For the last twenty years, rural communities have fought to get connected. We mapped unserved homes, documented gaps, applied for grants, built fiber, and made the case, over and over, that geography should not determine whether someone can participate in education, health care, business, or civic life. That work is not finished. Affordable and reliable broadband still matters, and there are still people who need it. But something important is changing. The question facing rural communities is no longer only how we get broadband here. Increasingly, we should be asking what we intend to build with it.

For Michigan’s Eastern Upper Peninsula, I believe that question could define the next chapter of our digital economy. Broadband connects people and businesses to the Internet, but beneath the services we use every day is another layer of infrastructure most of us rarely see: data centers, cloud platforms, fiber routes, Internet Exchanges, electrical infrastructure, and the physical computing systems increasingly powering artificial intelligence. If broadband was about reaching the network, the emerging opportunity is about what happens where networks, computing, energy, knowledge, and communities begin to intersect.

Broadband Was the First Layer

A useful place to start is the 2024 Navigating the Digital Future: Neutral Internet Exchanges as Catalysts for Ecosystem Growth study from DE-CIX. The study examines the expansion of Internet Exchanges in the United States and the industry’s shift toward distributed, carrier- and data-center-neutral exchanges. In the neutral model, the exchange is operated independently rather than controlled by one carrier or data-center company, allowing networks to connect across multiple facilities and reducing dependence on a single vendor or location.

An Internet Exchange, or IX, allows independently operated networks to exchange traffic directly. Instead of traffic always traveling through distant intermediary networks, Internet service providers, content networks, cloud providers, universities, enterprises, and other participants can interconnect more directly. The Internet Society similarly describes peering as one of the mechanisms that makes the Internet’s “network of networks” function and identifies Internet Exchange Points as infrastructure that can improve performance, reduce costs, strengthen resilience, keep local traffic local, and support local hosting and technical communities.

There is a useful distinction here. Fiber makes connection possible, but interconnection determines where networks actually meet. For decades, rural broadband policy has understandably concentrated on getting fiber to communities, homes, businesses, schools, and institutions. Having made enormous progress toward that goal, we can begin asking what opportunities emerge when we also invest in the places where networks connect to one another.

That does not mean every rural region needs an Internet Exchange. The Internet Society makes an important distinction between simply constructing exchange infrastructure and nurturing a sustainable local interconnection ecosystem. It emphasizes voluntary peering, neutral governance, diverse interconnection arrangements, and strong technical communities. That is an important lesson for the Eastern Upper Peninsula. The objective should not be to buy an IX. The objective should be to understand whether we can cultivate an ecosystem that would make greater regional interconnection useful and sustainable.

AI Is Changing the Infrastructure Equation

Artificial intelligence makes this conversation more urgent because computing is placing rapidly growing demands on the physical infrastructure behind the digital economy. The 2024 United States Data Center Energy Usage Report, produced by Lawrence Berkeley National Laboratory with support from the U.S. Department of Energy, estimated that U.S. data centers consumed approximately 176 terawatt-hours of electricity in 2023, representing about 4.4 percent of total U.S. electricity consumption. The study projected that consumption could rise to between 325 and 580 terawatt-hours in 2028, or approximately 6.7 to 12 percent of U.S. electricity consumption.

The international trend points in the same direction. The International Energy Agency’s Energy and AI report examines how rapidly expanding AI and data-center infrastructure is changing electricity demand, energy security, investment, and energy-system planning. Taken together, this work makes clear that the growth of AI cannot be separated from the physical systems required to power, connect, cool, and operate computing infrastructure.

Those trends present legitimate challenges. Large computing facilities need electricity, transmission infrastructure, fiber, land, cooling, equipment, and workers. Communities have every right to ask hard questions about electricity rates, environmental effects, tax incentives, resource use, and whether the economic benefits of infrastructure projects remain in the places that host them. The answer should not be to pursue data centers simply because data centers are growing.

