Pennsylvania is using existing permitting powers to introduce stricter requirements for data centre projects, putting energy costs, water use, community benefits, transparency and grid reliability at the centre of AI infrastructure development.

Pennsylvania has introduced a new approach to AI data centre regulation that could influence how other US states manage the rapid expansion of AI infrastructure.

Rather than waiting for a new law to be passed, Pennsylvania Governor Josh Shapiro has used existing state permitting authority to establish requirements for new data centre developments.

The approach links approval of new projects to a set of infrastructure and community requirements known as the Governor's Responsible Infrastructure Development (GRID) Requirements.

Under the new framework, data centre developers must agree to specific conditions before the Pennsylvania Department of Environmental Protection (DEP) will process their permit applications.

The policy addresses several of the biggest challenges created by the rapid growth of AI infrastructure—including electricity demand, water consumption, local community concerns and pressure on the power grid.

Pennsylvania Introduces New AI Data Centre Rules

Governor Josh Shapiro signed Executive Order 2026-05 in Harrisburg, directing the Pennsylvania Department of Environmental Protection to change how it handles permits for proposed data centre projects.

The order applies to new applications and requires developers to commit to the state's GRID requirements through a Consent Order and Agreement.

Developers must also obtain the necessary local approvals.

The policy does not technically prohibit companies from refusing to sign the agreement. However, projects that do not meet the required conditions cannot move through the state's permitting process in the same way.

This effectively makes compliance with the GRID framework a major requirement for new data centre development in Pennsylvania.

The state's Department of Revenue is also applying the framework when considering the sales and use tax exemption associated with data centre equipment.

In addition, data centre projects have been removed from Pennsylvania's permit fast-track programme.

What Are Pennsylvania's GRID Requirements?

The GRID framework is designed to make large infrastructure projects more accountable for the costs and impacts they create.

One of its central principles is straightforward:

Data centre growth should pay for the infrastructure and resources required to support that growth.

Under the framework, developers are expected to address areas including:

• Electricity generation

• Transmission infrastructure

• Distribution infrastructure

• Water conservation

• Local employment

• Workforce training

• Community engagement

• Community benefit agreements

• Environmental protection

• Public disclosure

This represents a significant shift in how AI infrastructure projects can be evaluated.

Instead of focusing primarily on investment and economic development, regulators are also considering the long-term costs that large data centres can impose on communities and utility systems.

Data Centres Must Pay Their Infrastructure Costs

One of the most important elements of the new framework concerns electricity.

Large AI data centres can require enormous amounts of power because AI training and inference depend on energy-intensive computing infrastructure.

Under Pennsylvania's GRID requirements, developers are expected to cover the full cost of the generation, transmission and distribution infrastructure required for their projects.

The objective is to prevent those costs from being shifted to ordinary households and businesses.

This issue has become increasingly important as technology companies build increasingly large computing facilities for artificial intelligence, cloud computing and other digital services.

If electricity demand rises sharply because of new data centres, regulators must determine who should pay for the additional infrastructure.

Pennsylvania's position is that the projects creating the demand should bear the associated costs.

Water Consumption Becomes Part of Data Centre Regulation

Electricity is not the only resource affected by large-scale data centre development.

Many data centres require substantial amounts of water for cooling their computing equipment.

Pennsylvania's framework therefore includes water conservation requirements.

Developers must consider the amount of water their facilities will consume and take steps to reduce unnecessary usage.

This is becoming an increasingly important issue for AI infrastructure because larger computing clusters can generate significant amounts of heat and require sophisticated cooling systems.

As data centres become larger, their environmental footprint is becoming an increasingly important consideration during project approval.

Communities Get a Bigger Role in Data Centre Development

The new Pennsylvania approach also places greater emphasis on local communities.

Developers are expected to engage with communities early in the planning process rather than waiting until major decisions have already been made.

Early engagement can give residents and local authorities an opportunity to influence issues such as:

• Facility design

• Environmental impact

• Traffic

• Water consumption

• Noise

• Electricity infrastructure

• Employment

• Local economic benefits

Developers are also expected to hire and train workers locally where possible and establish community benefit agreements.

The broader objective is to make sure communities receive tangible benefits from hosting large technology infrastructure projects.

Pennsylvania Targets Data Centre Secrecy

One of the more significant aspects of Pennsylvania's policy involves transparency.

The executive order prevents Commonwealth agencies from using nondisclosure agreements for data centre projects.

This is important because data centre development has historically involved considerable commercial confidentiality.

Companies may not always publicly disclose details about:

• Expected electricity consumption

• Water requirements

• Facility ownership

• Infrastructure requirements

• Planned capacity

• Development timelines

Pennsylvania's approach attempts to make more of this information available to the public.

The reasoning is that communities should have access to information about major infrastructure projects that could significantly affect local resources.

New Public Reporting Requirements

The transparency requirements go beyond simply publishing project information.

