ANALYSIS: Washington Wants a Single AI Law, but the States Have Already Passed 109

109 Refuse to Stay Silent

The Number That Bother Washington

Some numbers describe a movement. This one looks more like a barricade: as of July 1, 2026, 29 U.S. states had passed 109 laws on artificial intelligence. In addition, 28 other pieces of legislation targeted data centers. It’s not a perfect system. It isn’t even always a coherent policy. It’s the sound of a country that refuses to wait for a distant capital to finally decide who to protect, from what, and at what cost.

The White House would like a “minimally restrictive” national framework. The language is clean, almost gentle. But the gentleness ends where obligations begin: warning a child that a digital companion is not human, allowing a worker to challenge an automated decision, preventing a household from footing the bill for a data center’s electrical infrastructure.

The Patchwork and the Mirror

We call this a patchwork, as if it were a sewing mistake. Yet each piece exists because a specific risk encountered an institution still capable of taking action. This disorder did not arise from a provincial penchant for forms. It arose from a federal vacuum.

One hundred and nine laws do not prove that the states have found the right answer. They prove that Washington had not provided one.

The patchwork is not a flaw in the American system. It is its reflex when the center hesitates.

The Executive Order That Cannot Decree Everything

A National Ambition

On December 11, 2025, Donald Trump signed Executive Order 14365. Its stated goal is crystal clear: to uphold American dominance in artificial intelligence through a uniform and streamlined national policy. The industrial rationale is clear. A company deploying the same model across 50 jurisdictions may face differing definitions, deadlines, and obligations. A startup may pay a higher price for this complexity than a giant already surrounded by lawyers.

Recognizing this cost does not mean signing a blank check. A single standard can simplify compliance. It can also simplify the lack of protection. It all depends on what the central authority replaces, the strength of what it proposes, and the actual authority it possesses.

The Constitutional Limit

The executive order issues directives to the federal executive branch; it does not magically render every state law null and void. The legal analysis by White & Case emphasizes this point: preemption normally arises from an act of Congress, not from the president’s will alone. Agency powers, the Tenth Amendment, interstate commerce, and conditions attached to federal funds promise separate battles.

A president can open a front. He cannot declare on his own that he has already won before all the judges.

The executive order is powerful as a strategy; as a legal verdict, it remains a promise yet to be tested.

The Task Force as a battering ram

Thirty days to choose the targets

The executive order directed the Attorney General to establish, within 30 days, an AI Litigation Task Force whose responsibility would be to challenge state laws deemed incompatible with federal policy. The weapons cited are well-known: the Commerce Clause, preemption by existing federal rules, the First Amendment, and any other grounds the Department of Justice deems useful.

This choice changes the nature of the debate. The discussion is no longer merely about which protections are appropriate. The questions now are: Which state will agree to defend its law, how long will it be able to do so, and which company will wait for the ruling before complying?

Governing Through Uncertainty

A lawsuit does not need to succeed to have an effect. It imposes costs. It slows down regulation. It gives local opponents a new argument: why pass a law today that Washington will challenge tomorrow? Uncertainty thus becomes public policy without ever having been voted on as such.

But there is still no way to say which laws will be struck down, or even which ones will survive an initial challenge. The midyear assessment shows, above all, that the legislative process has continued.

The federal battering ram is already making its presence felt in public discourse. The judicial door, however, has not yet come crashing down.

The most immediate power of preemption may not be to overturn, but to cause hesitation.

Broadband Held Hostage

BEAD, an unexpected lever

The most brutal mechanism is not a constitutional doctrine. It is money. The executive order directs the Department of Commerce to specify under what conditions states that have enacted so-called “onerous” AI laws could lose access to certain non-deployment funds from the BEAD program, which is dedicated to broadband access. It also calls on agencies to review their discretionary grants and consider similar conditions.

The connection is politically explosive. A state might have to defend a rule regarding an algorithm while explaining to a rural community why its digital funding is at risk of becoming part of the battle. It’s no longer just Silicon Valley versus a local legislature. It’s a fiber-optic road being weighed against a transparency requirement.

The Threat and the Data

However, the available figures do not yet demonstrate a clear deterrent effect. The Tech Policy Press analysis finds no clear correlation between the adoption of laws, partisan control, and the average amount of BEAD allocations. California, while included in some projections, actually skews the averages significantly.

Making broadband access contingent on regulatory compliance would be an immense lever of influence. For now, its actual impact remains to be proven.

