Introduction
Electrical contractors find themselves at the center of one of the largest infrastructure investment cycles in U.S. history. Amid the deluge of data center construction announcements and the attendant infrastructure required to support them, demand for electrical contracting services is strong and growing.
History, however, offers a note of caution. The privately financed buildouts of new network capacity that came before this one – ranging from the railroads of the late 1800s to the fiber networks of the late 1990s – built well ahead of demand. The busts that followed were painful for the companies that financed them and the contractors that built them. Today’s buildout is larger relative to the economy than the fiber boom, increasingly financed with debt, and exposed to a handful of companies that make up a large share of the public equity market. With major IPOs on the horizon, this public equity concentration will only increase. As a result, a break in the AI narrative would be felt well beyond data centers.
For electrical contractors, none of this is an argument for avoiding AI work – supporting the buildout of a frontier technology is a tremendous opportunity, and no one can time the cycle perfectly. The question is how to be a valuable picks-and-shovels provider to the current buildout while maintaining a solid baseline business and exposure to other megatrends that – while not entirely uncorrelated with AI given the link between AI and overall macro at this juncture – represent attractive growth areas over the medium to long term.
Owners weighing an exit and buyers underwriting transactions in this sector need to ask: how durable are the demand drivers underpinning today’s earnings, how durable are the specific business models and end markets behind them, and who are the logical next buyers of these businesses when it is time for the next exit?
Below, we step through the current situation for electrical contractors, what past buildouts suggest about this one, what attributes make businesses in this sector stand out, and how the M&A landscape is taking shape.
Ground Truth
According to ServiceTitan’s analysis of user data, electrical contractors who were active on the platform in both 2024 and 2025 grew their average revenue year over year, with meaningfully outsized growth in certain states with high levels of data center activity:
Electrical contractor revenue growth in three data center states
Year-over-year revenue growth, 2025 vs. 2024, electrical contractors active on ServiceTitan in both yearsData centers and the power grids that support them require comprehensive electrical work – from new construction wiring through ongoing maintenance and testing. With little debate around the theoretical upside of the digital infrastructure buildout, investors are left to weigh the sustainability of current contractor performance and how best to underwrite companies in an industry where demand appears to be joined at the hip to a constantly evolving AI narrative.
Lessons from Past Buildouts
Relative to the size of the economy, the AI buildout is on pace to be one of the largest physical capital booms in U.S. history. Per Apollo, hyperscaler capex is expected to average ~1.9% of GDP over 2022–29 – ahead of the telecom and fiber boom of the late 1990s and second only to railroads:
The AI buildout is second only to the railroads as a share of the economy
Average annual investment as a share of U.S. GDP during each buildoutThe financing backdrop behind these buildouts is equally, if not more, relevant. Railroads and fiber networks were both privately financed and built by competitors racing for the same customers, and both were built well ahead of demand. Railroad construction exceeded 4% of GDP in 1870 and 1881; the Panics of 1873 and 1893 followed, and by the end of 1893 roughly a quarter of U.S. railroad mileage was in receivership (Bradstreet’s, December 1893).
In fiber, the amount of cable laid in the U.S. went from one million miles in 1996 to 10 million by 2000. By 2004, the cost of bandwidth had fallen more than 90%. As late as 2005, as much as 85% of U.S. broadband capacity was still unused (Goldman Sachs).
History tells us that in “arms race” style buildouts, the tendency is to overbuild, creating excess capacity and a “bust” or correction period. The AI buildout certainly has the characteristics of an arms race – whether that results in a bust remains to be seen.
