A proposed $420 billion merger between NextEra and Dominion is beginning to take shape as one of the most consequential corporate moves in the US AI buildout—less because it promises immediate consumer-facing change, and more because it would concentrate control over the physical bottlenecks that determine how fast data centers can be built, powered, and connected.
The deal, often referred to in early reporting and industry chatter as “Project Astra,” is framed around a simple but high-stakes reality: the modern AI economy runs on electricity. Not just any electricity—reliable, scalable power delivered on timelines that match construction schedules, permitting cycles, and the increasingly tight delivery windows demanded by hyperscalers and their supply chains. In that context, the phrase “data centre alley” has become shorthand for a set of US regions where new capacity is clustering. These are places where grid upgrades, transmission expansion, and interconnection queues are already under strain, and where the next wave of demand is expected to arrive faster than traditional utility planning cycles can comfortably absorb.
If the merger proceeds, it would not merely combine two large utilities. It would potentially reshape who controls the pace and structure of grid investment across a corridor that matters to AI infrastructure. That means the story is not only about corporate strategy; it’s about how essential infrastructure gets governed when the demand curve is steep, the capital requirements are enormous, and the political scrutiny is relentless.
To understand why this matters, it helps to look at what data centers actually require beyond “power.” Data centers need dependable capacity at specific voltage levels, with enough headroom to handle load growth without constant curtailment or costly retrofits. They also need interconnection agreements that are realistic about timelines. A hyperscaler can move quickly on servers and software, but it cannot simply “move quickly” on the grid. The grid is constrained by engineering, right-of-way acquisition, environmental review, equipment lead times, and—often most importantly—regulatory approvals.
That is where utility consolidation becomes more than a financial headline. When utilities merge, they can align planning processes, standardize procurement, and coordinate transmission and generation investments across a broader footprint. In theory, that can reduce friction and accelerate delivery. In practice, it can also concentrate decision-making power in fewer hands, raising questions about competition, rate impacts, and whether the public interest is being served as efficiently as investors expect.
Project Astra sits squarely in that tension.
What the proposed combination would change
NextEra and Dominion are both major players in US energy infrastructure, and each has its own strengths—generation portfolios, transmission capabilities, regulatory relationships, and operational expertise. A merger of this scale would likely be structured to preserve regulatory compliance while expanding the combined entity’s ability to plan and finance large projects.
For data center developers and hyperscalers, the most immediate question is whether the merged utility would be able to deliver capacity with fewer delays. In many markets, the bottleneck is not the existence of generation somewhere in the system; it’s the ability to move power to where it’s needed and to connect new loads without destabilizing the grid. Transmission upgrades can take years. Interconnection studies can take longer. Even when approvals are granted, construction schedules can slip due to supply chain constraints or legal challenges.
A larger, more integrated utility could, in principle, streamline some of these steps. It might also be better positioned to finance multi-year capital programs at scale, especially if regulators view the investments as necessary to meet demand growth. But the same integration could create a different kind of risk: if planning and prioritization become centralized, the allocation of capacity—who gets served first, how upgrades are sequenced, what terms are offered—could become more opaque. That opacity is not automatically harmful, but it becomes politically sensitive when the demand is driven by a sector that already attracts scrutiny for market power and energy consumption.
In other words, the merger’s impact would likely show up in the “how” of grid expansion: the sequencing of projects, the structure of interconnection commitments, and the degree to which capacity is treated as a public utility service versus a quasi-contractual asset negotiated with large customers.
Why “data centre alley” is more than a metaphor
“Data centre alley” is a useful phrase because it captures a geographic pattern: data centers are clustering where power infrastructure is relatively mature and where land, fiber connectivity, and permitting pathways make large-scale development feasible. These corridors are not uniform. Some areas have stronger transmission capacity than others. Some have faster permitting processes. Some face more complex environmental constraints. And some are already saturated with interconnection requests.
AI demand intensifies all of these differences. Training and inference workloads are driving not only incremental growth but also changes in how quickly capacity must be added. Hyperscalers increasingly treat data center buildouts as a rolling program rather than a one-time expansion. That means utilities face a moving target: load forecasts that can shift as model training strategies evolve, as hardware generations change, and as new product rollouts accelerate.