Data centre operators will be required to provide information including:

• Annual electricity consumption

• Natural gas consumption

• Estimated average electricity use during peak periods

• Total water consumption

• Maximum daily water demand

• Measures designed to protect air and water quality

This information can help regulators and communities understand the actual resource requirements of large data centre facilities.

It also creates a more transparent record of how AI infrastructure affects local resources.

Pennsylvania Creates a Public Data Centre Project Map

The state has also introduced a public-facing map showing proposed data centre projects that have engaged with the Department of Environmental Protection.

The map allows residents to follow projects as they move through the review process.

According to figures released alongside the order, more than 100 projects appeared in public databases.

Of those:

• 58 had approached the DEP

• 15 had submitted permit applications

• 5 had the requirements necessary to begin a first phase

Making this information publicly accessible gives communities greater visibility into the scale and location of planned data centre development.

Data Centres Could Face Power Curtailment Before Households

Another major element of the policy concerns electricity reliability.

Pennsylvania's Special Counsel for Energy Affordability has been directed to approach the state's Public Utility Commission with proposals concerning large data centre electricity demand.

The recommendations include requiring data centres to absorb costs associated with PJM reliability backstop auctions rather than passing those costs to other customers.

The state also wants more accurate forecasting and disclosure of data centre electricity demand.

Most significantly, the policy supports an approach in which data centres could be curtailed before residential customers when the electricity grid comes under significant pressure.

This could have major implications for AI infrastructure operators.

AI companies typically design data centres around high availability and continuous computing capacity. If facilities can be required to reduce their electricity consumption during grid emergencies, operators may need to design systems with greater flexibility.

Why Grid Reliability Matters for AI

AI infrastructure is becoming one of the fastest-growing sources of electricity demand in some regions.

Large computing clusters require substantial amounts of power, and demand can increase further as companies deploy increasingly powerful AI models.

The challenge for utilities is that electricity infrastructure must be able to handle these additional loads without compromising reliability for existing customers.

Pennsylvania's policy effectively establishes a principle:

Large technology projects should not increase the cost or reliability risks faced by ordinary electricity customers without taking responsibility for those impacts.

This could become an important consideration for other states competing for AI data centre investment.

Pennsylvania's Approach Does Not Depend on a New Law

Perhaps the most notable aspect of Pennsylvania's policy is how it was implemented.

The state did not need to pass an entirely new data centre law before taking action.

Instead, the executive order uses existing permitting authority and regulatory mechanisms.

Developers can agree to the requirements through a formal consent agreement, while the state uses its existing regulatory authority to determine whether projects can proceed through the permitting process.

This approach could attract attention from other state governments because it demonstrates that governments may have options beyond:

• Passing new legislation

• Introducing a moratorium

• Creating entirely new taxes

• Establishing new regulatory agencies

Existing permitting frameworks can potentially be used to introduce additional requirements for large infrastructure projects.

Political Debate Surrounds the New Framework

The policy has also attracted political criticism.

Pennsylvania's House passed bipartisan legislation designed to codify the GRID requirements, but the Senate did not advance the measure.

Supporters argue that large data centre developments should be responsible for the infrastructure and environmental costs they create.

Critics have questioned the governor's approach and argued that additional regulation could affect investment.

Environmental groups have also criticised the policy from the opposite direction, arguing that it does not go far enough and pointing to existing incentives available to data centre projects.

The debate illustrates the difficult balance governments face.

They want to attract billions of dollars in technology investment while also protecting electricity customers, natural resources and local communities.

Other US States Are Also Regulating Data Centres

Pennsylvania is not alone in dealing with the rapid growth of AI infrastructure.

States across the US are considering legislation focused on large electricity users and data centre development.

Several states have already adopted policies addressing issues such as:

• Electricity costs

• Data centre energy consumption

• Utility infrastructure

• Environmental impact

• Tax incentives

• Grid reliability

New Jersey, for example, has proposed requirements for large facilities that would make them responsible for a significant portion of their projected electricity costs.

Virginia has introduced an energy consumption tax for certain large electricity users.

Other states are considering additional regulations as communities become increasingly concerned about the impact of large-scale computing facilities.

Why AI Data Centre Regulation Is Becoming More Important

The growth of generative AI is changing the economics of data centre development.

AI models require increasingly powerful computing infrastructure.

That means companies need:

• More GPUs

• More servers

• More electricity

• More cooling capacity

• More network infrastructure

• More data centre space

The result is a race to build AI infrastructure.

But infrastructure expansion can create challenges for governments.

A new data centre may bring investment, construction activity, jobs and tax revenue.

At the same time, it can increase pressure on:

• Local electricity systems

• Water supplies

• Roads

• Land

• Environmental resources

• Utility infrastructure

This is why AI data centre regulation is becoming a major policy issue.

