Communities Have More Leverage Over the AI Buildout Than They Realize

Spending it on opposition wastes the moment.

Communities are skeptical of data centers.  They don’t know if they like the underlying technology, they certainly are not feeling the increase in power costs they attribute to this massive build, and they don’t feel like they need to provide tax assistance to the richest companies in the world.

That’s led to real impacts for how we can build out the infrastructure for the most powerful technological advancement of the last century and who benefits from this massive capital commitment to our future.  If we don’t answer those questions for communities in ways that are meaningful for them, our AI ambitions and its impact on economic growth, national security, and American competitiveness will suffer.  That will lead to the AI buildout taking more time, more money, and causing more pain than it needs to.

Developers face the most immediate risk to their ability to grow their businesses and build the projects that will create the data infrastructure of the future.  The secret is to not think of community engagement as an afterthought or a nuisance, but to recognize that communities are critical counterparties with real power and interests in negotiations.  While developers spend time and resources courting their tenants who pay the bills (the hyperscalers), they need to spend an equal amount of time thinking about productive relationships with gatekeepers of their ability to get the project off the ground in the first place. 

Like all negotiations, the first step is to understand the interests of the people across the table from you.  We must acknowledge the things communities are raising really scare people.  Major labor displacement plus increased power costs and decreased access to clean air and water is not a recipe for embrace from most American towns, counties, and cities.  It is not surprising then, that county commissioners in central Pennsylvania unanimously rejected a proposal to rezone eight hundred acres of farmland for a data center after months of organized resistance. Seattle, Denver, and other hubs of innovation are floating moratoria on data centers within city limits.

Researchers at the University of Michigan and MIT recently looked at 150,000 transcripts of local government meetings across 48 states. The share discussing data centers has grown ninefold since early 2023. Residents are showing up, and most of them are not pleased about what they hear. In a politically divided country, the instinct toward moratoriums is now spreading through Democratic primary politics and rural Republican county commissions alike.

There is precedent here for skepticism.  The history of major industrial buildouts in American communities, from petrochemical complexes to freight corridors to fossil generation to manufacturing at scale, is a history of communities accepting deals on terms set somewhere else and discovering that the costs landed locally and the benefits did not.

Today the math on who has the power and the mutual benefit is different than it ever has been. The structural conditions of this buildout differ from anything else the technology sector has had to navigate, and in ways that hand communities leverage to be equal participants in the process with titans in Silicon Valley. The path to the returns Silicon Valley has promised its investors now runs through county zoning meetings, state public utility commissions, state legislatures, grid operators, and a set of policy makers who are much closer to the people who live next to the substations than they are to the people wanting to build them.  In short, the ability of the largest and most profitable companies in history to continue to grow and return value to their shareholders runs, for the first time, through physical infrastructure.  Physical infrastructure is not built on computers in the Bay Area and Seattle, but in the suburbs and x-urbs of American cities from coast to coast.

This is also not the same community relationship developers are used to.  At Generate, we are very familiar with working with communities to create project alacrity by offering to buy the fire department an ATV or to build a fence for a neighbor of the farm where we will build a community solar project.  We understand the importance of supporting local community infrastructure like community colleges when we were trying to build an anerobic digestor.  Those types of transaction costs are the tools of the development trade.  This relationship is fundamentally different.  The impacts of these negotiations are much broader and much more resonant with the issues that are motivating a larger swath of Americans to pay attention and firm hard views on the subject.  Those views are animating the way people engage with their representatives and influence how they vote.  For a developer that means more headaches through cancelled projects, delays, and higher costs to do business.

Those views are forming, but are not yet fully developed.  Recent public polling found a split in respondants support and opposition for data centers in their communities with a healthy chunk of the public with their minds still open. The good news? There are ways to productively engage that lifts all boats. The bad news? Opposition and the forces that are informing them – both with the real concerns and with misinformation – are the loudest people at the party.  The time for developers and communities to create workable, mutually empowering models is right now. 

There are three paradigms that will define this moment.  Navigating them successful will determine the outcomes.

1. New Negotiation Partners

The AI buildout is physical and distributed. Software companies have spent forty years producing growth and returns for investors without building very much. That promise no longer holds. To produce the compute their business models now depend on, hyperscalers are building data centers that demand gigawatts in hundreds of communities across the country.  That translates to trillions of dollars in infrastructure investment by 2030. That capital cannot all land in one place. Improvements to power, water, and fiber infrastructure must be distributed across the map, often near population centers, in dozens of states.

This produces an asymmetry the technology sector has not previously faced. The Amazon HQ2 search of 2017 and 2018 was a bidding war between cities competing for one prize, and the terms of that competition got set by Amazon. The data center buildout is the inverse: hundreds of prizes that have to land somewhere, with the host community often holding more options than the bidder in any given local case. A community that turns down one data center may have a chance to host another but the developer who is turned down must absorb months of delay while they relocate, and that delay translates into real money. Communities that understand this have material negotiating power. Communities that do not are leaving it on the table.

