshold which BIS will likely increase through amendments of the regulations over time.
b. Three-Tiered Destination-Based System
The Framework established a three-tier destination-based system that will trigger different controls based on the end user:
Tier 1:
- Tier 1 is comprised of entities located in the United States and allied jurisdictions identified in paragraph (a) of supplement no. 5 of Part 740 (currently, Australia, Belgium, Canada, Denmark, Finland, France, Germany, Ireland, Italy, Japan, the Netherlands, New Zealand, Norway, Republic of Korea, Spain, Sweden, Taiwan, and the United Kingdom). BIS and interagency drafters of the Framework deem these governments to have “implemented measures to prevent the diversion of advanced technologies” and to have created “ecosystems that will enable and encourage firms to use advanced AI models to advance the common national security and foreign policy interests of the United States and its allies and partners.”[3]
- However, an entity headquartered outside these jurisdictions or whose ultimate parent is headquartered outside these jurisdictions would not be deemed to fall within Tier 1.
- These countries continue to retain almost unrestricted access to controlled ICs and will generally have no restrictions on their access to model weights and compute power.
Tier 3:
- Tier 3 is comprised of entities headquartered in, or whose ultimate parent company is headquartered in, Macau or destinations specified in Country Group D:5 (currently, China (including Hong Kong), Afghanistan, Belarus, Burma, Cambodia, Central African Republic, Democratic Republic of the Congo, Cuba, Cyprus, Eritrea, Haiti, Iran, Iraq, North Korea, Libya, Lebanon, Russia, Somalia, Republic of South Sudan, Republic of the Sudan, Syria, Venezuela, and Zimbabwe).
- The Framework maintains the current restrictions on the supply of advanced ICs to these countries continue to remain in place and adds additional restrictions. For example, model weights cannot be supplied to Tier 3 countries and data centers in Tier 3 countries are not eligible for VEU Authorization (discussed below).
- Security requirements for Tier 2 entities incentivizes Tier 2 countries to not only adopt a posture similar to the U.S. toward Tier 3 countries, but also, in the long run, to align themselves technologically with the U.S. In addition to de facto export controls, this would mean adoption of other elements of the U.S. technological landscape such as security controls, computing and other standards and integration with existing U.S.-origin technologies.
Tier 2:
- Tier 2 encompasses entities not specified in Tier 1 or Tier 3 and includes entities located in most countries in the world.
- The Framework permits the export of AI chips and associated compute power without a license up to a capped amount of Total Processing Performance (TPP). Data center companies in these jurisdictions can apply for a BIS license to access more compute power, subject to satisfying certain security and other requirements.
- Tier 1 entities may export their model weights to Tier 2 destinations, provided that the end user has instituted specified security measures that will reduce the risk of diversion.
- Notably, Tier 2 includes both countries that are otherwise considered close allies of the United States (including countries in NATO and the European Union, Israel and Singapore) and countries that are often treated by U.S. export controls and licensing decisions as posing higher evasion risks. The common thread among Tier 2 countries is the U.S. Government’s belief that their governments have not yet adopted the kinds of export controls on advanced AI chips and access to computing power that the U.S. and other Tier 1 countries have put in place in recent years.
III. The Framework Addresses Certain Evasion Activities to Limit Access to Advanced ICs and Model Weights.
a. Worldwide Licensing Requirements
According to BIS, Chinese companies have circumvented existing restrictions by using “foreign subsidiaries in a range of uncontrolled destinations to buy ICs subject to [Export Administration Regulations (“EAR”)] controls.” BIS views the risk of evasion through the use of subsidiaries in uncontrolled jurisdictions to be even greater for AI model weights as they can be sent anywhere in the world instantaneously once copied.
The Framework consequently imposes a license requirement to supply (i.e., export, reexport, or transfer (in-country)) controlled ICs and model weights to any end user in any destination. However, the Framework also provides license exceptions and other mechanisms to enable access to advanced IC to certain end users and designations which pose a comparatively low risk of diversion. The licensing requirements are thus based on the tier system described above, with permissive conditions for Tier 1 countries.