When money replaces reason, the federation begins to resemble a bill that people are threatening to stop paying.

Congress had, however, spoken

99 to 1

On July 1, 2025, the Senate removed a 10-year moratorium on state AI regulations from a budget bill. The vote: 99 to 1. Edward Markey, Maria Cantwell, and Marsha Blackburn had introduced the amendment. Democrats and Republicans hadn’t suddenly agreed on all digital regulations; they refused to allow states to be silenced for a decade.

This vote matters because it shatters the narrative of a national consensus. If uniformity were so indisputable, it would not have suffered such a crushing defeat. The Senate chose to preserve local autonomy, including when it comes to protecting children and consumers.

Coming Back Through the Side Door

The December executive order takes a different approach to an ambition that Congress had not enshrined. Targeted litigation, agency opinions, threats to funding, legislative recommendations: the strategy no longer resembles a head-on moratorium. It advances through multiple channels, each with its own technical language, each capable of producing a piece of the result.

What 99 senators refused to impose as a whole is now returning in administrative and judicial fragments.

A legislative defeat does not always kill an idea; sometimes, it teaches it to evade the spotlight.

Child protection as a revealing exception

Fourteen Laws on Digital Companions

By 2026, 14 states had passed laws affecting companion chatbots. Six were led by Democratic majorities, eight by Republican majorities. More than 100 proposals had been introduced. This geographic distribution is enough to break down partisan reflexes: concern about systems that simulate human presence, speak to minors, and may address self-harm or sexuality is not limited to any single political party.

The responses vary. Some bills emphasize age verification, others parental supervision, warnings, or risk assessment. There is not yet a common model. There is a shared sense of urgency.

What Washington Leaves Out

The executive order stipulates that the future legislative recommendation must not preempt otherwise lawful state protections regarding child safety. This exception is not a mere detail. It acknowledges that local regulations can be effective where the risk is concrete, politically evident, and cannot be indefinitely deferred to a federal commission.

Washington condemns the patchwork system, yet retains one of its most sensitive components. This contradiction reveals where the human pressure lies.

We sometimes discover the value of local authority at the very moment the central government promises not to crush it.

Colorado, a Testing Ground Under Threat

An Automated Decision, a Real Life

In 2024, Colorado passed SB24-205 on high-risk AI systems. Its core is not an abstract theory of machines. It addresses decisions with far-reaching consequences. The law requires developers and deployers to exercise due diligence against known or foreseeable risks of algorithmic discrimination. It mandates impact assessments, consumer notifications, the correction of inaccurate data, and—where technically feasible—an appeal with human review.

These are procedural obligations. But a procedure can deprive a person of a job, housing, credit, or an essential service. Compliance is not the opposite of humanity. It is sometimes the last way to reclaim it.

The Price of Uniformity

The decree specifically cites the rules that may impose constraints on model results or their disclosures. Legal analysts see Colorado as a likely battleground. This does not make the law perfect. Its timeline and scope have already sparked local debates. It simply means that the first major clash may occur where a state has dared to establish a right to challenge.

If national uniformity begins by eliminating human review, it does not simplify protection. It simplifies denial.

The American laboratory is worthless if only one experiment—the federal one—is allowed to survive.

Data centers emerge from the shadows

Twenty-eight bills for infrastructure

The 28 laws passed in 2026 concerning data centers tell the story of another awakening. For years, these facilities were primarily treated as projects to be enticed by tax exemptions. Now, states are examining electricity rates, the costs passed on to other customers, the water used for cooling, contracts with large consumers, tax breaks, and local moratoriums.

The so-called “intangible” machine has regained its weight. It consumes energy, occupies land, and requires power lines, power plants, pipelines, and rate-setting decisions. Artificial intelligence does not float above the country. It settles somewhere, and someone pays for it.

The Exception That Proves the Rule

In its future recommendations, the executive order excludes rules regarding computing infrastructure and data centers, except for certain general reforms to permitting. Here again, Washington implicitly acknowledges that place possesses its own knowledge and legitimacy. An Oklahoma power grid is not a federal abstraction. Neither is a water source in Idaho.

Computing power is sold as a cloud. But the bill lands in a city, on a meter, at someone’s home.

The more AI grows, the more policymakers must look at the ground on which its servers stand.

The False Choice Between Innovation and Protection

Speed as the Ultimate Argument

The administration presents light regulation as a prerequisite for American dominance. This argument deserves better than ridicule. The United States faces intense technological competition. Conflicting rules can slow deployments, discourage small businesses, and further concentrate the market in the hands of corporations capable of absorbing legal costs.

