Project Hazelnut Data Center Fight Turns $165 Million Into a Political Liability

Project Hazelnut has offered Hazle Township a $165 million benefits package, yet the proposed data center still faces determined local resistance. The Project Hazelnut data center fight now tests whether developers can purchase public acceptance after trust has already eroded.

NorthPoint Development proposes building 15 data center structures on approximately 1,280 acres near Interstate 81 in northeastern Pennsylvania. Its package includes household grants, community funding, and annual payments to the township. However, residents continue challenging the campus, its power infrastructure, and the secrecy surrounding a possible connection to Amazon.

That conflict reaches beyond one rural township. Pennsylvania leaders welcomed AI infrastructure as a source of investment, construction work, and technological relevance. They now face voters concerned about electricity costs, land use, environmental exposure, and decisions made before communities understood the scale involved.

The central contest is not simply development against environmentalism. It is NorthPoint’s compensation-based case against residents demanding disclosure, local control, and credible limits before construction proceeds. That same conflict is appearing across political lines and throughout the United States.

What the Project Hazelnut Data Center Fight Is Really About

Project Hazelnut changed from a development proposal into a test of whether financial benefits can repair a damaged public process.

NorthPoint’s proposed campus would place 15 data center buildings on a site covering roughly 1,280 acres. A hyperscale data center is a large computing facility designed to support extensive cloud or AI workloads. The project still requires essential township approvals.

NorthPoint announced a revised $165 million community benefits package in June 2026. According to the community package details, it divides the commitment into three main components.

A $45 million resident fund would provide $10,000 to each eligible Hazle Township household. Payment would follow the first data center building receiving an occupancy certificate. Residents could spend the money on household priorities, including improvements or energy upgrades.

Another $15 million would support locally directed programs such as workforce training, education, recreation, and youth initiatives. The remaining $105 million would reach the township through annual payments of $7 million across 15 years.

NorthPoint has said those payments could support a township police force and reduce residential trash collection fees. The developer describes the package as a binding, resident-focused commitment shaped by community feedback.

The proposal also includes assurances concerning the project’s electricity expenses. NorthPoint says service agreements and financial protections with PPL Electric Utilities would make the project responsible for necessary transmission upgrades.

Those commitments address several measurable concerns. They do not resolve the underlying dispute over public consent.

Some residents view the household grants as compensation for consequences they have not been allowed to evaluate fully. Payment also depends on the project crossing a major construction milestone. The structure therefore asks residents to accept the development before receiving the most visible personal benefit.

The sequence matters. Community members learned different pieces of the plan through letters, public meetings, legal proceedings, and right-to-know requests. By the time NorthPoint announced its largest package, opponents had already organized around transparency and property rights.

The offer consequently became part of the conflict instead of settling it. Opponents could portray the grants as an attempt to secure acceptance. Supporters could present them as an unusually direct way to share economic value.

Both interpretations start from the same fact: ordinary economic-development messaging was no longer enough. NorthPoint had to make benefits concrete, personal, and unusually large because residents had made local approval uncertain.

That is why Project Hazelnut explained only as an investment story misses the important change. The developer is no longer selling an abstract promise of regional growth. It is negotiating against an organized community with access to records, legal procedures, and statewide political attention.

Secrecy Turned Amazon’s Shadow Into a Trust Problem

Uncertainty about Project Hazelnut’s intended user became politically damaging because residents had to uncover key relationships themselves.

Amazon has not confirmed that it will operate or occupy the proposed campus. An Amazon spokesperson told Fortune that the company was evaluating multiple Pennsylvania sites and discussing the project with NorthPoint. The spokesperson said no definitive agreement was in place.

That qualification is important. Project Hazelnut should not be described as a confirmed Amazon data center. NorthPoint and local officials have also declined to identify a final user publicly.

However, records obtained through Pennsylvania’s right-to-know process revealed communications involving Amazon and state economic-development officials. According to Fortune’s investigation, those records showed discussions about keeping Amazon’s involvement undisclosed because of contractual concerns.

Amazon Web Services subsequently said it would stop entering nondisclosure agreements with government agencies for data center projects. That policy change addresses future transparency, but it does not erase how Hazle Township residents experienced this proposal.

Residents did not begin with a clear presentation naming the potential customer, computing purpose, power demand, and construction consequences. They assembled their understanding through document requests and related proceedings.

That discovery process changed the political meaning of every later assurance. A promise about jobs or utility protections now had to overcome questions about what else remained undisclosed.

