America is entering its biggest grid build-out since the post-war era. For two decades the network aged quietly, because there was no reason to expand it: demand grew just 0.4% a year between 2001 and 2020 as per US EIA data, and utilities spent their capital replacing what was already there. That period is over. AI data centers, reshored factories, and vehicle charging are adding load the grid was never sized for, and the constraint has shifted from generating power to physically delivering it. Equity markets are repricing the consequence: transformer and switchgear lead times run into years, contractor order books are at records, and a sector long treated as slow and regulated has become one of the market's most closely watched infrastructure trades.
The Grid's Quiet Decades
Every data center, every vehicle charging point, and every reshored factory depends on the same piece of infrastructure: the power grid. It is the least visible part of the energy system, millions of miles of transmission and distribution lines, substations, and transformers, and, for most of the past two decades, the least demanding. Electricity consumption in the United States barely grew, so the network carrying it was largely left alone.
The numbers explain why. Since industrialization became the forefront of growth, electricity demand was consistently high, and the years 1950s–1970s are an example of that. Between 1951–1970, net electricity generated grew at 7.7% CAGR. As the economy matured, the pace at which demand grew fell too. Between 1971–1990, demand grew relatively slowly at 3.4% CAGR. This period also marked the rise of outsourcing manufacturing to low-cost Asian countries, which also contributed to the slowdown. Finally, the rise of service-oriented and less energy intensive business models pulled demand down further during the 1990s. Between 2001–2020, demand grew by just 0.4% CAGR.
That calm period now appears to be ending. In four of the last five years, electricity growth has been 2% or more, a pace last seen in the early 2000s. The surge, especially post-COVID, has been shaped by three forces: a ramp-up in domestic manufacturing, electrification, and, most importantly, data centers. Depending on the forecaster, McKinsey and Bank of America, power demand is expected to grow 2.5% to 3.5% CAGR over the next 10-15 years.
Three Forces Behind Rising Demand
Reshoring fuels a manufacturing boom. Since the pandemic, the need for de-globalization has risen globally as countries look to avoid supply chain complications and reduce dependency on overseas production. The United States is no exception, having outsourced much of its manufacturing to Asia and Latin America. Construction of manufacturing plants advanced to $234 billion in 2024, double the average annual spending of $117 billion between 2016 and 2020.
Electrification, especially of vehicles, is another important source of demand. According to PwC, the number of EVs in the United States could grow almost twentyfold rising from 4.5 million units in 2023 reaching to 92 million units by 2040. That would mean EVs and related infrastructure could account for 9% to 12% of total grid capacity. As of 2023 PWC survey, 56% of large commercial and industrial customers had already started electrifying their fleets, and that share is expected to continue rising with another 30% already gearing up.
Data centers, probably the biggest driver. As per Goldman Sachs estimates, demand from data centers is expected to more than double between 2025 and 2027, rising from 31 GW to 66 GW. Their share of total electricity consumption could increase from 4.7% in 2024 to around 9.5%-15.3% by 2030, a near tripling from the current level, according to Berkeley Lab.
What This Means for the Grid Network
Building new generation or a new data center is the comparatively easy part. Connecting it to the grid is not. Much of the U.S. network was built between the 1950s and the 1970s, and as growth slowed after 2000, investment slowed with it, replacements happened reactively rather than proactively. According to a Bank of America report, 31% of transmission infrastructure and 46% of distribution infrastructure are now near or beyond their useful lives. Even in 2024, $63 billion went toward replacements against just $32 billion on new additions.
The deeper problem, though, is speed. The International Energy Agency notes that new generation or data center capacity typically takes one to five years to build, while the transmission and distribution infrastructure needed to connect it takes five to fifteen. The result is a queue: Lawrence Berkeley National Laboratory found roughly 2,060 gigawatts of capacity, about 1.5 times the country’s entire installed fleet, waiting for a grid connection at the end of 2025, with the median project now taking over five years to reach commercial operation. Supply crunch on the equipment side compounds the delay resulting in higher lead times. The IEA estimates global annual grid investment must rise roughly 50% by 2030, from about $400 billion today, simply to keep pace.
