Last updated 6 months ago. Our resources are updated regularly but please keep in mind that links, programs, policies, and contact information do change.
Four days before Donald Trump took office in January 2025, an Abu Dhabi royal signed a $500 million investment agreement in a Trump family cryptocurrency venture. Sheikh Tahnoon bin Zayed Al Nahyan, who also serves as the UAE’s national security advisor, directed $187 million of that sum to Trump family entities. Months later, the Trump administration approved the shipment of 500,000 advanced artificial intelligence chips to the United Arab Emirates—technology that Obama and Biden officials had specifically denied to the UAE because of concerns it could be diverted to China.
The United States government maintains an elaborate apparatus for reviewing foreign investments and technology shipments. Multiple federal agencies, formal committees with statutory authority, and established legal procedures are designed to flag risks and conflicts of interest. Yet the chip shipment sailed through with minimal documented opposition.
If the safeguards failed this spectacularly on a deal this obvious, they’re not safeguards at all.
Two Review Systems
American technology shipments face scrutiny through two separate review systems. The first examines the transaction itself: whether selling particular technology to particular buyers threatens the country. The second examines who’s buying: if a foreign investor is trying to gain control over American companies or technology through investment, that triggers a specialized review.
For semiconductor shipments to the UAE, the primary legal mechanism is the Export Administration Regulations, administered by the Commerce Department’s Bureau of Industry and Security. These regulations govern “dual-use” technologies—items with both civilian and military applications. Advanced AI chips fall squarely into this category. They can power weather modeling or accelerate autonomous weapons development.
When a company wants to ship such items, it files an application with BIS describing what they’re sending, where it’s going, who will receive it, and what they’ll use it for. BIS then refers the application to other agencies—State, Defense, Energy—for their input. Each agency gets roughly 30 days to recommend approval, conditional approval, or denial. If agencies disagree, the decision escalates through an interagency committee structure to the Commerce Secretary level, and if necessary, to the President’s Export Administration Review Board.
The UAE chip transaction triggered questions about a parallel system. If Sheikh Tahnoon or his investment entities were seeking to acquire control over American technology companies through capital, that would implicate the Committee on Foreign Investment in the United States—an interagency body chaired by Treasury that includes representatives from Commerce, State, Defense, and other agencies. CFIUS has broad authority to review foreign investments that could affect the country, particularly in critical technology sectors.
No public record shows CFIUS reviewed Sheikh Tahnoon’s World Liberty Financial investment. The crypto venture isn’t a traditional technology company, and CFIUS jurisdiction over crypto platforms remains legally unclear. More significantly, the Trump administration has made clear it prefers promoting rather than restricting foreign technology capital.
Sheikh Tahnoon’s capital went into one regulatory space. The chip transaction happened in another. Apparently nobody connected the dots.
Policy Shift Under New Administration
Career officials at BIS have historically represented one constituency: specialists trained in technology assessment, often skeptical of shipments that carry even moderate diversion risks. Political appointees representing the White House’s broader policy agenda represent another.
The Biden team tightened controls on AI semiconductors significantly. In January 2025, before the transition, Biden officials released an ambitious regulatory framework designed to govern global diffusion of AI technology. That “AI Diffusion Rule” would have created a worldwide licensing regime for frontier AI model weights and expanded controls on AI chips to most countries worldwide. Countries that posed diversion risks—particularly the UAE, given G42’s connections to Chinese firms—would face significant restrictions.
Within months of taking office, Commerce officials explicitly rescinded the Biden AI Diffusion Rule in May 2025, declaring it “overly bureaucratic” and harmful to American business. The stated approach prioritized commercial opportunity and competitive positioning against China over the more cautious stance of the predecessor.
David Sacks, appointed as the White House’s AI and crypto czar, emerged as a key architect of this new policy, advocating internally for aggressive technology approvals to Middle Eastern allies as a way to maintain American influence in the region and prevent China from establishing technological dominance.
When the UAE chip application reached BIS for review under new leadership, it encountered a reviewing committee with different priorities than it would have under Biden.
The Technology and Diversion Risk
The license covered Nvidia’s H200 and similar chips, capable of approximately 141.1 petaFLOPS in FP8 precision mode. These represent among the most powerful commercial processors available globally, designed specifically for training and operating large language models and other frontier AI systems.