Critically, the Framework conditions access to computing power on the closing of a loophole that has troubled BIS drafters for the last several years—how to control access by Tier 3 country governments and entities to AI Infrastructure-as-a-Service (IaaS) offered by data centers outside of Tier 3 countries. Through the imposition of this worldwide licensing requirement, and the Framework’s conditioning of parallel licensing exceptions for those procuring computing power in Tier 1 and Tier 2 countries on not providing AI IaaS to Tier 3 country users, BIS is hoping to close this loophole.[4]
b. The AI Model Weights Foreign Direct Product Rule
Tier 3 entities have increasingly turned to data centers outside the United States and cloud services to remotely access computing power as a result of regulations on advanced ICs introduced in October 2022 and October 2023, along with the financial and logistical challenges of obtaining large clusters of ICs through subsidiaries and other third parties.
Another way in which BIS attempts to limit Tier 3 access to AI IaaS is through its creation of a new foreign direct product (FDP) rule with breathtaking scope. BIS attempts to address this evasion, through an FDP rule that claims jurisdiction over closed-weight AI models trained anywhere in the world with the use of controlled ICs. Given that most advanced ICs manufactured globally remain dependent on at least some U.S. software, technology, and on items produced by U.S. software and technology, and that BIS has claimed jurisdiction over these advanced ICs through prior FDP rules, there will be few closed-weight AI models that would not be subject to the new licensing requirements imposed by this FDP rule.
BIS’s use of FDP rules in recent years arguably has far outpaced its ability to enforce its FDP rules outside of the United States, however, and it is unclear whether and how non-U.S. developers of new AI models will become aware of this new jurisdiction claim and whether they will submit to U.S. licensing authority.
IV. The Frameworks Offers Multiple Pathways for Exporting Advanced ICs
Despite BIS’s imposition of global licensing requirements on advanced ICs or closed weight AI models, the Framework offers several exclusions, exceptions, and favorable licensing policies that it argues will facilitate more responsible diffusion of advanced AI technology and benefits to certain end users and jurisdictions.
a. Exclusion for Open Weight Models
BIS “determined that a reasonable proxy for the performance of an AI model is the amount of compute—i.e., the number of computational operations—used to train the model.” Accordingly, the Framework places restrictions on the export of the model weights of the most advanced AI models.
The Framework’s new restrictions on the export of model weights only applies to closed-weight models trained with 10^26 computational operations or more. The Framework explicitly excludes “open” model weights of any AI model that have been “published” as defined in 15 C.F.R. § 734.7(a) and any closed models that are less powerful than the most powerful open-weight model.
b. License Exceptions
Existing License Exceptions (NAC/ACA)
The Export Administration Regulations (EAR) existing license exceptions for Notified Advanced Computing (NAC) and Advanced Computing Authorized (ACA) will apply to authorize the export of advanced ICs classified under ECCNs 3A090, 4A090, and corresponding .z items (except for 3A090.a items designed or marketed for use in a datacenter).
License Exception Low Processing Performance (LPP)
The Framework creates License Exception LPP to permit exports of advanced computing ICs and corresponding computing power up to a per-entity allocation of 26,900,000 TPP per-calendar year to any individual Tier 2 entity. This annual TPP limit applies to shipments to any individual Tier 2 entity even if the shipments are made by multiple exporters or reexporters or through more than one intermediate consignee.
License Exception Artificial Intelligence Authorization (AIA)
The Framework created License Exception AIA to permit exports of advanced computing ICs and corresponding computing power to Tier 1 countries.
The license exception also permits exports of otherwise controlled closed AI model weights, without an authorization, by companies headquartered in the United States and allies listed in paragraph (a) of supplement no. 5 and (i) the entities obtaining the items are located outside Macau or destinations specified in Country Group D:5; and (ii) the items will be stored in a facility that complies with the certain security standards that are set forth in supplement no. 10 to part 748.