Where Value May Accrue
The market has noticed. The Nasdaq Clean Edge Smart Grid Infrastructure Index, the leading benchmark for companies that build and equip the grid, has outperformed the S&P 500 over both the past one and three years. But that outperformance is not evenly shared. Some constituents have re-rated sharply while others have barely moved, which suggests investors are not buying "grid exposure" as a blanket theme. They are distinguishing between companies converting the investment cycle into order-book growth and those merely sitting inside a rising index.
Generation shows the same split from the other side, only starker. For most of the past decade, power producers ran excess capacity into a market with little pricing power: cheap natural gas kept wholesale prices low, and coal plants retired faster than new capacity was built. Vistra, now one of the largest U.S. power producers, traded near $10 a share as recently as 2016 and posted only single-digit gains most years through 2022. Over the last five years, however, the stock has compounded at roughly 50% annually, rising to around $140 per share. The cushion behind that reversal is gone industry-wide: the U.S. now has only about 3% excess generating capacity above reliability requirements, and ERCOT and PJM, the two fastest-growing grids, have none to spare.
New capacity is also expensive to add. As the CEO of NextEra Energy noted, a gas turbine that cost $785/kW to build in 2022 costs roughly $2,400/kW today. Even at that price, gas can supply only about 75 GW of the roughly 460 GW the U.S. needs by 2030.
Power producers show what happens when demand catches up to supply. But the more durable exposure sits a layer below them, with the companies that build and connect the grid itself. Whichever generation source ends up meeting the next unit of demand, gas, nuclear, or renewables, it still has to move through the same wires, transformers, and switchgear to reach anyone. These companies also sell to a wide base of utilities and generators rather than to a single buyer, which gives them a broader revenue base than any single producer they connect.
Here are some notable players in the power grid supply chain, across different parts of the value chain.
Eaton Corporation (NYSE: ETN) sits further down the chain, making the circuit breakers, switchgear, and backup power systems that link utilities to the buildings and data centers that draw power from them. Its electrical backlog grew 43% YoY in Q2 2026, and the company disclosed a U.S. data center backlog equal to roughly 15 years of construction at 2025 build rates, up from 12 years. Quarterly revenue hit a record $8.5 billion, up 21%, and full-year adjusted EPS guidance was raised to $13.40–$13.60. Switchgear lead times now run 52-80 weeks, roughly triple pre-pandemic norms, which makes Eaton a genuine constraint on project timelines, which also means a slowdown in orders would take years to show up in revenue. While Eaton may appear to be cheapest of the bunch at an EV/EBITDA of 28.9 times, it also carries a huge debt burden in its balance sheet of ~$20 billion. Its operating margin, which have been consistently increasing over the last 4 years and inching towards 20% is a plus point.
GE Vernova (NYSE: GEV) makes the equipment at the generation end of the chain, from gas turbines to grid systems, and is a product of General Electric’s 2024 spin-off. Its Q2 2026 backlog stood at $176 billion, up from $150 billion at year-end 2025, after orders grew 88% YoY organically to $24.2 billion. Management is now guiding toward $200 billion by early 2027. What makes that backlog durable is scarcity: only three companies worldwide build heavy-duty gas turbines, and buyers now pay non-refundable deposits years before a price is even set. The same scarcity cuts both ways, since much of the backlog consists of reservations rather than firm orders. At an EV/EBITDA of 82.0 times, this is also the most expensive of the bunch. While one reason could be that it has the brand name and the biggest order book, it also has the least amount of debt at $2.8 billion with a solid cash balance of $13 billion.