Prior policy had sharply restricted sales of such chips to certain destinations. The Biden team’s concerns about UAE access stemmed from documented business relationships between G42—the Emirati AI company ultimately expected to receive significant quantities of the approved chips—and Chinese technology firms, particularly Huawei and BGI Genomics.
G42’s CEO is Peng Xiao, a Chinese-born entrepreneur who previously led divisions of DarkMatter Group, a UAE surveillance company, and concluded deals with Huawei before his recent acquisition of UAE citizenship. That history created what officials viewed as significant diversion risk—a scenario in which American semiconductors could flow from the UAE through existing business channels into Chinese military or intelligence applications.
By 2023, responding to congressional pressure, G42 announced it was divesting from Chinese holdings, including an estimated $100 million stake in ByteDance. This divestment was presented as sufficient to address American concerns.
The baseline question for any decision is: Will this technology serve legitimate civilian purposes at the stated end-user, or does it pose unacceptable risks of diversion to military end-uses or hostile nations? On paper, G42’s stated purpose—building data centers to advance UAE AI capabilities—appears legitimate. In practice, officials had documented reasons to view the diversion risk as non-trivial.
Timeline and Interagency Process
Reconstructing the specific sequence of interagency decisions has proven difficult because much of the review process remains confidential. What is publicly documented reveals several striking features: the timeline was compressed compared to historical precedent for similar high-stakes technology decisions, and the documented interagency process shows minimal evidence of the kind of institutional resistance that typically characterizes sensitive decisions.
In May 2025, shortly after the $2 billion MGX-Binance crypto transaction using the USD1 stablecoin was announced, the White House indicated it had approved Microsoft’s request to ship Nvidia chips to the UAE. The Commerce Department announced these approvals in September 2025, stating they were based on “stringent safeguards.” By November 2025, Commerce made an additional announcement that it had authorized shipments of “up to 35,000 Nvidia Blackwell chips” equivalent to the approved quantity. The broader framework contemplated UAE access to approximately 500,000 chips through multiple tranches over time, though the full timeline and conditions for subsequent authorizations remained unclear.
Under the regulatory framework governing license processing, reviewing agencies must formally provide recommendations within 30 days of referral. When agencies disagree about whether to approve a sensitive transaction, the regulations provide for escalation through interagency committees. Ultimately, if consensus cannot be reached, the matter goes to the President’s Export Administration Review Board. There is no public evidence that such a procedure occurred in the UAE case. This suggests either the decision was made entirely within BIS without interagency referral, or the referral process was handled entirely in classified channels without documentation available to congressional oversight.
Department of Defense concerns about the chips potentially reaching China—the stated primary rationale for previous restrictions—would normally trigger Defense’s participation in any BIS license review. Similarly, the State Department’s Bureau of International Security and Nonproliferation participates in reviews involving countries where the U.S. maintains diplomatic sensitivities.
Yet no public record exists of Defense or State submitting written recommendations on this particular decision. No official statements from those departments were released explaining their position on whether the UAE represented an acceptable risk for AI chip shipments.
This contrasts sharply with documented interagency disagreements over other sensitive transactions, which generated documented disagreement between Commerce (which wanted to approve) and Defense (which raised concerns) before an escalated decision was made.
Conflicts of Interest
At the center of the procedural vulnerability was an extraordinary conflict of interest that formal government ethics rules should have prevented.
Steve Witkoff, who serves as the Middle East envoy—a position that directly involves advising the President on technology and policy toward the UAE—is also co-founder and significant financial beneficiary of World Liberty Financial. Witkoff’s relatives received approximately $31 million from Sheikh Tahnoon’s $187 million infusion into that venture.
Federal ethics regulations and the conflict of interest statute prohibit federal employees from participating in matters in which they have a financial interest. An employee in Witkoff’s position who is advising the President on policy toward an Emirati official who has directed $187 million to his cryptocurrency company is positioned to benefit from that same Emirati official receiving what had previously been denied American technology.
The formal requirement in such circumstances would be recusal—the official stepping away from decisions where their financial interest could create bias.
There is no public record of Witkoff recusing himself from discussions about the UAE chip transaction or any formal ethics review determining whether such recusal was required. Similarly, David Sacks, the White House’s AI and crypto czar who played a central role in negotiating the chip terms, maintains relationships with venture capital firms receiving capital from UAE sovereign wealth funds. This creates potential conflicts between his role advising on policy and his private financial interests in favorable outcomes for UAE technology deals.