The project’s early branding also contributed to confusion. Officials introduced the site as an innovation or technology campus, language that can suggest research offices or an academic setting. Residents later understood that the proposal involved warehouse-scale computing buildings and major power infrastructure.

The difference is material. A data center campus can reshape transmission planning, land use, emergency services, water management, noise conditions, and municipal finances. Communities need those details before deciding whether promised benefits outweigh local costs.

Amazon’s possible role also connects Project Hazelnut to a much larger state strategy. Governor Josh Shapiro announced a $20 billion Amazon investment in Pennsylvania data centers in July 2025. His administration described it as the largest private-sector investment in state history.

That announcement created a strong incentive for state officials to move projects forward. It also encouraged residents to ask whether local decisions had already been influenced by commitments made at higher levels.

The uncertainty leaves all sides exposed. NorthPoint cannot rely fully on Amazon’s reputation or resources because Amazon has not committed publicly. Amazon receives political scrutiny without formally owning the proposal. Township officials must evaluate a vast campus without a confirmed end user.

Residents face a different information gap. They are asked to judge long-term infrastructure while the ultimate operator, computing workload, development schedule, and total operating footprint remain unsettled.

This is the first major lesson from Pennsylvania data center opposition. Confidentiality that developers treat as routine can become evidence of exclusion when projects require public approvals and new utility corridors.

A community does not need proof of misconduct to lose trust. It only needs to discover that important conversations occurred before residents received a clear account of the project.

A Power Line Made the AI Boom Personal

The proposed transmission route transformed a distant debate about AI investment into a direct conflict over homes, land, and legal power.

PPL Electric Utilities proposed an 11.2-mile high-voltage line to supply the area associated with the data center development. According to residents interviewed by Fortune, 28 properties faced potential eminent-domain action connected to the route.

Eminent domain is the legal power to acquire private property for an approved public use while providing compensation. Its possible use created a sharp contradiction at the center of Project Hazelnut.

Some Hazle Township households were offered grants if the campus reached its first occupancy milestone. Nearby property owners were simultaneously confronting the possibility that a utility corridor could cross their land.

This contrast weakened the developer’s resident-benefit message. A household payment sounds different when another family is defending its property against infrastructure required by the same development.

Opponents formed the Alliance to Stop the Line and filed formal challenges with the Pennsylvania Public Utility Commission. Their concerns included property values, noise, electromagnetic fields, and the need for the project itself.

Administrative law judges later paused proceedings concerning extensions related to the power-line case. The underlying data center developments had not received approval and remained involved in litigation. That pause did not permanently reject the infrastructure, but it gave opponents a procedural victory.

The mechanism behind this conflict is easy to miss in national AI coverage. Software demand becomes computing demand. Computing demand becomes electricity demand. Electricity demand then requires substations, transmission capacity, generation, and utility proceedings tied to specific parcels.

Each step creates another point where residents, regulators, and local governments can challenge the buildout. A technology company can purchase processors quickly, but it cannot bypass every land-use and power-system constraint.

Project Hazelnut therefore reveals why data center development has become harder to separate from energy politics. The server buildings are only the visible portion of a larger physical network.

NorthPoint says the campus has secured necessary power capacity. It also says contractual protections would make the project pay for transmission upgrades needed to serve it. Those commitments directly address fears that existing customers will subsidize new infrastructure.

Yet cost allocation is only one question. Residents also want to know where infrastructure will go, whose land it will cross, and which risks will remain after construction.

This distinction explains why financial concessions have not ended the dispute. NorthPoint’s proposal answers, “What will the community receive?” Opponents are also asking, “What authority will the project exercise over the community?”

No benefits package can answer that second question by itself. It requires transparent route planning, credible regulatory review, local approval, and enforceable limits.

This dynamic pressures more than NorthPoint. Utilities must demonstrate that expansion is necessary and fairly financed. State officials must reconcile investment goals with property rights. Technology companies must accept public identification earlier in the process.

Developers elsewhere should notice the sequence. Once transmission infrastructure activates landowners, the political coalition expands beyond people living beside the proposed server buildings.

A project can then face opposition from homeowners, farmers, environmental advocates, electricity customers, and fiscal conservatives. Their reasons differ, but the same physical infrastructure brings them together.

Pennsylvania Data Center Opposition Crosses Party Lines

Project Hazelnut has become representative because data center concerns do not fit comfortably within America’s usual partisan categories.

