Per-MW pricing, regional variance, and cost drivers for owners scoping hyperscale & AI builds.
Salary benchmarks across the 14 mission-critical disciplines.
The buildout of AI and data centers is no longer just a technology story. It is now a power, infrastructure, financing, and delivery story.
That was the core message from a panel on powering the next industrial surge: energy demand is rising from multiple directions at once. Utilities are already managing grid decarbonization, renewable integration, and asset upgrades. Now, they must also plan for massive new AI-driven loads - often at a scale far beyond traditional commercial or industrial demand.
For construction, energy, and mission-critical leaders, the implications are immediate. This is not simply a question of adding more megawatts. It is a question of how to deliver generation, transmission, storage, and interconnection capacity fast enough, reliably enough, and affordably enough to support the next wave of development.
The panel brought together perspectives from utility operations, battery manufacturing, engineering and construction, and infrastructure finance. Taken together, their comments point to a broader conclusion: the bottleneck is no longer one technology. It is system coordination.
One of the most useful framing points from the discussion was that AI is arriving on top of an existing power transition - not instead of it.
Utilities are already being asked to:
Now add in hyperscale and AI data centers with extremely large power requirements, and the challenge becomes multiplicative.
That matters for owners, developers, and contractors because it changes how projects move through the market. In prior cycles, the question was often whether power would be available. Today, the question is more nuanced:
For mission-critical projects, this shifts energy from a late-stage utility coordination item to a front-end strategic constraint.
A particularly important point came from the utility perspective: the traditional utility model was not built for this pace.
One executive described a framework of simplify, speed, and scale. That phrasing is useful because it captures the operational reset underway across the sector.
Utilities are revisiting internal and external processes to remove friction. That includes:
For project sponsors, this suggests a new reality: the most successful utility partnerships will come from early, structured engagement, not last-minute demand requests.
Historically, some utilities added demand in fractional annual increments. That environment is gone in many growth markets. Large-load customers now expect infrastructure timelines closer to technology deployment schedules, not conventional utility pacing.
This is a major issue for construction leaders. If the power schedule lags the building schedule, the entire program can stall - even if sitework, shell, and MEP procurement are on track.
The panel highlighted that planning in 100 MW blocks is increasingly insufficient in some markets. Utilities are being asked to think in multi-gigawatt pipelines.
That has profound implications for:
For staffing and hiring leaders, this is where talent risk becomes inseparable from infrastructure risk. Scaling energy systems requires experienced people in grid interconnection, high-voltage electrical, commissioning, controls, energy storage, owner-side program management, and utility coordination.
The conversation around storage was one of the most practical parts of the panel. The key point was straightforward: batteries help match generation to demand more efficiently.
That matters in two ways.
Instead of sizing every part of the system to the absolute highest demand spike, storage can capture energy when it is available and discharge when needed. In theory and in many applications, this reduces the amount of capital that must be deployed solely to serve short-lived peaks.
For owners, that changes the economics of site power strategy.
The panel noted that data centers - particularly those supporting AI model training - can have rapid swings in demand. Those peaks and valleys occur on a faster cycle than many traditional industrial loads.
Batteries are well suited to that pattern because they can respond quickly to:
This is a critical takeaway for mission-critical operators. Storage is not just a renewable integration tool. It is increasingly part of the power quality and resilience conversation for large digital infrastructure.
Another useful insight from the discussion was how utilities evaluate storage. The value case is not one-dimensional.
The panel described three major categories of benefit:
Storage supports reliability by helping ensure sufficient power is available when needed. This is especially important as grids transition away from fully dispatchable legacy fossil generation.
As renewable penetration rises, some hours see very low - or even negative - pricing. Storage can charge during those periods and discharge when prices are higher.
Batteries can help stabilize systems with fast-changing loads or voltage/frequency fluctuations. For data centers, this capability may be as important as simple energy shifting.
The practical implication is that project teams should stop viewing storage as a single-line-item expense and start viewing it as a multi-function grid and facility asset.
Nuclear received significant attention, and for good reason. If AI demand keeps climbing while grids pursue lower-carbon generation, dispatchable low-carbon power becomes increasingly valuable.
The panel argued that nuclear capacity is expected to expand substantially through 2040. But optimism was paired with realism: the sector has a long record of cost overruns and schedule delays.
That honesty matters. Nuclear is often discussed in abstractions. The panel instead focused on what would have to change operationally.
Several themes stood out:
One speaker pointed to the need to learn from other sectors that have delivered large programs with tighter schedule and budget performance.
That lesson extends beyond nuclear. In mission-critical construction generally, projects accelerate when delivery teams lock in:
The message for owners is clear: speed does not come from starting faster; it comes from resolving uncertainty earlier.