The absence of documented ethics reviews, recusals, or conflict resolutions in matters involving such substantial financial interests represents a departure from the procedural safeguards that have historically governed sensitive interagency decisions. Even when no intentional misconduct occurred, the institutional practice has been to create a formal record demonstrating that conflicts were either avoided through recusal or that ethics counsel determined no conflict existed.
The absence of such documentation in the UAE case signals either that no such review occurred, or that reviews occurred but remained entirely confidential and outside normal congressional oversight mechanisms.
Comparison to Historical Precedent
When the Defense Department considered whether to allow Intel to maintain semiconductor fabrication operations in China in 2013, that review took several months and generated documented concerns from Defense and State before a conditional approval was issued. When Huawei attempted to purchase American technology in the early 2010s, interagency reviews often took many months and typically resulted in denials, with detailed explanations of concerns.
When Microsoft sought to deploy high-performance AI chips to Israel through licenses approved under Biden policy, the review process documented in the Federal Register and Congressional notices took several months. It included formal coordination among Defense, State, and Commerce. The specific performance thresholds, end-user verification procedures, and conditions all received documented interagency scrutiny.
The UAE approval shows compressed timing from application to decision. If a license application for 500,000 chips was filed in early 2025, a normal 30-to-90-day review timeline would project toward final decision in the April-to-June 2025 window. The actual announcement timing in May-September 2025 suggests the decision was made within roughly that window, if not faster.
Congressional oversight committees requested documentation of the review process from Commerce in late 2025 and early 2026, according to public statements from Senator Elizabeth Warren and Representative Mike Gallagher. Those committees sought “emails, documents, and communications” relating to how the chip decision was made. They also asked “whether the White House had coordinated on this deal with the president’s relatives or business interests.”
The White House provided limited information in response, asserting that the President had “no involvement” in business decisions and that his assets were held in a trust managed by his children, thus creating no conflict of interest.
Four Structural Failures
The UAE chip decision revealed several structural vulnerabilities in how American processes protect against conflicts of interest. These vulnerabilities were latent before this White House but became operational problems when decision-makers had significant financial entanglements with foreign investors receiving benefits from approvals.
First: The incomplete jurisdictional coverage between CFIUS and BIS. CFIUS reviews foreign capital in American companies and technologies. BIS reviews American shipments of technology and products. In the UAE case, these happened in sequence but potentially without integrated review. Sheikh Tahnoon’s capital into World Liberty Financial was reviewed neither by CFIUS nor documented to the public, occurring in an apparent regulatory gap between systems. Simultaneously, his anticipated receipt of American semiconductors through the G42 transaction went through BIS without CFIUS documenting whether the crypto infusion by the same investor should have informed the technology assessment.
This sequential rather than integrated review process created an information barrier. Reviewers couldn’t see the full scope of the transaction—crypto capital from a foreign official, followed by that same official’s technology firm receiving previously-denied American chips.
Second: The insufficiency of recusal and ethics procedures when officials have relatives with linked financial interests. Standard federal ethics rules allow recusal to resolve conflicts of interest. But modern business structures often locate financial interests with relatives rather than the federal employee directly, creating ambiguity about whether conflicts exist. Witkoff’s $31 million interest in World Liberty is held through relatives; therefore, some interpretations of ethics rules might suggest Witkoff himself has no direct “conflict” even though the financial benefit flows indirectly to him.
Yet the practical incentive alignment—decisions favorable to an Emirati official who benefits the relatives’ finances—creates the same conflict that the recusal requirement was designed to prevent. Formal ethics review procedures adapted to this reality do not appear to have been applied.
Third: The discretionary scope given to political appointees in decision-making. Career officials in the government have long understood certain technology shipments as presenting unacceptable risks requiring formal denial. But when political leadership reorients policy priorities—favoring commercial opportunity over cautious assessment—the career civil service faces a procedural challenge. They can argue the case, but if leadership wants to approve a transaction, formal legal processes typically allow it, provided basic requirements are met.
The UAE case shows this dynamic in operation: previous teams viewed the diversion risk as sufficiently serious to warrant denial; the new White House assessed the same facts differently and prioritized strategic alignment with the UAE over the diversion risk. Neither triggered the formal interagency disagreement and escalation procedures that would have created a documented record of the competing concerns.
