Virginia’s governor has signed an executive order that creates an artificial intelligence task force and directs state agencies to give local communities more weight in data centre decisions, a move expected to slow approvals.
Key takeaways
- The Verge reports that Virginia’s governor has ordered state government to take steps that could give local communities a larger say in how data centre projects are developed.
- The same order, identified as Executive Order 22, bars executive branch officials from signing nondisclosure agreements, a practice commonly used during site-selection talks.
- The order also establishes a state artificial intelligence task force, tying data centre policy explicitly to the growth of AI computing.
- Virginia matters disproportionately here because, as The Verge notes, it is already home to what is described as the data centre capital of the world.
- The order signals a shift in how a leading host state approaches data centre growth, but the practical effect on individual projects is not yet known.
What the executive order actually does
According to The Verge, the governor of Virginia has directed the state’s executive branch to take steps that could hand local communities more influence over how data centre developments proceed, and that could slow the pace at which such projects are approved. The order is identified as Executive Order 22.
Two elements are named in the reporting. The first is a ban on executive branch officials signing nondisclosure agreements. In practice, NDAs are routinely used when companies scout sites for large facilities: they keep the identity of the prospective tenant, and sometimes the scale and power requirements of the project, confidential while negotiations with officials are under way. Removing the state’s ability to enter those agreements changes what local residents and elected bodies can learn, and when.
The second is the creation of an artificial intelligence task force. The precise membership, remit and reporting timetable of that body are not described in the material available here, so they should not be assumed. What the pairing indicates is that the state is treating AI strategy and data centre siting as one policy question rather than two.
Why this is in the news now
Data centre construction has become one of the most visible physical consequences of the AI build-out. Training and serving large models requires dense clusters of specialised chips, and those clusters require buildings, cooling and, above all, electricity. Over the past few years that demand has turned what was once a quiet infrastructure category into a mainstream local political issue in the places where the buildings land.
Virginia is the sharpest version of that story. The Verge describes the state as already home to the data centre capital of the world, which means any change in its permitting posture is watched well beyond its borders. When a jurisdiction with that concentration signals that it intends to slow approvals and widen local input, other states and counties read it as a precedent.
The NDA element sharpens the news value further. Confidentiality in economic development deals is an established norm, and a state deciding unilaterally to stop signing such agreements is a departure from standard practice rather than a procedural tweak.
The background a newcomer needs
A data centre is a purpose-built facility full of servers, networking equipment and cooling systems. Some serve ordinary cloud computing: storage, websites, business software. A growing share is built for AI workloads, which are more power-hungry per square metre than traditional hosting.
Northern Virginia became a hub for historical reasons rather than a single decision. Early internet infrastructure and network exchange points concentrated there, fibre routes followed, and once a critical mass of connectivity existed, each new facility was easier to justify in the same place than anywhere else. Tax treatment and available land reinforced the cluster.
In the United States, land-use decisions are generally made locally: counties and municipalities control zoning, and they decide whether a particular parcel can host a large industrial building. State government influences that process indirectly, through utility regulation, tax policy, environmental permitting and the terms on which it participates in economic development negotiations. That division of labour explains why a state executive order works by empowering localities rather than by directly refusing projects.
Who is affected and how
Residents near proposed sites are the most immediate constituency. Their concerns typically centre on noise from cooling equipment, the visual impact of large windowless buildings and substations, traffic during construction, water use for cooling, and the transmission lines needed to feed the load. More disclosure earlier in the process gives them more time to organise or negotiate.
Local officials gain leverage if the state stops signing NDAs, because the information asymmetry between a developer and a county board narrows. Whether that leverage produces better outcomes or simply longer timelines depends on how it is used.
Electricity customers are affected less directly but more broadly. Large new loads can require generation and transmission investment, and how those costs are allocated between data centre operators and other ratepayers is a live regulatory question in several states.
Developers and cloud operators face schedule risk. For projects whose value depends on being energised by a particular date, an extra round of local scrutiny is a material cost. Construction workers, equipment suppliers and localities that want the tax base sit on the other side of the ledger.
Where informed people disagree
The central disagreement is about what data centres are worth to a community. Supporters point to property tax revenue that can fund schools and services from a facility that generates little traffic and few demands on public services, and argue that hosting computing capacity is strategically valuable. Critics respond that the jobs are few once construction ends, that the tax benefits are often reduced by exemptions, and that the costs — land, power, water, landscape — fall on neighbours who did not choose them.