Environmental concerns often attract Democratic voters. Property rights, government secrecy, and resistance to eminent domain often mobilize Republicans. Electricity affordability and distrust of large technology companies reach both groups.

A University of Pennsylvania survey found opposition to new data center construction increased by 12 percentage points within four months. The earlier survey, conducted during February and March 2026, found 49% opposition.

Political scientist Matthew Levendusky said researchers found resistance among Democrats, Republicans, independents, frequent AI users, and people who avoided AI. The Pennsylvania polling findings also indicated that opposition was not limited to residents living near a proposed site.

That result challenges the standard “not in my backyard” explanation. Proximity can intensify concern, but it does not fully account for broader anxiety about electricity costs, environmental pressure, surveillance, and AI’s social effects.

Project Hazelnut concentrates several of those grievances in one place. The campus carries an uncertain connection to Amazon, requires extensive power infrastructure, occupies a large site, and follows a state-led investment campaign.

It also shows how political framing can change depending on the speaker. Supporters describe domestic data centers as necessary for jobs, cloud capacity, and competition with China. Opponents describe them as industrial projects that transfer costs and decision-making power to rural communities.

Neither coalition aligns neatly with one party. Labor leaders can support construction employment while progressive activists oppose energy consumption. Conservative landowners can reject development while Republican national leaders promote American AI capacity.

The Associated Press found similar alliances in Nebraska, Texas, Wyoming, New Mexico, and other states. Its national opposition analysis reported that 60% of Republicans and 53% of MAGA voters opposed a local data center in a July poll.

At the same time, President Donald Trump continued presenting data centers as sources of jobs, tax revenue, and strategic advantage. Republican House Speaker Mike Johnson warned that a federal construction pause could endanger American competitiveness.

Democratic leaders face their own contradiction. They often support union construction jobs and technology investment, yet their voters may prioritize climate effects, utility affordability, and corporate accountability.

Governor Shapiro embodies that tension. His administration initially promoted fast approvals and enormous AI investment. It later imposed stronger conditions on projects seeking state priority, permits, or tax benefits.

Pennsylvania lawmakers have also introduced a proposed statewide moratorium on hyperscale construction and permitting. Senate Bill 1359 includes both Democratic and Republican sponsors, although it had not received a vote by October 11.

The bipartisan character of the resistance does not mean every opponent shares the same objectives. One group may want strict utility rules, another may demand better siting, and another may oppose AI infrastructure entirely.

That diversity can make the movement difficult to satisfy. A developer might resolve water concerns without answering property-rights objections. A state might protect ratepayers while leaving noise, land use, or disclosure to local authorities.

It also makes the politics durable. Opposition does not depend on a single ideology that can be dismissed after an election. It is rooted in the physical effects of development and the process used to approve it.

For AI companies, the message is uncomfortable. Public enthusiasm for useful AI products does not automatically translate into support for the facilities that run them.

The cloud once felt remote and nearly weightless. Project Hazelnut makes it visible as land clearing, transmission corridors, industrial buildings, and negotiations over who bears the cost.

The $165 Million Package Cannot Resolve Every Risk

NorthPoint’s package is substantial, but its size should not be mistaken for independent proof that the project’s full costs are covered.

The package creates identifiable community benefits. Household grants would deliver direct value. Township payments could finance services, while workforce funding could help residents qualify for construction or technical roles.

NorthPoint has also tied its proposal to Pennsylvania’s responsible-development standards. Those standards emphasize local approvals, infrastructure planning, cost protections, labor considerations, and environmental expectations.

Supporters can reasonably argue that rejecting such commitments removes potential resources from a community. Data centers can generate construction activity, expand the local tax base, and support cloud services used across the economy.

The proposal may also establish a better baseline than developments offering vague promises. Written funding schedules and utility protections are easier to evaluate than general claims about future prosperity.

However, several uncertainties remain.

The campus lacks final local approval. Amazon has not confirmed itself as the end user. The construction schedule, final operating configuration, and precise long-term employment totals remain unclear publicly.

The $165 million also combines several kinds of value distributed over different periods. A grant available after the first occupancy certificate is not equivalent to money delivered before approval. Fifteen years of township payments are not a single immediate transfer.

Readers should therefore avoid treating the headline amount as a simple payment to residents. Only $45 million belongs to the proposed household grant fund. Another $15 million supports community initiatives, while $105 million would reach the township over time.