Fourth: The opacity of interagency committee deliberations when matters touch Presidential interests. When a Chief Executive has significant financial interests in foreign investor relationships, the normal transparency mechanisms of interagency government become constrained. Agencies cannot easily document dissent from the President’s preferred policy through formal channels without appearing to challenge the President directly.
The result can be a muted interagency process where agencies that might otherwise formally register concerns instead acquiesce quietly. This creates documentation of approval without evidence of the underlying debate.
Applicable Legal Framework
The Export Control Reform Act of 2018 governs the legal authorities the Commerce Department exercises in administering controls. ECRA contains explicit conflict-of-interest protections prohibiting the Commerce Department from charging fees or imposing conditions based on anything other than foreign policy. The law was designed to prevent the government from using licenses as leverage for commercial advantage or to benefit particular officials.
Yet ECRA does not explicitly require recusal of executive officials with financial interests in the outcomes of decisions. It assumes that general federal ethics law and presidential ethics regulations will prevent such conflicts from arising in the first place. Those general ethics rules prohibit federal employees from participating in matters affecting their financial interests. They require ethics counsel review when such interests exist. They provide for formal recusal procedures.
But those rules apply at the individual employee level. When leadership prioritizes a particular outcome, enforcement of ethics rules against the relevant officials becomes politically delicate.
The Foreign Investment Risk Review Modernization Act of 2018, which gives CFIUS its authority, contains similar language and similar assumptions about ethics rules preventing conflicts. It empowers CFIUS to block or condition foreign capital that threatens the country, including capital by foreign government officials in sensitive technology sectors.
Yet there is no evidence CFIUS reviewed or even formally evaluated whether Sheikh Tahnoon’s crypto capital should have triggered review. He is a foreign government official making a large infusion into an American technology-adjacent company that was about to benefit from his other business entities receiving major U.S. technology shipments.
Congressional Response and White House Position
Congressional oversight committees are attempting to establish whether the interagency process functioned as intended or failed due to conflicts of interest. Senator Elizabeth Warren has called for investigation into whether the chip decision was influenced by the crypto transaction. She requested detailed documentation of the decision-making process and communications involving officials who benefited from the crypto deal. Representatives on the House Select Committee on China have similarly requested information about how concerns regarding Chinese access to the chips were evaluated.
Commerce Department and White House officials have declined to provide detailed documentation, citing presidential privilege, confidentiality of licensing decisions, and assurances that no conflict of interest occurred because the President’s assets are held in a trust. They argue that the UAE chip transaction represented a sound strategic decision based on legitimate grounds—maintaining American technological leadership in the Middle East, securing important Emirati partnership on Middle East diplomacy, and preventing China from establishing technological dominance in a critical region.
The procedural question remains distinct from the substantive policy question. Even if the chip transaction was strategically sound, the process by which it was approved appears to have omitted standard safeguards. The absence of documented interagency review, the apparent lack of formal recusal procedures addressing obvious conflicts of interest, and the integration of crypto capital and technology approvals in a compressed timeline all suggest that institutional procedures designed to protect against corruption proved insufficient to their task.
Reform proposals being discussed include Congressional legislation requiring advance notice to Congress before approving sensitive technology shipments to foreign government officials or entities they control. Other proposals would expand ethics recusal requirements to explicitly cover federal officials with relatives’ financial interests in foreign investor relationships affecting policy outcomes. Still others would require integrated review procedures where CFIUS examines foreign capital and BIS approvals together when the same investor is involved, rather than as sequential separate processes.
The White House has not supported such reforms, arguing they would impede its ability to promote American technology internationally and compete effectively against China. The position rests on the view that controls have historically been too restrictive, harming American business while failing to achieve goals because foreign adversaries simply build their own alternatives. From this perspective, the UAE chip transaction represents wise policy—strengthening an important ally while maintaining American technological presence and preventing China from becoming the exclusive source of semiconductors to the region.
Regardless of whether the UAE represented an acceptable risk for chip shipments, the interagency process that approved the transaction omitted the formal institutional safeguards that exist precisely to ensure such decisions are made based on criteria rather than personal financial benefit. The machinery designed to prevent corruption in technology decisions has a gap wide enough to drive 500,000 AI chips through. Until that gap is closed—through either executive discipline or legislative requirement—the committee that’s supposed to review these deals before they happen isn’t reviewing much of anything.
Our articles make government information more accessible. Please consult a qualified professional for financial, legal, or health advice specific to your circumstances.