But the absence of rules also leads to concentration. When a citizen cannot understand an automated decision, only the dominant players possess the data, the model, and the vocabulary to defend it. When an incident remains secret, the market learns too late. When infrastructure costs are passed on to households, innovation receives an invisible subsidy.

Two Obligations to Fulfill

The right question, then, is not whether to choose between accelerating progress and protecting people. It is to determine which rules truly protect without stifling technology, and which level of government can enforce them competently. This requires a federal government capable of establishing a credible baseline, then allowing states some flexibility when their risks differ.

Innovation without protection creates victims. Protection without intelligence can create monopolies. We must reject both forms of complacency.

A serious country doesn’t just ask how far its machines can go, but who will step in when they cross the line.

Convergence Hidden Beneath Conflict

The Same Fears in Different States

The mid-2026 tally does not show a surge in formal bipartisanship. About 31% of the laws passed had co-sponsors from both parties—nearly the same as the previous year—and the proportion of primary sponsors from both sides had even declined. Yet a convergence on certain issues is emerging: children, consumers, insurance, dynamic pricing, regulated professions, energy, and data centers.

This is less spectacular than a grand compromise in Washington. It may be more profound. Opposing majorities face the same consequences and craft different responses around a similar core.

Federalism as a Learning Process

This diversity allows for comparison. Which disclosure requirement works? Which right of appeal becomes impractical? Which audit truly sheds light? Which rule unintentionally benefits the giants? A federation can turn 50 jurisdictions into chaos. It can also transform them into experiments that can be observed, corrected, and replicated.

But we must still allow the learning process to take place and honestly publicize the failures. No state should be treated as sacrosanct simply because it acted before Washington.

Divergence is not always a problem. When properly managed, it becomes the way a democracy learns without risking everything at once.

Before unifying the rules, perhaps we should consider what their differences teach us.

Companies Facing Two Uncertainties

The visible cost of the patchwork

For developers, insurers, employers, and platforms, the proliferation of laws means compliance checklists, incompatible definitions, varying effective dates, and differing enforcement regimes. This cost is real. To deny it would be to replace analysis with slogans. It can slow down a startup more than a multinational corporation and encourage a national approach.

But the promised federal framework does not yet exist. There is no general law passed by Congress that properly supersedes local statutes. There is an executive order, agency timelines, threats of litigation, exceptions, and a forthcoming recommendation.

The Hidden Cost of Waiting

Companies must therefore comply with existing laws while anticipating those that Washington will attempt to overturn. This is a second source of uncertainty, created not by the states but by the federal strategy itself. White & Case, in fact, advises continuing to comply with state obligations as conflicts take shape.

The federal government promises to cure regulatory uncertainty by piling constitutional, financial, and judicial uncertainty on top of it.

Simplicity is not merely proclaimed; it is demonstrated by a clear rule that survives the next hearing.

What States Can Do Wrong

Virtue Is Not Automatic

Defending states’ authority does not mean endorsing every piece of legislation. A legislature may draft provisions that are too broad, give in to moral panic, protect a local actor, impose a technically impossible obligation, or multiply audits that reveal nothing. Federalism distributes wisdom; it also distributes error.

Colorado itself has faced pressure to adjust its regulations. Certain laws regarding chatbots vary significantly on issues of age, consent, and supervision. Legislation targeting pricing or insurers can produce unintended consequences. Good intentions do not constitute an impact assessment.

The Challenge Facing Our Side

This fact troubles those who view any local regulation as heroic resistance to the White House. Resistance does not excuse imprecision or legislative theater. A rule must identify a harm, impose an enforceable duty, provide a remedy, and be open to evaluation.

States deserve the right to act. They never deserve the right to escape accountability for what they have done.

Rejecting federal preemption does not require us to applaud the first form that includes the words “artificial intelligence.”

What Trump Gets Right, Then Undermines

The Need for a National Floor

Trump sees a real problem: the United States cannot allow its AI policy to remain forever a mere sum of local responses. Issues of national security, competition, cutting-edge technology, and trade transcend state lines. A federal framework is necessary. On this point, his assessment deserves credit.

We need common definitions, investigative powers, transparency commensurate with the risk, liability rules, and a baseline of rights that no state can lower. A company should know what it is required to do everywhere; a citizen should know what rights cannot be taken away from them anywhere.

