Designing Data Centers for the Communities and Natural Environments Where We Operate

When Oracle evaluates locations for a new AI data center, we start with a simple truth: we're joining a community. And we know as a new member of the community, how we design and build our AI data centers is critical to the long-term relationship we wish to establish.

We believe our AI data centers play a critical role in supporting America's leadership in the AI era. They deliver substantial economic benefits, including thousands of jobs both during construction and ongoing operations, community infrastructure upgrades, and ongoing tax revenue.

But jobs and economic benefits are only one part of how a community may view an AI data center. How we intend to care for the land, use water and mitigate environmental impacts is very important. We also recognize that communities have concerns about the potential impact of AI data centers on daily life. Our approach is to design our campuses with care from the start, taking into account land use, water and the natural environment.

For example, at our 672-acre Port Washington, Wisconsin site, 172 acres will be preserved or enhanced, including protecting wetlands and planting more than 2,000 native trees to increase biodiversity. At our Saline Township, Michigan AI data center, three-quarters of the site will remain farmland, wetlands, and open space. We also designed the data center campus so its buildings sit below the road level, along with ample setbacks and natural buffers from trees and landscaping. We use sensor-controlled, downward-facing lighting to reduce unnecessary light pollution while meeting our safety and security goals. It's also quiet-just 55 decibels at the property line-which is about the same volume as a normal conversation.

We've also heard from community members and local leaders who are worried about a large AI data center's impact on local water supplies. We share this concern. Water is a precious resource, and we take our responsibility to protect it very seriously.

To this end, our data centers use what are called closed-loop, non-evaporative cooling systems. These operate much like a home air conditioner, just at a larger scale. The system requires an initial fill of water, and that's it. At our site in Doña Ana County, New Mexico, where drinking water resources are especially on the minds of community members, the initial fill will be made with non-potable water drawn from an existing nearby well. It doesn't touch any of the community's drinking water. Likewise, we know communities in Michigan and Wisconsin are concerned about protecting the Great Lakes. So are we. Our closed-loop water system keeps us from ever needing to continually drawn water from those important natural resources for cooling purposes.

Across all our AI data centers, whether in the dry areas of Texas or New Mexico, or the Great Lakes states of Michigan and Wisconsin, our daily water use is similar to that of a typical office building. It's used only for the typical restroom, kitchen and janitorial services that you'd expect. It's more efficient for us, and better for the environment to use these closed-loop systems.

In Michigan, once complete, the data center's new stormwater management systems will reduce runoff into the Saline River to levels lower than current conditions. To help support the community near our AI data center project in New Mexico, we have committed $50 million to help modernize Doña Ana County's water system. And in Wisconsin our AI data center is helping fund the Valley Creek Corridor Revitalization Project to restore and protect local waterways.

While these design choices help minimize environmental impacts, we know that reliable and affordable energy are a priority for local communities. We pay our own way on energy so ratepayers' bills and electric grid reliability are never impacted by our AI data center. Oracle - not ratepayers - will fund any electrical infrastructure upgrades needed to service our campuses. We also rely on clean energy sources whenever possible, including wind and solar. Abilene, Texas for example is one of the windiest cities in America where the wind often goes underutilized as a power resource. Our AI data center in Abilene taps into the excess, low-cost renewable energy from the local grid, helping to stabilize the grid and utilize wind power that would otherwise be wasted due to low demand.

While these steps are important, we know the work isn't done. We are committed to protecting the land, water and natural environments in the communities where we build AI data centers, and being a good neighbor with the community.