License Exception Advanced Compute Manufacturing (ACM)
The Framework created License Exception ACM to permit exports of advanced computing ICs to “private sector end users” for the purposes of “development,” “production,” and storage (in a warehouse or other similar facility) of such ICs. However, the license exception does not cover exports for the purpose of training an AI model or exports to Tier 3 countries.
“Private sector end user” defined as either (1) an individual who is not acting on behalf of any government (other than the U.S. Government), or (2) a commercial firm (including its subsidiary and parent firms, and other subsidiaries of the same parent) that is not wholly owned by, or otherwise controlled by any government (other than the U.S. Government).
Summary of New Licensing Requirements, License Exceptions and Exclusions and Licensing Policy
ECCN | Tier | License Exceptions and Exclusions | License Application Review Policy |
Advanced ICs
(ECCNs 3A090.a, 4A090.a, and corresponding .z items) |
Tier 1 | ACA* / AIA / ACM / LPP | Presumption of Approval |
Tier 2 | ACA* / ACM / LPP | Presumption of Approval up to TPP cap
Presumption of Denial in excess of TPP cap |
|
Tier 3 | NAC* | Presumption of Denial | |
Advanced ICs
(ECCNs 3A090.b, 4A090.b, and corresponding .z items) |
Tier 1 | N/A (not restricted) | N/A (not restricted) |
Tier 2
D:1 and D:4 countries, excluding destinations also specified in A:5 or A:6 |
ACA* | Presumption of Approval | |
Tier 3
D:5 countries, excluding destinations also specified in A:5 or A:6 |
NAC* | Presumption of Denial | |
Closed-Weight AI Models
(ECCN 4E091) |
Tier 1 | AIA / Open-weight AI models | Presumption of Approval |
Tier 2 | Open-weight AI models | Presumption of Denial | |
Tier 3 | Open-weight AI models | Presumption of Denial |
* Except for 3A090.a items designed or marketed for use in a datacenter.
c. License Review Policy (Tier 1 Presumption of Approval; Tier 2 Per-Country Allocation)
The Framework creates a presumption of approval for exports of advanced ICs and closed-weight AI models to Tier 1 countries.
The Framework provides a favorable license review policy for exports of advanced ICs and corresponding computing power up to a per-country allocation of 790,000,000 TPP for Tier 2 countries for the period from 2025 to 2027. BIS will review applications for the supply of advanced ICs to Tier 2 countries under a presumption of approval, up to this amount.
d. Revamped VEU Authorizations
In October 2024, BIS introduced a Data Center VEU Authorization to facilitate the supply of advanced ICs to end users in destinations that do not raise national security or foreign policy concerns. The Framework bifurcates the Data Center VEU Authorization into a Universal VEU (UVEU) Authorization and National VEU (NVEU) Authorization.
The UVEU Authorization is available only to Tier 1 entities, and, subject to certain geographic allocation limits, enables UVEU to deploy data center in Tier 2 destinations. Specifically, a UVEU cannot transfer or install more than 25% of its total AI computing power—i.e., the AI computing power owned by the entity all its subsidiary and parent entities—to or in locations outside of Tier 1 countries. U.S. UVEUs are required to maintain at least 50% of their total AI computing power in the United States. Moreover, UVEUs cannot transfer or install more than 7% of its total AI computing power to or in any single Tier 2 country. This kind of AI computing power location requirement has no precedent within the EAR, but more broadly tracks policy efforts by the Department of Commerce and interagency partners to induce the world’s leading advanced IC manufacturers to locate more of their production capacity in the United States.
The NVEU Authorization, on the other hand, is available to Tier 2 entities on a per-company, per-country basis (i.e., separate authorizations required for each Tier 2 country, even if undertaken by the same company), subject to quarterly caps. BIS explains that these allocation caps represent computing power clusters that are approximately 12 months, or one generation, behind the cluster size it believes will be needed to train the most advanced dual-use AI models. Importantly, the total amount of computing power authorized for exports to NVEUs will not count towards the amount of computing power allocated to a Tier 2 country.
