It’s easy to talk about artificial intelligence as something abstract–software in the cloud, algorithms behind a screen, innovation happening somewhere far away. But for many Black communities across the country, AI is no longer theoretical. It’s loud. It’s physical. And it’s showing up next door in the form of massive data centers that demand enormous amounts of electricity, water, and land–often in communities that had little say in whether they arrived at all.
That reality helps explain why Microsoft recently felt compelled to promise something that it’s calling a “community first” approach to AI data centers. The announcement came amid growing local resistance from residents worried about rising utility costs, environmental strain, and whether AI data centers actually deliver long-term economic benefits to their communities.
Microsoft vice chair and president Brad Smith acknowledged what communities have been saying out loud: public trust is strained. In response, Microsoft laid out a set of explicit commitments aimed at easing that skepticism. The company says it will cover electricity costs so data centers do not raise consumer utility bills, work with utilities to expand power supply rather than strain local grids, reduce water use and replenish more water than it consumes, invest in local construction and operations training , expand AI education efforts, and–importantly–avoid asking communities for property tax breaks.
On paper, it sounds like progress. But the reason these promises are necessary tells a much deeper story about how Black and lower-income communities have once again found themselves carrying the weight of “innovation”–a pattern that long predates any “community first” pledge.
Across the South and Mid-Atlantic, resistance to data centers has been building for some time, as communities confront the growing infrastructure demands required to support these projects. In Memphis, residents raised alarms about environmental and health impacts tied to a data center owned by xAI, the artificial intelligence company founded by Elon Musk.
Similar concerns are emerging in South Fulton, Georgia. Residents are organizing against a wave of data center expansion they say is reshaping their neighborhoods without their consent, reviving familiar concerns about environmental racism and unequal land-use decisions.
In Maryland, Prince George’s County residents have pushed back against a proposal to build a massive hyperscale data center at the former Landover Mall location, prompting local lawmakers to pause permitting and explore stronger community engagement and land uses.
For residents in these communities, the math is simple. Data centers use massive amounts of electricity and water. That demand can drive up utility costs, strain aging infrastructure, and lock communities into long-term environmental tradeoffs. Meanwhile, many data centers create relatively few permanent jobs once construction ends, even as local governments are often pressured to offer tax incentives to secure them.
So when you hear stories about Black communities pushing back against data centers being built in their neighborhoods, it is not an anti-technology stance. It’s an honest assessment of how much these projects take from communities—and how little they tend to give back.
One reason this conversation feels new is because AI has been marketed primarily as innovation, especially by the current administration. We hear about smarter tools, faster answers, and revolutionary breakthroughs—but rarely about zoning boards, transmission lines, and thousands of tons of water required to facilitate them.
The truth is that AI is infrastructure. And like other forms of infrastructure in this country—highways, power plants, industrial corridors–has often been placed where political resistance is weakest. Too often, that has meant Black communities absorbing the costs while others reap the benefits.
Against that backdrop, Microsoft’s announcement matters because it signals that as community opposition grows louder and more organized, tech companies are being compelled to address the local consequences of AI infrastructure upfront–rather than after harm has already occurred.
But voluntary promises are not the same as accountability.
Whether Microsoft’s “community first” approach becomes meaningful will depend on what happens after the press release: how these commitments are enforced, whether communities have real power in negotiations, and whether transparency around energy use, water consumption, and economic impact is sustained over time. Without clear standards, protections risk becoming uneven–available to communities with leverage, but not to those historically ignored.
More broadly, this moment forces a question tech companies can no longer sidestep: if AI is becoming part of our everyday infrastructure, what does real accountability to the communities absorbing its impact actually require?
At its core, the backlash against AI data centers isn’t about rejecting the future. It’s about demanding fairness. Black communities are naming a familiar imbalance at the heart of AI’s expansion: the costs are local and long-term, while the benefits are limited or flow elsewhere. Transparency, cost containment, and real local investment are not optional demands–they are the baseline requirements when communities are asked to absorb a certain level of impact.
If AI is going to shape the next generation, it cannot be built on the same old patterns of extraction and exclusion. Microsoft’s announcement may mark a shift in tone, but the real test will be whether this moment leads to structural change or whether “community first” becomes just another promise communities are expected to trust without proof.
Danielle A. Davis, Esq., is Director of Technology Policy at Joint Center for Political and Economic Studies, Host of The Miseducation of Technology Podcast and a Public Voices Fellow on Technology in the Public Interest with The OpEd Project.