The more important lesson is that computing is becoming infrastructure. Increasingly, decisions about computing have to be considered alongside telecommunications, electricity, workforce development, land use, economic development, and community benefit. For rural communities, that opens a much larger conversation than broadband access alone.

What If We Turned the Map Around?

The Eastern Upper Peninsula is usually described as remote, and it certainly can appear that way when we measure our geography according to our distance from Detroit, Lansing, or Chicago. Networks, however, give us another way to think about geography. The EUP occupies a distinctive position within the Great Lakes, immediately adjacent to Ontario, with fiber systems, utilities, educational institutions, public infrastructure, businesses, Tribal Nations, and communities distributed across a large geographic area.

None of those characteristics automatically makes the Eastern Upper Peninsula the next great data-center market, nor do they prove that we should build an Internet Exchange here tomorrow. Those conclusions would run ahead of the evidence. But our geography and the infrastructure investments already underway justify asking a different question: instead of thinking only about how the Eastern Upper Peninsula connects to somebody else’s digital economy, what parts of the digital economy could we build here?

That could eventually include carrier-neutral interconnection facilities, regional content caching, edge computing, cloud connectivity, cybersecurity infrastructure, research computing, regional data platforms, AI infrastructure, or an Internet Exchange if network demand can support one. The point is not that every rural community needs all of those things. It is that rural communities can begin thinking beyond Internet access toward Internet infrastructure.

The broader interconnection market suggests that this is not an unreasonable question. Internet Exchanges continue to develop around the world as geography, data-center development, local content demand, and network architectures change. The lesson for the EUP is not that we should imitate a major metropolitan exchange. It is that the architecture of the Internet continues to evolve, and regions have an opportunity to examine where they might fit within that changing topology.

A Distributed Great Lakes Digital Gateway

I have been using the phrase Distributed Great Lakes Digital Gateway to describe what this larger regional strategy might eventually become. This is not a conclusion from the DE-CIX study or the Internet Society. It is my interpretation of what these larger infrastructure trends could mean for the Eastern Upper Peninsula.

The word distributed is important. I am not suggesting that the Upper Peninsula should attempt to recreate Northern Virginia or compete to host the largest possible hyperscale data center. Bigger is not automatically better, particularly when an infrastructure project consumes enormous quantities of energy without producing comparable local value.

A different model is worth investigating. Several strategically located regional facilities could be connected through diverse fiber routes, with different locations serving different purposes. One might support carrier-neutral interconnection, another research or AI computing, and another edge services, public-sector infrastructure, or regional data resources. Schools, colleges, universities, utilities, governments, Tribal Nations, telecommunications providers, businesses, cooperatives, and economic-development organizations could participate in different parts of the ecosystem rather than relying on a single organization to build and control all of it.

Some of the most important infrastructure in that model would not be inside a data center. It would be the fiber technicians, electricians, network engineers, cybersecurity specialists, data professionals, building-automation technicians, researchers, educators, and entrepreneurs capable of building and operating the systems. The economic-development opportunity is therefore not simply to host digital infrastructure. It is to develop the regional capacity to own, operate, understand, maintain, and create value from it.

That distinction matters. Hosting somebody else’s infrastructure and developing regional capability are very different economic-development strategies.

Who Owns the Infrastructure Everyone Needs?

That brings us to an institutional question deserving just as much attention as the technology: who owns the infrastructure everyone else depends upon? Rural communities have encountered versions of this problem before. Rural electrification and rural telecommunications both required communities to find ways to aggregate demand, pool resources, and build infrastructure that was difficult to justify one household or community at a time.

There may be a lesson in that history for the AI and digital-infrastructure era. Some portions of next-generation regional digital infrastructure might make sense cooperatively or through another shared, neutral institutional structure. That does not require government to own everything, nor does it mean replacing private investment. A neutral regional organization could instead provide common infrastructure through which private businesses, public agencies, Tribal Nations, educational institutions, utilities, cooperatives, nonprofits, and entrepreneurs can connect and invest.