Quanta Services (NYSE: PWR) does the physical work of building and maintaining transmission and distribution networks. Its Q2 2026 backlog reached $53.4 billion, up 49% YoY, with the electric segment accounting for 82% of the total backlog. Quarterly revenue rose 41% to $9.56 billion, and full-year guidance was raised to $39.3–$39.7 billion. Its real advantage is labor rather than order size: the constraint on rebuilding the grid is skilled workers, and Quanta is the industry’s largest employer of craft-trained electricians. The labor scarcity, however, pressures its own margins even as the backlog grows. Quanta Services currently trades at an EV/EBITDA of 33.8 times but also has almost double the amount of debt at $6.1 billion compared to GEV.
Valuation is an area where all three companies appear relatively expensive. The NASDAQ Clean Edge index trades at a P/E of around 29x, much lower than 74x for Quanta Services and 44x for Eaton. GE Vernova’s exceptional gain and tax benefit skews the P/E in its favour. The recent surge these companies have witnessed, and a consistent double digit order book growth also resulted in such inflated valuations.
Together, the three sit at different points of the same build-out: turbines, construction, and connection. None of them need to bet on which generation technology wins, only that more of the grid gets built, equipped, and wired up.
The opportunity is not confined to the United States. Grid upgrades have been underway for years in South-East Asia, and Europe faces rising investment needs as renewables account for a larger share of generation. In April 2025, Spain and Portugal faced country-wide shutdowns due to grid-related issues and have since raised their planned investments. Beyond the companies above, larger players such as Siemens, ABB, and Hitachi Energy operate globally at far greater scale.
What We Are Watching
Investor-owned utilities plan $1.4 trillion of capital spending through 2030, up from $1.1 trillion projected a year earlier, with 2025 spending reaching a record $204.1 billion and forecast to rise 17% to $238.8 billion in 2026, according to the Edison Electric Institute. What is far less certain is how much of it converts into energized infrastructure, and on what timeline.
Not every announced project will materialize. Data center timelines depend on regulatory approvals and equipment availability, clean energy manufacturing has seen repeated delays, and vehicle electrification faces a slowdown as subsidies taper. Berkeley Lab's analysis of connection queues found that, of all the capacity entering between 2000 and 2020, only around 13% has reached commercial operation. The typical wait is now around 61 months, against 36 months in 2015 and 22 months in 2008. Announced capacity, in short, is a poor predictor of delivered capacity.
- Order-book execution. A large backlog is only a claim on future revenue. What matters is the pace at which it converts to delivered, billed revenue. Slippage here is often the first sign that announced capacity will not arrive on schedule.
- Cash conversion and working capital. Where a company books a large share of revenue on credit, revenue recognized on the income statement can outrun cash collected. Receivables growth, days sales outstanding, and working capital swings are worth tracking alongside revenue growth itself.
How Evalueserve Can Help
Capital is already following the build-out. The First Trust Nasdaq Clean Edge Smart Grid Infrastructure Index ETF drew $3.1 billion of inflows in Q2 2026, its fourth straight quarter of inflows, according to Morningstar. Over the last 12 months it has attracted a total inflow of $7.5 billion. Institutions are moving too. Hodes Weill and Cornell University's 2026 survey of 142 institutions found energy has overtaken digital as their top sector priority, with about a third planning to raise energy. At the company level, 13F filings show the number of funds holding GE Vernova rose by more than 1,100 over the past year and Eaton's by about 400. The flip side is crowding: as a trade becomes consensus, the margin for error on execution narrows.
Evalueserve helps buy-side investors and asset managers separate announced opportunity from realizable opportunity, through deep analysis of business models, order book quality, and execution capability. We deliver robust financial models, disciplined assumptions, backlog-to-revenue conversion analysis, and scenario-based valuation frameworks that test what happens when project timelines slip, supported by domain experts across power markets, grid infrastructure, data centers, and industrials.
That is what makes this opportunity both attractive and difficult. The need for investment is visible, the capital is being allocated, and the beneficiaries span multiple parts of the value chain. What remains uncertain is the pace of delivery, and identifying the businesses able to convert a multi-decade capex cycle into sustained returns is where careful analysis earns its keep.
