There is a moment in every transformation when the battle lines become clear. In AI infrastructure, that moment has now arrived. For the last two years the headlines have been about megawatts, leases, GPUs, sovereign partnerships, and hyperscale announcements, but the real story sits beneath all of that noise. The bottlenecks that will determine where the next generation of colocation and edge facilities get built are no longer technical. They are interconnection, water, and local politics. Everything else is downstream.
I learned this watching the country’s most unlikely places turn into front lines. A farmer in Pennsylvania asking what a data center even is. A meeting in Arizona where residents packed a room to protest a facility’s water permit. A tribal community weighing the promise of economic renewal against the risk of losing autonomy. A Texas county preparing for winter storms while approving load that rivals entire cities. None of these stories look like Silicon Valley, but together they tell the truth about where AI infrastructure is actually being shaped. The future is not being decided inside GPU labs. It is being decided inside planning commissions, water boards, county supervisor meetings, and interconnection queues that stretch longer than the timelines for most private equity funds.
This is the new terrain for investors and developers. It is fragmented, political, and deeply local. It is also where the opportunity lives for those who know how to read it. The three battlegrounds below will define every colocation and edge site selection strategy between now and 2030, and ignoring them means underwriting blindfolded.
Interconnection Is the New Gatekeeper
Most investors still think land strategy begins with proximity to power lines. They think transmission and distribution look roughly interchangeable. They do not understand that being next to a line means nothing if you cannot actually interconnect to it, and they do not understand that queues have become the real choke point in the system.
Across the country, transmission-level projects face delays measured in years, not months. Developers are discovering that the line they thought they could tap is either fully subscribed or requires upgrades that cost more than the land. In the mid-Atlantic, PJM warned of unprecedented load growth. In Texas, ERCOT winter-storm risk appears in the same headlines as AI expansion. In Pennsylvania, utilities testified that Amazon’s data center load rivals the electricity consumption of Pittsburgh. This is the kind of context that changes how you value land.
Being adjacent to distribution offers quick interconnection but limits scale. Being adjacent to transmission offers scale but guarantees delay. The winning strategy is not guessing which is better. It is knowing which counties have already funded upgrades, which substations have room to grow, and which lines are scheduled for reinforcement. This is where Chevron’s entry into the sector becomes so important. Chevron is not building a data center. It is building a power plant to serve one. It understands that the shortest path to compute is generation, not entitlement.
Google’s decisions in Texas follow the same logic. Forty billion dollars of development is being steered toward counties where load can actually be absorbed. This is not a technology story. It is an interconnection story dressed as a technology story.
The Water Battleground Is Only Just Beginning
For years, water was considered a nuisance detail in data center planning. Cooling systems mattered but not enough to drive politics. That era is ending. Arizona’s Project Blue showed the country what happens when a state realizes data centers can undermine long-term water security. Public opinion shifted overnight. Residents turned a permitting issue into a political movement. Arizona regulators tightened standards. Governors weighed in. Suddenly water was no longer a backend engineering detail. It was a front-page political constraint.
Arizona is not an outlier. It is the preview. Every major AI facility that leans on evaporative cooling is walking into a public arena where water is sacred. States like Pennsylvania, where hydropower and river systems once felt endless, are now facing residents who worry that data center growth will stress local resources. Headlines in York County shifted from tax gains to concern over water tariffs. People see the cooling systems, hear the roar of fans, and imagine their utilities rising to subsidize corporate demand.
Some operators respond by shifting to liquid cooling. Others push efficiency narratives. Some emphasize on-site solar or nuclear partnerships to offset environmental criticism. But none of these responses matter if a community feels cornered by outsiders building infrastructure they do not understand. The political mobilization around water will shape permitting for the rest of the decade, and investors who ignore this will learn the hard way that cooling strategies determine entitlement timelines.
Local Politics Will Decide Everything
The most misunderstood battleground is the local one. Many investors assume that federal incentives and state policies drive the market. In reality, the decisive action takes place at the smallest scale. A township meeting in York County draws hundreds of residents concerned about noise, water use, and grid costs. A planning commission in West Manchester debates whether a golf course should become a server farm. A tribal nation weighs a data center proposal against questions of sovereignty, environmental stewardship, and long-term bargaining power. These decisions rarely reach national news, but they determine the fate of nine figure investments.
People underestimate how powerful a zoning board can be when it decides to draw a line around farmland. They underestimate how quickly small towns can organize when they hear a data center might consume the electricity of an entire municipality. They underestimate how a single county supervisor, like those in Arizona and Pennsylvania, can shift an entire state’s incentives policy by pushing back against large tax exemptions.
And they underestimate how federal permitting reform is slowly but surely changing the sequence of these negotiations. The shifts to NEPA review timelines, transmission siting rules, and environmental justice criteria add friction on one side and acceleration on the other. They create a new map of where AI infrastructure can be built cleanly, quickly, and defensibly.
This is where tribal land becomes so important. It sits at the convergence of sovereignty, energy resource potential, workforce opportunity, and federal partnership. Some tribes are leaning into data centers as an economic development strategy. Others see them as a threat. Investors who do not understand the cultural and political dynamics here are walking into a world they are not prepared for.
The Analyst Who Connects the Dots
These stories look local. A zoning fight in a township. A water permit in Arizona. A tribal negotiation in the Southwest. A winter-storm warning in Texas. A state incentive report in Connecticut. A community meeting in Indiana. Each story looks small, but together they form the real landscape of AI infrastructure in America. This is where my work sits. The through-line is that the future of colocation and edge development will not be won by capital alone. It will be won by those who can translate local friction into national insight.
My thesis is simple. The GPs who win from 2026 to 2030 will master fragmentation rather than fight it. They will understand that interconnection queues tell you future land value. They will understand that water politics tell you which cooling technologies will attract or repel a community. They will understand that zoning battles tell you which counties are future clusters and which are future dead ends. They will understand that incentives are shifting from tax giveaways to cost protection for residents. And they will build portfolios that align with this new reading of the map.
The national story is not about AI growth. It is about who can navigate the constraints that growth creates. Those constraints will decide the winners. Those constraints will decide the clusters. Those constraints will decide which developers shape the next decade and which ones get stuck with stranded projects. This is the new reality. This is the battleground. And this is where the opportunity sits for anyone willing to analyze the country one county at a time.