At this stage, the principle matters more than the precise organizational form. Infrastructure intended to connect an ecosystem should not become another silo. The national growth of neutral Internet Exchanges reinforces that principle. Neutral infrastructure can increase network diversity and allow participants greater choice among carriers, facilities, and routes rather than making everyone dependent upon the same gatekeeper. From a broader public-interest perspective, neutral governance and diverse interconnection arrangements are also characteristics of sustainable local peering ecosystems. For rural communities accustomed to limited infrastructure choices, those principles deserve serious attention.

Start With the Evidence

None of this means we should begin by pouring concrete, buying servers, or announcing an Internet Exchange. We should begin with evidence. We need to understand where Internet traffic from the Eastern Upper Peninsula actually goes, where our major fiber routes connect, which networks serve the region, where those networks interconnect, where the nearest carrier-neutral facilities and cloud on-ramps are located, how resilient the routes connecting us to them are, and where adequate electrical capacity exists.

We should also examine what computing workloads could realistically make technical and economic sense here, what workforce would be required to support them, how our proximity to Ontario might influence a broader Great Lakes network topology, and what infrastructure gaps stand between the region we have today and the one we might want to build. Most importantly, we should ask what economic value could remain in the region rather than simply flowing through it.

Those are questions we can investigate with data. That should come before choosing a site, recruiting a hyperscale data center, or deciding that an Internet Exchange is the solution. If the evidence tells us the opportunity is smaller than we imagine, we should know that. If it reveals an opportunity we have not yet recognized, we should know that too. Regional strategy should emerge from understanding the system rather than chasing the latest technology.

Broadband Was the Beginning

For a generation, rural communities have fought for the right to participate in the digital economy. That work mattered, and it still matters. Broadband was never valuable simply because we could put fiber in the ground. It was valuable because connectivity gave people, businesses, institutions, and communities the ability to do things they could not do before.

Perhaps we are arriving at another one of those moments. We have spent a generation trying to bring the Internet to rural communities. The next generation gives us an opportunity to ask whether some of the Internet itself should begin to live here: not simply fiber passing through the region or households consuming services hosted hundreds of miles away, but computing, interconnection, research, cloud access, data infrastructure, technical skills, entrepreneurship, and businesses built around those capabilities.

Broadband was not the destination. It expanded what was possible. Our next responsibility is to decide what we want to do with that possibility. For the Eastern Upper Peninsula, the question I hope we begin asking is no longer simply whether the network reaches us. It is what we can build together, now that it does, at the intersection of networks, energy, computing, knowledge, and community.

That may be where the next chapter of rural Michigan’s digital economy begins.

Sources and Further Reading

  • DE-CIX. Navigating the Digital Future: Neutral Internet Exchanges as Catalysts for Ecosystem Growth. 2024. https://www.de-cix.net/en/about-de-cix/news/neutral-ixs-the-backbone-driving-the-growth-of-digital-ecosystems-in-the-us.
  • Internet Society. Internet Exchange Points and Local Interconnection Ecosystems. 2026. https://www.internetsociety.org/resources/policybriefs/2026/ixps-and-local-interconnection-ecosystems/.
  • Internet Society Pulse. IXP Tracker. https://pulse.internetsociety.org/en/ixp-tracker/.
  • Lawrence Berkeley National Laboratory. 2024 United States Data Center Energy Usage Report. December 2024. https://eta-publications.lbl.gov/publications/2024-lbnl-data-center-energy-usage-report.
  • U.S. Department of Energy. DOE Releases New Report Evaluating Increase in Electricity Demand from Data Centers. December 2024. https://www.energy.gov/articles/doe-releases-new-report-evaluating-increase-electricity-demand-data-centers.
  • Lawrence Berkeley National Laboratory. United States Data Center Energy Usage Report: 2025 Update. 2026. https://eta.lbl.gov/publications/united-states-data-center-energy-2025.
  • International Energy Agency. Energy and AI. 2025. https://www.iea.org/reports/energy-and-ai/.