Today, the pace of the AI buildout continues to accelerate. Consensus estimates for hyperscaler capex in 2027 have more than doubled in the past year:
Expectations for 2027 hyperscaler capex have more than doubled in a year
Consensus capex estimates, September 2025 vs. October 2026, $ billionsIncreasingly, these investments will be financed with debt. Capex at Oracle, Microsoft, Amazon, Meta, and Alphabet is projected to exceed their combined operating cash flow for the first time in 2026 (Van Nieuwerburgh), and Morgan Stanley estimates a ~$1.5 trillion external financing gap for data centers through 2028, ~$800 billion of which it expects private credit to fill. AI already accounts for about half of U.S. investment-grade bond net issuance this year:
AI accounts for about half of U.S. investment-grade net issuance this year
Share of 2026 year-to-date U.S. corporate bond net issuance, AI vs. non-AISeparately, per J.P. Morgan, a basket of 42 AI-related stocks has accounted for 60–80% of S&P 500 earnings growth, price returns, and capital spending since January 2024, and technology now accounts for roughly a third of U.S. real GDP growth. The IMF’s latest Global Financial Stability Report names an AI-driven market correction as one of three key risks to the financial system – noting that household exposure to equities is at historic highs, and that 14 circular financing deals among AI companies since 2023 paired ~$300 billion of equity and ~$650 billion of debt with ~$1.4 trillion of future revenue commitments. Per the BIS, the end of past investment booms has been associated with a slowdown in GDP growth of more than one percentage point on average – the largest of which came after the dot-com boom. AI and overall macro are now linked to such a degree that if the AI narrative breaks, the effects will not be contained to data centers and tech companies.
Turning back to contractors, these businesses will not be insulated simply because they do not own the assets they help build and maintain. The dot-com boom provides a relevant case study.
During the dot-com boom, the carriers that owned the fiber – Global Crossing, Williams Communications, Metromedia Fiber Network, and others – drove the buildout of internet infrastructure. When internet traffic fell short of forecasts, several of the largest carriers filed for bankruptcy in 2002. The contractors that installed that fiber did not own the networks, but they were still exposed. When carriers stopped spending, revenue at Dycom, MasTec, and Quanta fell 19% to 37% from peak, and it took four to eight years to get back. Here is how each was positioned going into the bust:
All three fiber contractors were hit hard; the most diversified held up best
Share of 2001 revenue by customer industry; calendar-year revenue decline and recovery, 2000–08
−19%5
−37%8
−30%4
Quanta’s electric power and gas work cushioned its fall, but with 44% of revenue still tied to telecom and cable – and telecom construction spending down 26% from 2001 to 2003 – it took five years to regain its peak. MasTec, with ~74% of revenue in telecom and cable, fell twice as far and took eight. Dycom, almost entirely telecom, is the exception – it fell less than MasTec and recovered fastest, helped by its 2002 acquisition of Arguss Communications and a rebound in cable work. In 2001, diversification softened the blow, but it did not replace the lost growth.
Revenue does not tell the whole story, however. Quanta, the most diversified of the three, carried significant debt – it had to bring in a new investor, First Reserve, at $3 a share and amend its agreements with lenders and noteholders to get through the bust period. The company’s debt balance was the cost of building its diversification: it had bought more than 80 contractors since its 1998 IPO, refinancing the bank borrowings behind those deals with $210 million of secured notes and $172.5 million of convertible notes in 2000 – at the top of the market – and when telecom earnings fell, so did its headroom under the covenants attached to that debt. MasTec had financed its own acquisitions the same way, and S&P cut its bonds to junk in March 2003.
Dycom took its own hit when a major cable customer, Adelphia, stopped work and filed for bankruptcy in 2002, but it entered the downturn with net cash, and that same year bought a competitor, Arguss Communications, with its own stock.
Six years later, the roles reversed. By 2008, Quanta held more cash than debt and earned most of its revenue from electric power and gas, and MasTec had pushed into energy. Dycom still earned three-quarters of its revenue from telecom – when housing and then the broader economy collapsed, its customers cut spending and its shares fell further than either peer’s:
Both end market exposure and capital structure play a role in navigating a downturn
Share price indexed to 100 at each market peak, monthly closes; revenue mix going into each downturnThese time periods provide evidence that both end market exposure and capital structure play a role in navigating a downturn. In 2001, Quanta was relatively diversified, but that diversification was built through an unsustainable approach to acquisitions and capital structure. In 2008, Dycom’s modestly leveraged balance sheet could not protect it from the impact of the GFC on its customers and end markets.
