The AI Preemption Executive Order’s BEAD Strategy Faces Steep Legal Hurdles

On Dec. 11, 2025, President Trump signed an executive order titled “Ensuring a National Policy Framework for Artificial Intelligence” that, as part of a broader attempt to block state regulation of AI, directs the National Telecommunications and Information Administration (NTIA) to condition an estimated $21 billion in broadband funds on states rolling back AI regulations the administration deems “onerous.” But the administration’s chosen legal mechanism faces a significant statutory obstacle: The Broadband Equity, Access, and Deployment (BEAD) Act does not clearly allow NTIA to condition BEAD funding on state AI policy.

The state attorneys general preparing to challenge the executive order have the better reading of the BEAD statute—and even if a court found the text ambiguous, interpretive canons would likely break the tie against the administration. The major questions doctrine requires clear congressional authorization before an agency asserts power over questions of vast economic and political significance. Federalism-protecting canons require the same before Congress conditions federal funds on state policy or displaces traditional state authority. Section 5(a) implicates each of these concerns: It would leverage a broadband infrastructure program to reshape national AI policy, targeting state laws governing disclosure, liability, and consumer protection. Yet BEAD—a statute about deploying service and connecting locations—never mentions AI. The statute does not offer the clarity these canons require.

How BEAD Works

Congress created BEAD in 2021 as a $42.45 billion formula grant program “to bridge the digital divide.” Millions of Americans, disproportionately in rural and low-income areas, lack access to reliable high-speed internet. Congress designed BEAD to close that gap by funding broadband infrastructure to connect unserved and underserved locations. The BEAD statute sets up a system in which each state’s allocation is fixed by formula, NTIA must approve state plans for using that allocation, and any money that is not used is redistributed under the same formula.

Each state’s allocation depends on its connectivity gap. Under subsection (c), the assistant secretary is required to allocate BEAD funds among the states according to each state’s share of locations lacking broadband access, with adjustments for high-cost areas. The availability of those allocated amounts is “subject to” NTIA’s approval of a state’s letter of intent, initial proposal, or final proposal, so states must submit plans before they can draw down their formula-based allocations. 

NTIA’s approval power in subsection (e) evaluates each state’s proposed spending plan against specific statutory criteria. Before receiving funds, a state must submit an initial proposal and then a final proposal for how it will use its allocation. For each proposal, NTIA must determine whether “the use of funds proposed” is “in the public interest” and “effectuates the purposes of this Act,” and whether the proposed use “complies with subsection (f).” 

Subsection (f) specifies what BEAD money can buy. It authorizes spending in six categories: deploying service to unserved and underserved locations; connecting eligible community anchor institutions; data collection, broadband mapping, and planning; installing broadband infrastructure in multifamily residential buildings; broadband adoption and affordable devices; and “any use determined necessary by the Assistant Secretary to facilitate the goals of the Program.” The list is framed in terms of broadband deployment, connections for particular kinds of locations, and related planning and adoption activities.

When states fail to use their allocations, the statute directs reallocation through the same formula that governed the original distribution. If a state fails to secure approval for a required proposal or fails to use its full allocation by the applicable deadline, NTIA “shall reallocate” the unused amounts to other grantees “in accordance with the formula” in subsection (c). The result is a program in which Congress determines each state’s allocation, NTIA reviews how states propose to spend those allocations, and any money that moves among states does so under the same statutory formula.

The Trump Administration’s BEAD Reforms

Under the Biden administration, NTIA allocated funds to every state under the BEAD formula. Many states received approval for their initial proposals and began drawing down support for planning and early deployment, with significant portions of their allocations earmarked for nondeployment activities: mapping, digital literacy programs, and devices to help residents get online. Three states—Louisiana, Delaware, and Nevada—had received final proposal approval and were preparing to begin deployment.

In June 2025, however, NTIA announced the “Benefit of the Bargain” reforms, a major restructuring of the program. The name captures the administration’s theory: The Biden administration had layered requirements onto the BEAD program—including a preference for fiber-optic technology, rate regulation, and various labor and climate provisions—that allegedly inflated costs and slowed deployment, shortchanging taxpayers. The reforms stripped those requirements and opened competition to fixed wireless and satellite providers. To implement this shift, NTIA voided all previously approved final proposals, required states to conduct a new round of subgrantee selection under the revised rules, and rescinded approval for nondeployment uses like digital literacy programs and workforce training.The amounts freed up by those revisions were consolidated into a “remaining funds” pool, estimated at over $21 billion. This pool sits outside the formula-based allocations described above. Although the legality of these reforms is contested on multiple fronts, this piece addresses only the executive order’s attempt to condition BEAD funds on state AI regulation.