In order to receive NVEU Authorization, a data center operator that owns its advanced computing capacity must apply to BIS and go through an intensive application process that will be subject to interagency review. The criteria for NVEU approval are extensive and span over two pages of the Federal Register (published at least initially in a smaller font size) and include requirements on data center ownership, physical security, supply chain security, personnel security, and acceptable use criteria that are subject to documentation, auditing and reporting requirements. Approved applicants for the NVEU Authorization will be listed in the EAR.
V. Conclusion
The Framework is a sweeping and unprecedented attempt to regulate the global diffusion of AI technology, especially to countries that pose national security and foreign policy challenges to the United States. It reflects the U.S. Government’s recognition of the strategic importance of AI as a transformative and disruptive technology that can have profound implications for military, economic, and social domains. It also reflects the U.S. Government’s determination to maintain its leadership and competitive edge in AI innovation, while preventing the misuse and exploitation of AI by adversaries and competitors.
The Framework drafters have acknowledged the heavy lifts that stakeholders impacted by these new regulations will be required to make by staggering its implementation dates. The new worldwide licensing requirements on computing power and on closed-weight AI models, and associated license exceptions will not be implemented until May 15, 2025, and the significant security and other requirements associated with the VEU license exceptions will not be required to be in place until January 15, 2026. Even if new Commerce officials opt to postpone the implementation of these new regulations by two months under authority of President Trump’s Regulatory Freeze Pending Review Executive Order, the sweeping nature of these new regulations will require many companies and other industry stakeholders to begin taking steps now to reflect new controls and to be in a position to take advantage of the different authorizations the Framework makes available.
The Framework, however, is not without its challenges and uncertainties. It is likely to face significant critique by at least some types of AI sector stakeholder during the public comment period, which is currently scheduled to close on May 15, 2025. It is also likely to face technical, and political hurdles, as well as potential backlash and countermeasures from affected countries and entities, as it is implemented. On a technical level, its efficacy is premised in part on widespread awareness of the new export controls it puts in place, especially by stakeholders located outside of the United States, and it is unclear whether the Trump Administration will expend the resources required to publicize and educate non-U.S. persons on how the regulations will work, including convincing those training AI models outside the United States that their closed-weight AI models may be subject to U.S. licensing controls based on the computing power and infrastructure required to train them. There are also data center business model and associated contract-related changes that some Tier 2 country stakeholders would be required to make to take advantage of the NVEU and LPP authorizations that will take many months to implement.
The perceived efficacy of the Framework is also challenged by the release of the Chinese company DeepSeek’s r1 open-weight AI model, which DeepSeek released after BIS published the interim final rule establishing the Framework. In allowing exports of open-weight models without a license, BIS “assess[ed] that the most advanced open-weight models are currently less powerful than the most advanced closed-weight Models.” However, the availability of powerful open-weight AI models like DeepSeek that approximate the capabilities of the most advanced closed-weight, and the possibility that powerful AI models could be trained with lower levels of compute power, both challenge key assumptions underlying the Framework..
Moreover, implementation of the regulations will also require extensive coordination and cooperation among U.S. allies. On the intergovernmental level, the BIS and U.S. Government agency partners such as the Department of State will need to act and think multilaterally to continue to enjoy the cooperation of Tier 1 countries and to attract support for a U.S.-lead AI ecosystem in Tier 2 countries, which may be more difficult to do amidst threats and other actions that the Trump Administration may take unilaterally to advance the America First Trade Policy.
[1] Framework for Artificial Intelligence Diffusion, 90 Fed. Reg. 4544 (Jan. 13, 2025) (hereinafter the Framework).
[2] President Donald, J. Trump, Regulatory Freeze Pending Review Executive Order, Jan. 20, 2025 (available at https://www.whitehouse.gov/presidential-actions/2025/01/regulatory-freeze-pending-review/).
[3] Framework at 4548.
[4] For example, the Framework requires Tier 2 compute providers to institute exacting security measures, including new security measures in supplement no. 10 to part 748, to prevent the use of their compute power by Tier 3 entities.
