Getty Images/iStockphoto
What CIOs need to know about community-first AI infrastructure
Data center construction is growing, leading infrastructure providers to adopt a community-first AI infrastructure approach, aimed at building trust and long-term viability.
AI demand is driving the planned construction of thousands of data centers across the U.S., but communities are voicing concerns about the impact on communities at large and their resources.
Because of this, more organizations are responding with a "community-first" AI infrastructure approach.
Microsoft is touting its community-first infrastructure, and other providers are taking the approach.
"We've been very hands-on with involvement in the communities," said Wes Cummins, CEO at Applied Digital, a digital infrastructure provider based in Dallas. "When we went into Ellendale [N.D.], our first large facility, we worked very closely with the community, and it was very welcoming."
Applied Digital's second campus, in Harwood, N.D., was also very welcoming, Cummins said. But the industry has started to garner more attention as it has grown and there has been more pushback -- not from the communities themselves, but from external naysayers.
Communicating with local community leaders and holding public forums early in the process helped make the projects successful. If done right, the project will benefit the community and create jobs without disrupting the community, Cummins said.
What community-first AI infrastructure means
Community-first planning involves considering the impact a data center will have on a community, its residents and resources. By working with local leaders and community members, companies hope to minimize the negative effects while creating the greatest possible benefit -- including job creation and tax base enhancement.
"It's such a perfect match, with what we are doing inside of these small towns, so the projects fit really well," Cummins said. "They have massive benefits to the community, but they're the right size."
Seeking community input early can allow for some tailoring of the project. Because they are being located in agricultural areas -- where water resources are crucial -- the data centers are designed with closed-loop water systems that enable data centers to use less water than the average household.
Applied Digital's approach to community-first AI infrastructure includes selecting sites in a way that benefits the grid operation of the region, investing in technology such as the closed-loop water systems and running the most efficient infrastructure possible.
"If you do all of these things the right way you create substantial benefits all the way around," Cummins said.
Why Microsoft is taking this approach now
In early January, Microsoft announced its community-first AI infrastructure initiative, which outlined five commitments the company made to residents of the communities where Microsoft locates its data centers.
According to Microsoft's blog, the company will make five commitments to be a "good neighbor" in the communities where they build, own and operate data centers:
- Electricity. Microsoft plans to make data centers more efficient and collaborate with utility companies to ensure that electricity cost of servicing its data centers isn't passed on to residential customers.
- Water. Microsoft's data centers will minimize water use and replenish more water than they use. This is done by investing in local water systems and water replenishment projects and by optimizing water usage for cooling systems.
- Job creation for residents. Microsoft plans to invest in partnerships to train local construction workers and expand its Datacenter Academy program.
- Increase local tax bases. Microsoft will not ask for tax abatements and will support policies to invest its added tax dollars in services that are important to the local community.
- Investment in local AI training and nonprofits. Microsoft will partner with schools, colleges, businesses and libraries to provide AI training at all age levels.
What this means for the future of AI infrastructure
As the demand for AI, cloud computing and digital transformation grows, so does the need for more capacity, faster processing and data centers.
"It's clear that AI has potential and provides a value proposition, so the industry keeps investing in it. But these data centers are big and the easiest places to build data centers have already been taken," said Sebastien Jean, CTO at Phison Electronics.
Tech companies need to be good corporate neighbors if they want to continue to receive permission to enter a community. In the future, if it becomes apparent the data centers are harming communities, there could be changes at the political or legal levels that make it difficult for data centers to operate, Jean said.
Instead of being reactive and trying to make changes after they've affected community members, tech companies are choosing to think of the community and not just focus on delivering technology results, he said.
"We have to really consider the impact we're having on communities and put some real thought into being a good citizen and community member," Jean said.
Community-first infrastructure approaches should be a collaborative effort, where community groups can reach out to IT companies with their concerns, he said. Organizations can deliver on their sustainability promises using the same systems-based approach that technology companies regularly use.
Practical implications for enterprise IT leaders
Ten years ago, most people were unaware of what data centers were or that internet service requires a physical location, said Jenny Gerson, senior director of sustainability at DataBank.
"This is new, that all of a sudden people are realizing what these buildings are, and obviously there's a lot more of them as people are trying to keep up with cloud and AI-based needs," Gerson said.
A data center must operate 24 hours a day to ensure institutions such as hospitals and police stations have uninterrupted access to their data. Community-first AI infrastructure initiatives aim to balance that need with the needs of communities.
All stakeholders must work together to address all the concerns and common interests, and all voices are heard from the start of construction, Gerson said.
DataBank uses sustainable building materials and focuses on renewable power as well, she said. The company ended 2024 with 62% renewable energy using a procurement and grid mix. Its goal is to be 100% carbon-free by 2030.
Along with greater visibility comes greater accountability, and there are several external frameworks for documenting the organization's sustainability efforts:
- Environmental Social Governance (ESG). Disclosures of a company's environmental, social and governance metrics.
- CP (formerly carbon disclosure project). Reports on environmental impact focused on climate change, water security and deforestation.
- GRESB (formerly Global Real Estate Sustainability Benchmark). An annual assessment that evaluates and benchmarks the ESG performance of global real estate and infrastructure investments.
"These are all either investor or customer-driven, and many of them are publicly available, so it's putting more information out into the public view," Gerson said.
More stringent frameworks exist that require organizations to commit to goals. For example, the Science Based Targets initiative [SBTi] requires target setting for carbon emissions.
Some states are adopting required reporting as the data center industry grows. It's expected that Europe will soon require reporting as well, and a lot of European investors require that documentation.
"More and more we're seeing investors that are requiring it, so it's becoming something that many companies have to fulfill," Gerson said.
Risks CIOs should watch
Increased community involvement brings additional risks. Miscommunication about shared community goals or failure to maintain an efficient data center could result in reputation damage that's difficult to overcome. It may even hamper an organization's efforts to locate additional data centers in other areas.
There are also internal risks. Organizations that can't document sustainability results may lose foreign and domestic investors who require the information.
Key questions CIOs should ask providers
Time to market is important when developing data centers -- especially if an organization is trying to capture the value of the current interest in accessing compute interest, said Brad Johnson, utilities engineering expert at Bentley Systems. Organizations need developers with a history of delivering projects on or ahead of time.
"The ones who succeed at that are the ones who have figured out how to have this positive interface with the communities that they're seeking to build in," Johnson said.
It's important to look for an organization that can provide examples where positive public consultation maintained or accelerated the timeline from initial concept to permit, he said.
Technology companies must understand that they're operating in a utility environment, not a tech company environment. They need positive interactions within the community to deliver the project on time and maintain support for future changes or expansion.
"The good data centers -- the ones really worth getting behind and entrusting -- are the ones that figure that out," Johnson said.
Julie Hanson is a freelance writer who has reported on local news across Massachusetts and New Hampshire.