A second dispute concerns transparency. Defenders of NDAs argue that companies will not engage with a jurisdiction that cannot keep commercial plans confidential, and that banning them simply moves projects elsewhere. Opponents argue that public bodies making irreversible land and energy commitments should not do so under secrecy that prevents residents from evaluating the deal.
A third runs through energy policy: whether new demand accelerates investment in generation, or crowds out other users and complicates decarbonisation commitments. Reasonable analysts differ, and the answer varies by grid.
What it means in practice
The practical effect is procedural rather than prohibitive. Nothing in the reporting suggests data centre construction is being banned. The mechanism is friction and information: more disclosure, more local participation, and therefore, plausibly, longer timelines.
For anyone following a specific project, the immediate change is what can be known and when. Without state NDAs, details that previously surfaced late may surface while decisions are still open.
For the wider industry, the significance lies in the precedent. If the most concentrated market in the world adds process, developers may weigh other states more heavily, and those states may then face the same local debates on a compressed schedule. The alternative outcome is that the industry adapts, accepting more disclosure as the cost of access to existing fibre and power infrastructure. Which of these happens is not yet determinable.
What to watch next
Three things will indicate whether this is a durable shift. The first is implementation: executive orders direct agencies, and the substance emerges in the guidance, model ordinances or review criteria that follow. The second is the AI task force’s output, including who sits on it and what it is asked to recommend — details not available in the current reporting.
The third is the response of localities and developers. Whether counties use any new leverage to impose conditions, and whether projects continue to be announced at the previous pace, will show more than the order’s text does. Legislative follow-up, and how utility regulators handle cost allocation for large loads, are the other indicators worth tracking.
Frequently asked questions
What is Executive Order 22 in Virginia?
The Verge reports that it is an executive order issued by Virginia’s governor directing the state government to take steps that could give local communities a larger role in data centre development and could slow approvals. It also bans executive branch officials from signing nondisclosure agreements and establishes an artificial intelligence task force. Further operational details of the order are not available in the reporting summarised here.
Does the order stop data centres being built in Virginia?
No. Nothing in the available reporting indicates a ban or moratorium. The described mechanism is procedural: increasing the say of local communities and, as a consequence, potentially slowing the approval process. Land-use approval in the United States generally rests with counties and municipalities, so a state order works mainly by changing the information and leverage available to those local decision-makers rather than by rejecting projects itself.
Why do nondisclosure agreements matter in data centre projects?
NDAs are commonly used during site selection, keeping the prospective tenant’s identity and sometimes a project’s scale or power requirements confidential while officials negotiate. That confidentiality limits what residents and local bodies know while decisions are being shaped. Banning executive branch officials from signing them is intended to reduce that information gap. Supporters of NDAs argue confidentiality is necessary for companies to engage at all.
Why is Virginia called the data centre capital of the world?
The Verge refers to Virginia as already hosting the data centre capital of the world. The concentration is usually explained by history: early internet infrastructure and network exchange points clustered in Northern Virginia, fibre routes followed, and each additional facility became easier to justify where connectivity, land and power were already available. Tax treatment and proximity to existing operators reinforced the cluster over time.
Why are data centres controversial with local residents?
Common objections include noise from cooling systems, the visual impact of large buildings and associated substations, water use, construction traffic, and new transmission lines. Residents also question whether promised tax revenue outweighs those effects, given that large facilities employ relatively few people once built. Supporters counter that the tax base is substantial and that the facilities place limited demand on public services such as schools.
How does AI increase demand for data centres?
Training and running large AI models requires dense clusters of specialised processors, which consume considerably more power per unit of floor space than conventional hosting. That raises the electricity, cooling and site requirements of new facilities and increases the number being proposed. The result is that AI strategy and data centre siting have become a single policy question in host jurisdictions, which is reflected in pairing an AI task force with data centre measures.
Sources and further reading
- The Verge — policy reporting on the Virginia executive order, the data centre provisions and the AI task force.
- Official state of Virginia publications — the text of executive orders and any implementing guidance issued by state agencies.
- County and municipal planning departments in Northern Virginia — zoning applications, staff reports and public hearing records for individual data centre proposals.
- State utility regulators and grid operators — filings and reports on large-load interconnection and how associated costs are allocated among customers.
Surfaced from the rss:verge signal “state data centre restrictions”. AI-assisted draft, editorially reviewed.