One of the strongest practical lessons came from the battery manufacturing side. The speaker described a repeated pattern across large plant delivery: when teams broke ground before the design was sufficiently mature, later changes drove outsized schedule and cost pain.
That observation is highly relevant to data centers, advanced manufacturing, utility-scale energy, and mission-critical healthcare or life sciences projects.
Under schedule pressure, many organizations convince themselves that early mobilization is always the faster choice. But this panel reinforced the opposite: premature construction can lock in rework at a multiple of the original planning cost.
For project executives, that means:
This is not a theoretical warning. It is one of the most consistent root causes of delivery failure across large capital programs.
One of the biggest shifts discussed was the emergence of hyperscalers as active infrastructure finance participants.
Traditionally, energy infrastructure financing relied on utilities, public-sector support, classic project finance, or regulated capital recovery. But AI growth is introducing a new class of counterparties with:
According to the panel, these firms are not just buying power. They are helping underwrite:
This matters because financing often determines whether a project reaches final investment decision, not just whether the technology works.
When capital providers are also the load drivers, project development can move faster. But it also raises new questions:
For employers and owners in mission-critical sectors, the implication is strategic: energy partnerships are becoming part of development strategy, not merely a utility procurement process.
A recurring concern throughout the discussion was fairness. If AI and data centers consume a growing share of power, who pays for the generation, wires, substations, and upgrades?
The panelists suggested that the answer is evolving away from older economic development models. In the past, some data centers benefited from discounted rates. The newer approach described by the utility representative was very different: large-load customers paying for the transmission and generation needed to serve them through incremental tariff structures.
That distinction is important.
If done correctly, such arrangements can:
A cited example involved a large data center agreement in which the customer would fund its transmission and generation requirements, contribute to local education, use limited ongoing water, and generate substantial property tax revenue.
Whether every project will follow that model was not specified in the video, but the trend is notable: the industry is moving toward frameworks where hyperscale demand must more clearly pay its full freight.
The panel also touched on an issue many energy conversations underplay: power is only part of the community footprint.
Water use, land use, and local social acceptance all matter - particularly for AI campus development. One speaker described concerns about data centers taking premium sites and competing for resources. Another argued that modern data center development, if properly structured, can reduce those concerns through:
For owners and developers, this reinforces a practical lesson: community acceptance will increasingly depend on resource transparency and local value-sharing, not just corporate messaging.
One of the most underrated themes from the discussion was continuity.
Several speakers pointed to the damage caused when energy policy swings sharply over time. In sectors like nuclear, long gaps between projects can erode expertise, disrupt supply chains, and increase costs when markets try to restart.
That observation applies more broadly. Industrial capacity is not just steel and concrete; it is also:
Stable policy and visible long-term demand help preserve and rebuild these capabilities. Without that continuity, every new wave of projects starts with avoidable friction.
For hiring leaders, this is especially relevant. Labor shortages in mission-critical construction are not just a recruiting problem. They are often a policy and market consistency problem expressed through staffing.
For firms delivering data centers, advanced manufacturing, healthcare campuses, energy facilities, or other high-consequence projects, the panel points to several practical implications.
Waiting until late in site selection or design development is increasingly risky. Owners should evaluate:
Generation, storage, interconnection, and facility design cannot be handled in separate silos anymore. The best outcomes will come from coordinated planning across utilities, developers, engineers, contractors, and financiers.
In fast-growth markets, speed matters. But mature design still wins. Teams that lock design assumptions, standardize systems, and use modular or prefabricated solutions will likely outperform teams that rely on brute-force acceleration.
The industry may have enough capital and demand, but many regions do not yet have enough experienced people to deliver at this pace. The winners will be organizations that secure leaders who understand both mission-critical delivery and energy infrastructure interfaces.
The strongest thread running through the panel was cautious optimism.
Yes, the demand outlook is daunting. Yes, project schedules, supply chains, and affordability are real constraints. But the panelists also described a market with unusual momentum:
In other words, AI is not just increasing power demand. It is forcing a redesign of how energy infrastructure gets planned, financed, and delivered.
That is the real takeaway for construction and mission-critical leaders. The next industrial surge will not be powered by one technology alone. It will depend on better execution across the full value chain - from design maturity and financing structures to storage integration, policy stability, and workforce development.
The organizations that understand that early will be better positioned to build on time, secure capacity, and avoid becoming stranded behind the power queue.
Source: "The Future of Energy Infrastructure in the AI Era | Conference of Montreal 2026" - IEFA TV, YouTube, Jul 13, 2026 - https://www.youtube.com/watch?v=_qre_nD4lxo