The Method of Threats

But a floor cannot be established by threatening broadband access, launching a task force against pioneers, and asking agencies to find ways to preempt after the moratorium failed in Congress. This method transforms a need for unity into a struggle for dominance.

Trump is right to want a country capable of speaking with one voice. He is wrong if that voice demands that 50 others remain silent.

Democratic unity is not the silence of those below; it is an agreement strong enough not to fear their experience.

The courts will not replace politics

Years to Draw the Line

The lawsuits that have been announced may clarify the respective powers of the federal government and the states. They may define the scope of the commerce clause, the limits of government agencies, and the legality of financial terms. This work is necessary when institutions clash. It will also be slow, costly, and fragmented.

In the meantime, models will evolve. New products will emerge. Teenagers will talk to AI companions. Employers will buy screening tools. Electric grid operators will negotiate with data centers. Political life cannot be put on hold until the final ruling.

The void the judge does not fill

A court will rule on whether a rule is permissible. It will not dictate which national policy America should choose. It will not single-handedly design a balanced framework of security, innovation, recourse, and transparency. That responsibility remains with Congress, and ultimately with the elected officials who must vote in their own names.

A judge can shift a boundary. He cannot write, on behalf of a nation, the contract that the nation refuses to debate.

When all politics ends up in the courthouse, it is often because the legislature has left its seat empty.

The real patchwork is that of responsibilities

What protects us before uniformity does?

One hundred and nine laws. The number keeps coming up, but it doesn’t mean that all one hundred and nine solutions are good. It means that one hundred and nine decisions were made while the national framework remained incomplete. Some protect a child. Others regulate an insurer, a price, infrastructure, or a high-risk model. Others will be amended or will disappear.

The federal government can challenge them. States can resist. Businesses can demand clarity. Citizens, for their part, are already living with the decisions produced by the systems. They don’t live in a legislative future. They live in the present reality.

A federation worthy of its technology

America needs a national law strong enough to provide protection everywhere and humble enough to learn from what works locally. It needs a Congress that passes legislation, agencies that stay within their mandates, states that tailor their own rules, and courts that adjudicate without being turned into a permanent public policy workshop.

The issue is not about saving the patchwork. It is to prevent uniformity from becoming a euphemism for a void imposed on everyone.

One hundred and nine laws do not constitute harmony.

But they are an answer.

And as long as Washington offers nothing better than a threat, why should the states put down their needles?

Signed, Maxime Marquette, columnist

Columnist’s Transparency Disclaimer

Editorial Stance

I am not a journalist, but a columnist and analyst. My expertise lies in observing and analyzing the geopolitical, economic, and strategic dynamics that shape our world. My work consists of dissecting political strategies, understanding global economic trends, contextualizing the decisions of international actors, and offering analytical perspectives on the transformations that are redefining our societies.

I do not claim to possess the cold objectivity of traditional journalism, which is limited to factual reporting. I aim for analytical clarity, rigorous interpretation, and a deep understanding of the complex issues that affect us all. My role is to make sense of the facts, place them within their historical and strategic context, and offer a critical interpretation of events.

Methodology and Sources

This text respects the fundamental distinction between verified facts and interpretive analyses. The methodological rule is consistent: factual information is published only if it is supported by a verifiable source, and the sources actually used in this article are listed under “Sources,” never here.

Categories of primary sources used by the publication, when applicable: official communiqués from governments and international institutions, public statements by political leaders, reports from intergovernmental organizations, and dispatches from recognized international news agencies.

Types of secondary sources: specialized publications, internationally recognized news media, analyses from established research institutions, and reports from sector-specific organizations.

When an article cites statistical, economic, or geopolitical data, it is sourced from data-producing institutions (intergovernmental organizations, central banks, national statistical institutes), and the specific institution is listed under “Sources.”

Nature of the Analysis

The analyses, interpretations, and perspectives presented in the analytical sections of this article constitute a critical and contextual synthesis based on available information, observed trends, and expert commentary cited in the sources consulted.

My role is to interpret these facts, contextualize them within the framework of contemporary geopolitical and economic dynamics, and give them coherent meaning within the broader narrative of the transformations shaping our era. These analyses reflect expertise developed through continuous observation of international affairs and an understanding of the strategic mechanisms that drive global actors.

This article describes a situation as documented at the time of publication, not a prediction: subsequent developments may alter these perspectives. No updates are promised in advance; when an article is corrected or supplemented, the change is dated within the text.

ANALYSIS: Washington Wants a Single AI Law, but the States Have Already Passed 109