NextEra’s latest move with Dominion is being framed as a mega-merger, but the more revealing story may be what the deal is really buying: a physical corridor for power, land, and grid capacity that can be turned into “data centre alley” at industrial speed. The headline number circulating in coverage—$420 billion—signals not just ambition, but a recognition that the next era of AI infrastructure won’t be constrained by chips alone. It will be constrained by electricity delivered reliably, at scale, to the exact places where compute demand is expected to cluster.
In other words, this isn’t only a corporate transaction. It’s an attempt to pre-build the bottleneck. And if the market is right, the bottleneck is about to get tighter.
What’s at the center of the plan is the idea of a dedicated stretch of infrastructure—an alley—where data centres can be sited close to generation, transmission, and distribution upgrades already planned or accelerated through the transaction. The “alley” concept matters because data centres are not flexible loads in the way many people assume. They are designed to run continuously. Their economics depend on predictable power availability, stable voltage and frequency, and fast recovery from disturbances. That means the grid requirements are not merely “more capacity,” but more capacity with the right characteristics, delivered with the right reliability standards and timelines.
For NextEra and Dominion, the strategic logic is straightforward: AI workloads are driving a surge in demand for electricity, and the companies that can convert that demand into permitted, financed, and grid-connected capacity will shape where the next wave of compute lands. The $420 billion figure—reported as the scale of the transaction—also reflects how expensive it has become to secure the full stack: generation, transmission, interconnection, land, and the regulatory approvals that determine whether a project can move from blueprint to megawatts.
But the unique angle here is that the deal appears to treat “data centre alley” as a system, not a collection of individual projects. That distinction changes everything about execution risk.
Why “alley” beats scattered buildouts
The traditional approach to data centre growth has often been incremental: a developer identifies a site, negotiates power, applies for permits, waits for interconnection studies, and then builds. In a world where AI demand is accelerating faster than grid timelines, that model becomes fragile. Interconnection queues can stretch for years. Transmission upgrades can be delayed by permitting, procurement cycles, or local opposition. Even when power exists somewhere in the region, it may not be deliverable to the specific node where a data centre wants to connect without costly upgrades.
A “data centre alley” approach tries to reverse the sequence. Instead of waiting for each facility to find its power, the infrastructure provider plans the corridor first—aligning grid upgrades, land strategy, and permitting pathways so that multiple facilities can plug in as demand materializes. This is closer to how industrial parks are developed: you don’t build one factory and then hope the roads arrive in time; you build the roads and utilities so the factories can arrive quickly.
That’s why the corridor concept is being described as central to the NextEra–Dominion arrangement. It suggests a coordinated effort to create a repeatable pathway for data centre development, reducing the time between “interest” and “operational capacity.” In AI terms, time matters because compute demand is not static. Training runs, inference rollouts, and enterprise deployments all have their own schedules. If power delivery lags, capital sits idle—or worse, the customer chooses another region.
The market has learned this lesson repeatedly. When power is uncertain, data centre operators hedge by building backup systems, negotiating long-term power purchase agreements, or shifting to locations with clearer grid readiness. But hedging is expensive. It also doesn’t solve the fundamental issue: AI economics increasingly depend on energy cost and reliability, not just on the availability of servers.
So the “alley” is a bet that the next wave of AI infrastructure will reward regions that can deliver both.
The $420 billion signal: not just consolidation, but capacity underwriting
Mega-mergers in utilities are often interpreted through a familiar lens: scale efficiencies, bargaining power, and regulatory leverage. Those factors may matter here, but the $420 billion magnitude points to something else: capacity underwriting.
When electricity demand rises sharply, the utility’s challenge is not only to generate more power. It must also move it. Transmission and distribution upgrades require long lead times—engineering, equipment procurement, construction, and regulatory approvals. Utilities also face financial constraints: they must fund projects before revenue arrives. If demand is uncertain, utilities hesitate to invest aggressively. If demand is clear, they can justify investment and recover costs through regulated mechanisms.
AI-driven load growth changes the demand clarity equation. Data centres are large, persistent customers. Their load profiles are relatively steady compared with many industrial users. That makes them attractive from a planning perspective, but it also makes them politically sensitive. Communities worry about traffic, water use, noise, and—most importantly—whether the grid will be strained or whether ratepayers will bear the cost.