Data Centre Developers Face a New Cost Equation

For years, companies selecting locations for data centres primarily considered factors such as:

• Land availability

• Electricity prices

• Fibre connectivity

• Tax incentives

• Climate

• Workforce

• Proximity to customers

Those factors remain important.

However, regulation is becoming another major consideration.

Developers may increasingly need to calculate:

How much will it cost to obtain community support and meet environmental, energy and infrastructure requirements?

Pennsylvania's policy makes this particularly clear.

The cost of developing an AI data centre is no longer simply the cost of land, servers and electricity.

It can also include the infrastructure required to support the project and the commitments necessary to gain regulatory and community approval.

What Pennsylvania's Policy Could Mean for AI Companies

AI companies and cloud providers planning new infrastructure may need to rethink how they approach data centre development.

Future projects could require more detailed planning around:

Energy

Companies may need to demonstrate how they will obtain electricity without placing unreasonable costs on existing customers.

Water

Developers may need to invest in more efficient cooling systems and water conservation measures.

Community Relations

Technology companies may need to engage with local residents earlier and provide clearer information about projects.

Transparency

Operators may increasingly have to disclose information about resource consumption and infrastructure requirements.

Grid Flexibility

AI data centres may need systems capable of reducing electricity consumption during periods of grid stress.

Could Pennsylvania Become a Model for Other States?

Pennsylvania's approach is significant because it combines several regulatory objectives within an existing permitting framework.

It does not simply focus on stopping data centre development.

Instead, it attempts to establish conditions under which development can proceed.

That distinction could make the framework attractive to other governments facing similar challenges.

States looking for a middle ground between unrestricted data centre expansion and outright moratoriums could study Pennsylvania's approach.

The biggest question will be whether the framework successfully balances investment with public interests.

The Future of AI Infrastructure Regulation

The AI industry is entering an infrastructure-intensive phase.

The next generation of AI systems will require enormous amounts of computing capacity, which means more data centres and greater demand for electricity and cooling resources.

Governments therefore face a difficult question:

How can they support AI investment without allowing the costs of that expansion to fall disproportionately on local communities and existing utility customers?

Pennsylvania's GRID framework provides one possible answer.

Its approach focuses on making developers responsible for infrastructure costs, increasing transparency, protecting resources and involving communities earlier in the development process.

Whether other states adopt the same model remains to be seen.

But one thing is becoming increasingly clear: AI infrastructure can no longer be treated purely as a technology-sector issue.

It is also an energy, environmental, economic and public-policy issue.

Conclusion

Pennsylvania's new approach to AI data centre regulation represents a significant development in the growing debate over how governments should manage AI infrastructure.

By using existing permitting authority, the state has introduced requirements covering energy costs, water conservation, community engagement, workforce development, transparency and grid reliability.

The policy could influence how other states approach the rapidly growing demand for AI computing infrastructure.

For data centre developers, the message is equally clear.

Building the next generation of AI infrastructure will require more than securing land and power.

Companies will increasingly need to demonstrate that their projects can operate responsibly, contribute to local communities and avoid shifting their infrastructure costs onto the public.

As AI continues to expand, consent, transparency and infrastructure responsibility may become just as important to data centre development as electricity, land and connectivity.

Frequently Asked Questions

What is Pennsylvania's new AI data centre regulation?

Pennsylvania has introduced requirements for new data centre projects through the Governor's Responsible Infrastructure Development (GRID) framework. The requirements address infrastructure costs, energy, water, community engagement and transparency.

What are Pennsylvania's GRID requirements?

The GRID requirements are a set of conditions covering areas such as electricity infrastructure, water conservation, local hiring and training, community benefits and public disclosure.

Who pays for the infrastructure needed by new data centres?

Under Pennsylvania's framework, data centre developers are expected to pay the costs associated with the generation, transmission and distribution infrastructure needed to support their projects.

Why is water consumption important for AI data centres?

Large data centres generate significant heat and often require cooling systems that consume water. As AI computing capacity increases, managing water consumption has become an important environmental consideration.

Can Pennsylvania data centres be required to reduce electricity use?

The state's proposed approach gives regulators a mechanism to prioritise data centre curtailment during periods of grid stress, potentially requiring facilities to reduce electricity consumption before residential customers are affected.

Why are governments regulating AI data centres?

AI data centres can require large amounts of electricity, water and infrastructure. Governments are increasingly evaluating how these projects affect utility customers, communities, natural resources and grid reliability.

Could other US states adopt Pennsylvania's approach?

Yes. Pennsylvania's use of existing permitting authority provides a potential framework for other states looking for ways to regulate data centre development without immediately creating entirely new legislation.

Will AI data centre regulations slow AI development?

Regulation could increase development costs or add additional approval requirements, but policymakers argue that responsible regulation can help ensure AI infrastructure growth does not create disproportionate costs for communities and existing utility customers.