When demand is uncertain, utilities typically rely on forecasting models and staged investment plans. But AI-related load growth can be difficult to predict with the same confidence as traditional industrial or residential demand. That uncertainty can lead to conservative planning—building enough to meet near-term needs while reserving additional capacity for later. If the merged utility is perceived as having greater ability to plan ahead, it could reduce the risk of underbuilding. If it is perceived as overbuilding or using consolidation to justify higher rates, it could trigger regulatory pushback.
This is why the merger is being watched not just by investors, but by regulators, grid operators, and large customers who want clarity on timelines and costs.
The regulatory question: coordination versus concentration
Utility mergers in the US are rarely just about corporate efficiency. They are also about regulatory oversight and public interest obligations. State regulators typically scrutinize rate impacts, service reliability, and whether the merger will benefit customers. Federal authorities may also weigh in depending on the structure and the scope of transmission assets.
The central regulatory dilemma is straightforward: consolidation can improve coordination, but it can also reduce competitive pressure and concentrate leverage. In essential infrastructure, leverage matters. If a utility controls the path from generation to load, it can influence the terms under which capacity is delivered. Even if the utility remains regulated, the practical effect can be significant—especially when large customers negotiate long-term arrangements.
Supporters of consolidation often argue that scale enables better planning and faster execution. They point to the complexity of grid modernization and the sheer capital required to expand transmission and upgrade substations. They also argue that fragmented ownership can slow down decision-making, create duplication, and complicate project management across boundaries.
Critics counter that consolidation can lead to higher costs, less transparency, and a tendency to prioritize investor returns over customer affordability. They also worry that mergers can create “too big to fail” dynamics, where regulators feel pressured to approve deals to avoid disruption—even if the benefits are not fully proven.
Project Astra lands in this debate at a moment when AI-driven demand makes the stakes feel urgent. That urgency can cut both ways: it can speed up approvals if regulators believe the grid needs rapid investment, but it can also intensify scrutiny if the public perceives that the urgency is being used to justify market power.
How the deal could affect hyperscalers and data center developers
For hyperscalers, the value of a utility relationship is not only the availability of power but also the predictability of delivery. A data center is expensive to build and even more expensive to delay. If a utility can offer clearer timelines for interconnection and upgrades, it reduces the risk premium that developers otherwise price into projects.
A merged utility could potentially offer more standardized pathways for large customers, including clearer upgrade schedules and more coordinated planning across transmission and distribution layers. That could help developers secure financing and reduce uncertainty in construction schedules.
However, there is another side. If capacity becomes more tightly controlled, large customers may find themselves competing for priority access. Priority access can be beneficial if it reflects objective grid readiness and engineering constraints. It can be problematic if it becomes a negotiation game where the largest players secure the best terms while smaller developers face longer waits.
This is where the merger’s structure matters. If the combined entity commits to transparent capacity planning and publishes clear criteria for interconnection prioritization, it could mitigate concerns. If those criteria remain opaque, the merger could be interpreted as shifting leverage toward the utility and away from customers.
There is also the question of how the merged utility handles load growth beyond data centers. AI demand is not the only driver of electrification. Electric vehicles, heat pumps, industrial electrification, and broader decarbonization policies also increase load. Regulators will likely ask whether the utility’s investment plan balances these needs or whether data center demand crowds out other priorities.
The “AI revolution” framing can obscure that balancing act. Electricity systems do not care whether the load is for a factory, a home, or a server farm. They care about stability, reliability, and cost recovery. The merger will be judged on whether it improves outcomes across the board, not only for the most visible customers.
Capital intensity and the timeline problem
Even if a merger is approved, the grid does not transform overnight. Transmission projects require long lead times. Equipment procurement can be delayed by manufacturing capacity. Construction depends on labor availability and permitting. Legal challenges can stall projects. And the grid must be upgraded carefully to avoid reliability issues during transitions.
So what does a merger change in the short term? It can change planning horizons and financing structures. It can also change how quickly the utility can coordinate across regions and manage project portfolios. But it cannot eliminate the physical constraints of building infrastructure.
That means Project Astra should be evaluated on two timelines at once: the corporate timeline (regulatory approvals, integration planning, capital allocation decisions) and the infrastructure timeline (interconnection studies, transmission upgrades, substation expansions, commissioning).
If the merged utility uses the deal to accelerate planning and pre-position projects, it could shorten the time between demand signals and actual capacity delivery. If it instead focuses on integration overhead and longer-term restructuring, the benefits