The real opportunity is to be a community that is open to understanding the mutual upside of a data center with the developer and the hyperscaler (when they’re different) and being a firm and far partner in the negotiation to make sure interests are truly mutual.  The opportunity for developers and hyperscalers is to offer real value to communities that reflect the realities of the people that live there are understand that their cooperation and partnership are the road to quicker returns and a competitive advantage against your rivals. 

2. The Need for (and value of) Speed

Speed has acquired enormous quantifiable value to the companies building these facilities, and that value can be shared.

Standard data-center economics put traditional facility revenue at roughly $3.3M per MW per year. AI workloads run materially higher: A deployed gigawatt of AI inference capacity generates an estimated $10-12 billion per year in end-market revenue. Pulling that revenue forward by six months is worth roughly $4.2 to $5.4 billion, assuming 80-90% utilization (Occam’s Edge​). The same logic compounds across a developer’s portfolio of sites, where every project waiting in the queue benefits when the queue moves faster.

Some of that surplus accrues to the hyperscaler as competitive advantage in a race against other hyperscalers and they are certainly saying the right things to demonstrate they understand the role communities play in delivering that value. Anthropic’s recent commitment to absorb the electricity rate increases associated with its data center load follows similar commitments from Microsoft and OpenAI. Like in any good deal, the real upside should be split between the counterparties.  So, where the hyperscalers recoup billions through faster processes to permit and connect to the grid and developers get value in being able to deliver projects on time and on budget, communities should get real value at the scale of the savings as well.  Some examples:

• Bill credits funded by the project itself for residential and small business customers in the affected utility territory, calibrated against any rate impact the new load creates.
• Substation and transmission capacity sized for community load growth and resilience, not just for the data center.
• Distributed renewable resources and battery storage that improve local power reliability during heat waves and storms.
• Workforce pipelines for electricians, instrumentation technicians, and plant operators that survive past construction and tie into operations and the broader regional grid.
• Community benefit agreements with enforcement mechanisms, so what gets promised at the rezoning meeting is the same thing that exists in five years.

These are commercially priced terms a developer will accept, because the alternative is twelve more months of delay in a market where each month is worth nine figures. That is a more useful conversation with communities than the typical stock response that residents always see through about quickly expiring construction jobs in warehouses full of computers designed to replace their labor.  That is the current conversation.  It’s not productive but it can be shifted when developers look to communities as partners who deserve respect and value through negotiation. 

3. Power infrastructure that lifts everything connected to it

This is the biggest opportunity families and small businesses have ever had to offset the costs of building a truly reliable and affordable grid.  Communities that understand that and can negotiate in good faith with developers and hyperscalers can help build that future with the savings they provide by allowing for quicker connection to a clean and reliable grid. 

The American electricity grid is old. Large stretches of transmission infrastructure are operating well past design life. Distribution networks in many regions are vulnerable to weather events that are growing more frequent and more severe. Generation queues are clogged with renewable projects that cannot interconnect on usable timelines. None of this gets fixed without sustained capital investment, and sustained capital investment for grid modernization has been politically and financially difficult to mobilize for the better part of two decades.

The data center buildout creates a contracted revenue base, in the form of multi-decade power purchase agreements from credit-worthy buyers, that can finance grid upgrades.  There are many interesting ideas on how to do that at a national level but there are opportunities at a more local and regional level to use this new moment to create conditions for building the local resiliency that can benefit communities.  The same transmission corridor that connects a data center to a renewable resource zone connects everything else along that corridor. The same substation upgrade that supports a hyperscaler’s load gives the surrounding community headroom for new manufacturing, new housing, electrified transport, and resilient backup power. Battery storage can both smooth data center load profiles and backstop the grid in heat waves and storms.

Designed deliberately, the AI buildout becomes the largest grid modernization the country has seen since rural electrification. If we let inertia, opposition, and purity guide this build out we risk creating a petrochemical corridor that works against communities by creating increased air pollution from single-cycle gas plants on the cheapest available rural land, that costs residential ratepayers without any tradeoff and does nothing meaningfully to reinforce the grid.  In short, creating another instance of infrastructure build out that cost communities and benefit elites. 

Both outcomes are technologically available today. Which trajectory materializes depends on whether communities, regulators, and developers negotiate the terms of the buildout deliberately, on the front end, while leverage is still distributed.

The window

That is a strange moment, and it will not last. The companies needing to build are organizing fast through procurement standards, model legislation, financing structures, and political infrastructure to make sure their interests are represented at every level where decisions get made. Advocates need to help communities benefit from this moment, not be a reflexive “no” and convince communities they have nothing to gain.  It isn’t true and it won’t help the people climate, environmental, and community organizing groups claim to represent.  Communities that organize with equal seriousness, around what they want from this buildout rather than around whether they want it at all, will end the decade with infrastructure their grandchildren still benefit from.  That infrastructure can be clean, affordable, and reliable. Communities that spend the leverage on saying no will get what they always have: whatever is left over after capital sets the terms.  Developers that try to undercut productive efforts to build that kind of resilience in communities will end up losing.  Both sides need to realize that the relationship can be productive and valuable.