The energy protections also require scrutiny. Saying a project will pay for transmission upgrades does not establish how every generation, capacity, or system cost will be allocated. Those questions depend on utility contracts and regulatory decisions.

Environmental concerns likewise extend beyond a single commitment. Residents have raised questions about habitat disturbance, construction near contaminated sites, noise, water management, and cumulative industrial development.

Pennsylvania has begun examining those issues systematically. The bipartisan Center for Rural Pennsylvania opened a public hearing series covering siting, electricity demand, regulation, environmental effects, and local-government readiness.

That work matters because many townships lack the technical staff needed to evaluate hyperscale projects. A proposal can require expertise in zoning, utility regulation, engineering, tax agreements, environmental permitting, and emergency response.

Small governments may face experienced developers and large potential operators while depending on part-time officials or outside consultants. Financial benefits can increase the pressure to approve projects before all consequences are understood.

The strongest skeptical case is therefore not that every promised benefit lacks value. It is that benefits and safeguards must be evaluated independently.

A household grant does not validate a transmission route. A police fund does not confirm that utility customers are protected. An environmental commitment does not replace permit review. A possible Amazon connection does not guarantee stable long-term employment.

The reverse is also true. Community resistance does not prove that the project cannot operate responsibly. Some concerns can be addressed through enforceable conditions, route changes, monitoring, and public reporting.

The decisive question is whether those protections exist before approval and remain enforceable afterward. Trust cannot depend on voluntary promises that a future owner or operator can reinterpret.

This is where Project Hazelnut explained as a simple corporate-versus-community conflict becomes incomplete. The practical outcome will emerge from zoning decisions, utility cases, court proceedings, contracts, and state permitting rules.

Money can support a compromise, but only governance can make that compromise credible.

Three Signals Will Decide What Happens Next

The next phase will show whether Project Hazelnut becomes a model for negotiated development or a warning about deals offered after public trust disappears.

The first signal is Hazle Township’s approval process. The campus cannot advance on its proposed scale without clearing local land-use requirements and related legal challenges.

Approval with detailed conditions would strengthen NorthPoint’s claim that direct benefits and infrastructure protections can produce a workable agreement. Another delay or rejection would show that the $165 million package failed to address the community’s central objections.

The content of any approval matters as much as the vote. Watch for binding limits covering noise, construction hours, environmental monitoring, emergency planning, financial guarantees, and future expansion.

The second signal is the power-line case. PPL’s proposed 11.2-mile route connects the campus to landowners who may receive none of its direct advantages.

If regulators allow the line after documenting need, route alternatives, and cost protections, the project will clear a major infrastructure obstacle. A longer pause, route redesign, or denied application would weaken the campus regardless of NorthPoint’s community package.

That proceeding will also test a broader claim made throughout the AI infrastructure boom. Developers say large customers can cover the network investments they require without shifting costs onto existing users.

Regulators will need evidence rather than assurances. Their decisions can influence how other Pennsylvania utilities structure service for large computing loads.

The third signal is whether Amazon formalizes a relationship with Project Hazelnut. Public confirmation would clarify the project’s customer, strategic purpose, and likely operating expectations.

It would also make Amazon directly responsible for explaining the campus to residents. That accountability could improve transparency, but it could intensify opposition to Big Tech’s physical expansion.

If Amazon selects another location, NorthPoint would need to identify a different customer or demonstrate that the campus remains viable without one. Continued ambiguity would prolong the trust problem surrounding the proposal.

These three signals should be read together. A named customer cannot substitute for zoning approval. Local approval cannot eliminate the need for power infrastructure. A utility authorization cannot guarantee that the campus secures a tenant.

The outcome will influence development far beyond Hazle Township. Companies are learning that national arguments about AI leadership do not automatically prevail at a township meeting.

Communities want specific answers about land, electricity, water, noise, taxes, and enforceable responsibility. They also expect those answers before major decisions become difficult to reverse.

The Project Hazelnut data center fight has already changed the political calculation. A developer placed an unusually large offer on the table, but opponents kept asking who decided, who benefits, and who carries the risk.

Those questions now unite voters who disagree on much of American politics. The next test is whether developers respond with deeper transparency and enforceable limits, or simply larger checks.

Anyone tracking AI infrastructure should follow the township vote, the transmission case, and Amazon’s position. Together, they will reveal whether Pennsylvania’s backlash changes how data centers are built or stops this one entirely.