Gibson Dunn’s lawyers are available to assist in addressing any questions you may have regarding these issues. For additional information about how we may assist you, please contact the Gibson Dunn lawyer with whom you usually work, the authors, or the following leaders and members of the firm’s International Trade practice group:
United States:
Ronald Kirk – Co-Chair, Dallas (+1 214.698.3295, [email protected])
Adam M. Smith – Co-Chair, Washington, D.C. (+1 202.887.3547, [email protected])
Stephenie Gosnell Handler – Washington, D.C. (+1 202.955.8510, [email protected])
Donald Harrison – Washington, D.C. (+1 202.955.8560, [email protected])
Christopher T. Timura – Washington, D.C. (+1 202.887.3690, [email protected])
David P. Burns – Washington, D.C. (+1 202.887.3786, [email protected])
Nicola T. Hanna – Los Angeles (+1 213.229.7269, [email protected])
Courtney M. Brown – Washington, D.C. (+1 202.955.8685, [email protected])
Amanda H. Neely – Washington, D.C. (+1 202.777.9566, [email protected])
Samantha Sewall – Washington, D.C. (+1 202.887.3509, [email protected])
Michelle A. Weinbaum – Washington, D.C. (+1 202.955.8274, [email protected])
Karsten Ball – Washington, D.C. (+1 202.777.9341, [email protected])
Hugh N. Danilack – Washington, D.C. (+1 202.777.9536, [email protected])
Mason Gauch – Houston (+1 346.718.6723, [email protected])
Chris R. Mullen – Washington, D.C. (+1 202.955.8250, [email protected])
Sarah L. Pongrace – New York (+1 212.351.3972, [email protected])
Anna Searcey – Washington, D.C. (+1 202.887.3655, [email protected])
Audi K. Syarief – Washington, D.C. (+1 202.955.8266, [email protected])
Scott R. Toussaint – Washington, D.C. (+1 202.887.3588, [email protected])
Lindsay Bernsen Wardlaw – Washington, D.C. (+1 202.777.9475, [email protected])
Shuo (Josh) Zhang – Washington, D.C. (+1 202.955.8270, [email protected])
Asia:
Kelly Austin – Denver/Hong Kong (+1 303.298.5980, [email protected])
David A. Wolber – Hong Kong (+852 2214 3764, [email protected])
Fang Xue – Beijing (+86 10 6502 8687, [email protected])
Qi Yue – Beijing (+86 10 6502 8534, [email protected])
Dharak Bhavsar – Hong Kong (+852 2214 3755, [email protected])
Arnold Pun – Hong Kong (+852 2214 3838, [email protected])
Europe:
Attila Borsos – Brussels (+32 2 554 72 10, [email protected])
Patrick Doris – London (+44 207 071 4276, [email protected])
Michelle M. Kirschner – London (+44 20 7071 4212, [email protected])
Penny Madden KC – London (+44 20 7071 4226, [email protected])
Irene Polieri – London (+44 20 7071 4199, [email protected])
Benno Schwarz – Munich (+49 89 189 33 110, [email protected])
Nikita Malevanny – Munich (+49 89 189 33 224, [email protected])
Melina Kronester – Munich (+49 89 189 33 225, [email protected])
Vanessa Ludwig – Frankfurt (+49 69 247 411 531, [email protected])
© 2025 Gibson, Dunn & Crutcher LLP. All rights reserved. For contact and other information, please visit us at www.gibsondunn.com.
Attorney Advertising: These materials were prepared for general informational purposes only based on information available at the time of publication and are not intended as, do not constitute, and should not be relied upon as, legal advice or a legal opinion on any specific facts or circumstances. Gibson Dunn (and its affiliates, attorneys, and employees) shall not have any liability in connection with any use of these materials. The sharing of these materials does not establish an attorney-client relationship with the recipient and should not be relied upon as an alternative for advice from qualified counsel. Please note that facts and circumstances may vary, and prior results do not guarantee a similar outcome.