Rather than killing data center construction projects, the American public is conditioning them on social license to operate. There are problems with procedure: local residents feel “railroaded” by timelines and decision-making processes that leave them with little or no influence. And this affects the quality of relationship between the community and the project together with its sponsors, investors, developer, and/or operator. There is a contractual way to improve the relationship by customizing new institutions for sharing decision-making between companies and communities, each suited to local context. This is the Good Neighbor Agreement, a type of community benefit framework. This post describes examples of this kind of agreement and provides assessment questions for investors, sponsors, developers, and operators.

Reacting to widespread opposition against data centers in recent months, leading hyperscalers have created a new vulnerability. They have overfixated their community relations on resource distribution (who spends what, who gains what, who loses what). But in prioritizing the delivery of benefits to communities in exchange for their permission to build, developers risk losing the broader culture struggle. In turn, this oversight could drive the business costs of conflict up, not down. A central problem for data center developers — one that operators could inherit — is that they are doing little as an industry to show the public how decisions about distributions of costs and benefits are being made, and neither are they emphasizing how community governance contributes to those decisions.

The Community-First AI Infrastructure initiative of Microsoft Corporation (NASDAQ: MSFT) is assessed for its capacity to preserve social license. Implications are discussed for investors and companies that design, build, own, and operate AI data centers.

There is growing perception in communities that the AI infrastructure build-out is “rigged” against them. To derisk community opposition and actually get to a constructive outcome, developers need separate empathetic approaches for addressing people who feel swindled and mobilizing supporters.

In June 2026, Oracle Cloud Infrastructure announced a redesign of Project Jupiter in response to community concerns and described the changes in the Santa Fe New Mexican, a newspaper reaching residents of Dona Ana County where the facility is sited. Since local opposition is de facto for this case and explicitly led to a redesign, it’s a good example to evaluate for strength of public relations strategy.

“The water consumption challenge for data centers is largely solved,” Josh Parker, CSO at Nvidia, told Axios in June 2026.

Nvidia had just announced on its company blog that its latest AI servers, when running in favorable climates, can run on liquid-cooling architecture that doesn’t need chillers, reducing a data center’s water consumption for cooling purposes to near zero.

Though Nvidia’s announcement called this “one of the biggest efficiency leaps in data center history,” let’s consider some possible adverse consequences from a public relations viewpoint. What could follow from this effort at mobilizing support for the AI infrastructure buildout by lauding Nvidia?

The conversation about AI data center build-out sites has started to focus on the communication dynamics driving the controversies. Companies and investors with financial stakes in AI infrastructure construction must have a strategy for managing material risks from community opposition. Insights into misinformation and opposition trends are given.