The Preemption Executive Order

The preemption executive order leverages this pool to advance the administration’s AI priorities. Section 5(a) directs the secretary of commerce to issue a policy notice “specifying the conditions under which States may be eligible for remaining funding” under BEAD. The policy notice “must provide that States with onerous AI laws … are ineligible for non-deployment funds, to the maximum extent allowed by Federal law.” The order cites 47 U.S.C. § 1702(e)–(f)—the approval and use-of-funds provisions described above—as the statutory basis for this authority. The administration’s implicit theory is that these provisions give NTIA discretion to consider state AI policy when evaluating whether a state’s proposed use of funds is “in the public interest” and “effectuates the purposes” of BEAD.

The order frames this linkage as serving BEAD’s mission. Section 5(a) requires the policy notice to “describe how a fragmented State regulatory landscape for AI threatens to undermine BEAD-funded deployments, the growth of AI applications reliant on high-speed networks, and BEAD’s mission of delivering universal, high-speed connectivity.” The theory, in short, is that AI applications drive demand for high-speed networks, so state AI regulations that burden those applications threaten the value of BEAD-funded infrastructure.

Reviewability

Before reaching the merits, states will have to navigate BEAD’s unusual review structure. Subsection (o) exempts NTIA’s decisions from the Administrative Procedure Act, including the provisions governing rulemaking and judicial review. That means states likely cannot raise arbitrary-and-capricious challenges to the executive order’s policy notice. In addition, subsection (n) gives the U.S. District Court for the District of Columbia exclusive jurisdiction over any challenge to “a decision of the Assistant Secretary made under this section” and directs that court to “affirm” unless the decision was “procured by corruption, fraud, or undue means,” involved “actual partiality or corruption,” or reflected comparable “misconduct.” The administration might argue these provisions foreclose challenges that NTIA exceeded its authority in issuing the policy notice. 

States might respond that an action exceeding statutory authority isn’t a decision “under this section” at all, so subsection (n) doesn’t govern. But in NRC v. Texas, the Supreme Court recently noted that this type of nonstatutory review is narrow and rejected it where a separate statutory review scheme was available. Still, states could contend that subsection (n)’s standard—limited to corruption, fraud, and misconduct—provides no meaningful avenue to test these claims on the merits. Courts are especially reluctant to read such provisions as foreclosing constitutional claims, such as the Spending Clause challenges discussed below. And interpreting these provisions to preclude all substantive review would imply that Congress granted NTIA sweeping authority over national AI policy while simultaneously insulating that authority from judicial oversight.

The Core Statutory Dispute

Assuming courts reach the merits, here’s the core dispute: Can NTIA condition access to BEAD funding on state AI policy? Section 5(a) of the executive order assumes it can, citing subsections (e) and (f) of the BEAD statute and directing the Commerce Department to explain how state AI regulations threaten BEAD’s mission. But states challenging this would have the stronger argument—both as a matter of statutory text and, more decisively, under interpretive canons that require clear congressional authorization before an agency claims such sweeping power.

The administration’s theory—though not yet fully articulated in court—appears to rest on BEAD’s broad approval language. Section 1702(e) authorizes NTIA to approve state proposals only if “the use of funds proposed” is in “the public interest,” “effectuates the purposes of this Act,” and “complies with subsection (f).” The executive order’s framing suggests the argument that BEAD’s “purposes” include supporting applications running on federally funded networks. Because AI applications rely on high-speed connectivity, state laws that burden AI development could reduce demand for BEAD-funded infrastructure.

But subsection (e) is better read more narrowly. General statutory terms like “public interest” are interpreted in light of neighboring provisions, and here that term sits alongside the requirement that proposals comply with subsection (f)’s connectivity-focused spending categories. That context cabins the inquiry to the spending proposal itself—whether proposed uses fit BEAD’s categories—not into a freewheeling assessment of state policy. 

BEAD’s statutory structure reinforces this reading. The BEAD statute separates allocation in subsection (c) from approval in subsection (e). Congress chose to fix each state’s allocation by formula and provided that any funds a state “fails to use” must be “reallocate[d]” according to the same formula. Exercising its approval power, NTIA can reject a proposal because costs are excessive, technology is inadequate, or projects do not serve unserved locations. But it cannot deny a state its formula-based allocation altogether because that would allow the agency to override the congressionally chosen formula. If NTIA can use its general approval discretion to override the formula, the mandatory reallocation provision is left with little independent work to do. Courts generally resist interpretations that allow a general grant of authority to negate a more specific statutory command.

Nor does the catch-all in § 1702(f)(6) support such a broad assertion of authority. That subsection authorizes “any use determined necessary by the Assistant Secretary to facilitate the goals of the Program.” Courts typically read such provisions, while broad, in light of the specific items that precede them. The first few categories in § 1702(f) are all related to connectivity mechanics: deploying service, connecting institutions, mapping, and infrastructure. State AI laws are categorically different from these connectivity-focused categories.