A transaction of this scale can be read as an attempt to align incentives across the value chain: to ensure that the entity controlling the grid and the planning process has the balance sheet and regulatory posture to invest early, while also capturing the long-term value of delivering power to AI clusters.
In that sense, the merger is less about “owning” data centres and more about owning the conditions under which data centres can exist.
The grid reality behind AI’s appetite
To understand why the corridor matters, it helps to translate AI demand into grid terms. A modern data centre can draw tens to hundreds of megawatts depending on size and design. Multiply that by dozens of facilities and you quickly reach the scale where transmission constraints become the limiting factor. Even if generation exists, the grid may not be able to deliver it to the right location without congestion relief.
Utilities plan around load growth using forecasting models, but AI introduces a new kind of uncertainty: demand can spike based on announcements, funding cycles, and procurement decisions by hyperscalers and enterprise buyers. Unlike some traditional industrial loads that expand gradually, AI-related buildouts can accelerate rapidly once a customer commits to a campus.
That’s why “interconnection” is such a critical word in these discussions. Interconnection studies determine whether a generator or load can connect to the grid without violating reliability standards. For large loads, the studies can be complex and may require upgrades. If those upgrades are not planned in advance, the queue becomes a bottleneck.
A corridor strategy can reduce that bottleneck by coordinating upgrades ahead of time. It can also standardize the process for multiple developers, making it easier to allocate capacity and manage the technical requirements of connecting large loads.
There’s also the question of resilience. AI infrastructure is increasingly treated as mission-critical. Outages are not just inconvenient; they can be financially catastrophic. That pushes data centre operators to demand high reliability and fast restoration. Utilities, in turn, must ensure that the corridor’s design supports redundancy and stability, not just peak capacity.
This is where the “alley” framing becomes more than marketing. It implies a deliberate engineering approach to reliability and deliverability.
Regulatory path: the real timeline
Even if the strategic logic is compelling, the regulatory path will determine whether the corridor can be built on the timeline implied by the $420 billion figure. Utility mergers typically face scrutiny from regulators concerned about competition, consumer costs, and service reliability. In the US context, state-level approvals and federal oversight can both play roles, and the specifics vary depending on corporate structure and jurisdiction.
For a deal tied to AI infrastructure, regulators will likely focus on several questions:
First, who pays for the upgrades? If the corridor requires major transmission and distribution investments, regulators will want clarity on cost recovery mechanisms and whether ratepayers are protected from excessive risk.
Second, how will capacity be allocated? If the corridor is built to serve data centres, regulators may ask whether access is fair and whether the utility will favor certain customers or affiliates.
Third, what are the reliability commitments? Regulators may require performance metrics tied to outage frequency, restoration times, and grid stability.
Fourth, what about environmental and community impacts? Transmission lines and substations can face local opposition. Data centres themselves raise concerns about land use and water consumption. Even if the merger is primarily about grid infrastructure, the downstream effects will be part of the conversation.
The “alley” concept could help regulators because it offers a structured plan rather than a patchwork of individual projects. But it could also raise new concerns if it concentrates influence over where data centres can locate.
In practice, the regulatory timeline may be the most important variable. AI customers can move quickly, but utilities cannot. That mismatch is precisely why corridor planning is attractive: it attempts to compress the time between regulatory approval and physical buildout.
Capacity allocation: the hidden chessboard
One of the most consequential details to watch is how capacity will be allocated across regions. The phrase “data centre alley” suggests a geographic concentration, but AI demand is not uniform. Some areas will attract more hyperscaler investment; others may see more enterprise deployments. The corridor strategy could either reinforce regional specialization or create a new imbalance where one area becomes the default destination.
If the merged entity controls the corridor, it may have leverage over which developers get access to power and on what terms. That could shape the competitive landscape among data centre operators and among AI infrastructure providers.
There’s also the question of flexibility. AI loads can evolve. A campus built for training might later shift toward inference-heavy workloads, changing power density and cooling requirements. A corridor designed for one type of load profile may need upgrades to accommodate future changes. The best corridor strategies anticipate that evolution by designing for modular expansion and by maintaining pathways for additional capacity beyond the initial build.
Another factor is the relationship between long-term contracts and grid planning. Data centre operators often seek long-duration power purchase agreements to stabilize energy costs. Utilities prefer contracts that provide revenue certainty to justify capital expenditures. The corridor strategy could align these interests