The administration’s reading also lacks a limiting principle. If “the goals of the Program” extend to maximizing the value of every application that runs on a high-speed network, the agency could just as easily condition funds on other policy preferences. For instance, NTIA could withhold BEAD dollars from states with restrictive cryptocurrency regulations on the theory that blockchain applications also drive demand for high-speed networks. The statute offers no textual basis for distinguishing AI from these other use cases. While NTIA retains discretion to set technical standards for funded networks, extending that discretion to regulate the downstream applications that run on those networks would effectively grant the agency plenary power over the digital economy.

Why Ambiguity Isn’t Enough to Sustain Section 5(a)

But even if a court concludes that BEAD’s text is ambiguous as to whether the executive branch can condition BEAD funds on state AI regulation, three interpretive canons—the major questions doctrine and two federalism-protecting clear-statement rules—should break the tie against the administration. 

The major questions doctrine requires “clear congressional authorization” before an agency asserts authority over matters of “vast economic and political significance.” The doctrine has been applied primarily in regulatory contexts, but the core concern—that agencies should not claim sweeping powers from vague statutory language—applies here as well. Section 5(a) fits that description: It leverages an estimated $21 billion to intervene in a contested national debate over AI regulation, while resting on general terms like “public interest,” “effectuates the purposes,” and “any use determined necessary.” The BEAD statute is about broadband infrastructure; it never mentions AI. A court applying this doctrine would likely demand a clearer textual hook.

Federalism-protecting clear-statement rules point in the same direction. The Supreme Court has held that conditions on federal grants must be stated “unambiguously” in the statute, enabling states “to exercise their choice knowingly, cognizant of the consequences of their participation.” And the Court has held that when federal law threatens to displace traditional state authority, Congress must “make its intention to do so unmistakably clear in the language of the statute.” Section 5(a) implicates both concerns: It would condition BEAD funds on state AI policy, but the statute never mentions AI. And it targets state laws governing disclosure, liability, and consumer protection—classic exercises of state police power. When the statute can plausibly be read either to authorize or to forbid conditioning broadband funds on such laws, these canons favor the reading that preserves state regulatory authority. The statute is ambiguous at best for the administration—and under these canons, ambiguity is not enough.

Constitutional Objections Are Unlikely to Be Dispositive

States might also raise constitutional objections, but the statutory arguments are likely to be dispositive. Under South Dakota v. Dole, Congress may attach conditions to federal grants only if they (a) serve the general welfare, (b) provide clear notice, (c) relate to the federal interest in the program, (d) do not independently violate another constitutional provision, and (e) are not so coercive as to “pass the point at which pressure turns into compulsion.” 

The administration appears poised to satisfy three of these requirements. Courts give Congress wide latitude to define the general welfare, and promoting AI innovation plausibly qualifies. The independent-bar prong seems unlikely to apply, though a final assessment must await the policy notice itself. And coercion likely isn’t a problem for the administration: NFIB v. Sebelius invalidated a threat involving over 10 percent of a state’s overall budget. Even states with the highest relative exposure face losses of roughly 2 to 3 percent of annual expenditures. West Virginia’s estimated nondeployment pool, for instance, represents about 2.8 percent of its $20.4 billion budget. These one-time amounts fall well below the level the Court found coercive in NFIB.

Relatedness is a closer question, but it is unlikely to be dispositive. A court that embraces the administration’s statutory theory will likely embrace its relatedness arguments as well. The statutory argument relies on the theory that AI applications drive demand for high-speed connectivity. If a judge accepts this link to satisfy the “purposes” of the BEAD statute, that finding effectively supplies the factual predicate for constitutional relatedness. If the states win on the statutory issue, by contrast, the constitutional question is practically irrelevant. 

Dole’s notice requirement follows a similar logic. The policy notice will presumably specify any AI policy conditions before states decide whether to accept the remaining BEAD funds. But the Spending Clause requires that states be able to identify the condition from the statute itself, either in the text or in implementing rules the statute clearly authorizes. The question is thus whether BEAD’s text clearly delegates authority to condition broadband funds on state AI policy. That is substantially identical to the statutory inquiry discussed above, so the statutory and constitutional notice questions should track each other even more closely than with the relatedness inquiry. 

This analysis assumes courts can reach the statutory question. If subsection (n) poses greater obstacles for statutory claims than constitutional ones, these constitutional issues could become more significant.

* * *

AI is a topic of rapidly increasing economic and political significance. And Section 5(a) is unprecedented. That combination—a novel assertion of authority over a major emerging policy area—should trigger the major questions doctrine. The federalism stakes point in the same direction: Conditioning federal funds on state regulatory choices invokes clear-statement rules that require Congress to speak unambiguously. The BEAD statute is—at best—ambiguous for the administration. But these canons require more than ambiguity to sustain Section 5(a). The BEAD statute falls short of that standard.