The theoretical contractor that would have best navigated both cycles would have paired a diversified business model with conservative financing. In other words, contractors exposed to a booming industry need the right business model, financed the right way.
The lesson for today’s contractors is not to avoid AI work, but to make sure it’s not the only thing holding the business up, and that the underlying capital structure of their business remains on terra firma while pursuing this work. The rest of this note looks at what that means in practice.
Peeling Back the Demand Story
The best electrical contracting services businesses can address investor concerns around earnings quality by maintaining exposure to attractive non-data center areas of the market and ensuring revenue is generated from programmatic new builds, retrofits, and maintenance services rather than episodic, one-off engagements. Companies with this profile begin to offer a compelling investment proposition – they remove the physical real estate risks of data center construction, retain the upside of providing a critical service to the AI infrastructure buildout, and provide programmatic exposure to other high-growth megatrends.
Exposure beyond pure-play data center demand
Data centers are far from the only source of opportunity for contractors – as one example, the power generation and transmission infrastructure being constructed to support AI requirements is likely to be relevant regardless of how data center demand ultimately plays out.
Per McKinsey analysis, up to 75 gigawatts of coal and other thermal generation capacity is expected to retire by 2030 as aging plants surpass 65 years of operation.
Up to 75 gigawatts of coal and other thermal generation capacity is expected to retire by 2030
Cumulative capacity of plants with retirement dates by 2030 and aging plants older than 65 years, 2025–30, gigawattsAs a result, current AI forecasts may not need to ultimately be realized for utility and high-voltage contractors to see robust demand. If data center demand shows up, the near-term risk is weighted toward underbuilding, and the pace of supporting infrastructure investment will likely accelerate. If it falls short, the same generation and transmission capacity being built for data centers has the potential to be used to replace capacity exiting an already tight grid and serve ordinary commercial and industrial load.
Unlike in 2001, contractors today may have somewhere else to go if AI demand slows. Every major segment of electrical contracting is forecast to grow through 2030, investor-owned utilities plan to spend ~$178 billion on transmission over 2025–28 (EEI), and the AI infrastructure buildout sits alongside myriad other megatrends unrelated to large language models (i.e., grid modernization and interconnection, industrial reshoring and domestic manufacturing, and building electrification and retrofits) – all competing for a limited pool of electrical labor.
Not every alternative offers the same protection, however. If a pullback in AI is severe enough to take the broader economy with it, cyclical markets like commercial construction and discretionary retrofits would slow alongside it. The work most likely to hold is regulated utility spending and maintenance that owners cannot defer. The following matrix summarizes how the major segments of electrical contracting compare:
Not all non-data center work offers the same protection
Relative assessment of electrical contracting segments; darker is more favorable| Growth outlook | Independence from AI | Recession resilience | Recurring or programmatic | |
|---|---|---|---|---|
| End market | ||||
| Data centers | High | Low | Low | Medium |
| Utility transmission and distribution | High | Medium | High | High |
| Power generation and interconnection | High | Medium | Medium | Low |
| Industrial and manufacturing | Medium | Medium | Medium | Medium |
| Health care and life sciences | Medium | High | Medium | Medium |
| Commercial and multifamily | Low | High | Low | Low |
| Type of work | ||||
| New construction | High | Medium | Low | Low |
| Repair, retrofit, and electrification | Medium | High | Medium | Medium |
| Service, testing, and maintenance | Medium | High | High | High |
Exposure to post-construction revenue streams
While new construction is expected to be the fastest-growing segment of electrical contracting through 2030 (6.9% CAGR, FMI Consulting), repair and retrofit will comprise the largest segment of the market (~43%) over that period. Service & maintenance will also hold a meaningful share.
New construction grows fastest, but most electrical spend goes to existing buildings
Cumulative U.S. electrical services spend by segment, 2026–30, $ billions, and segment CAGR, 2026–30Businesses solely focused on greenfields are not only constraining their addressable market by excluding ~63% of total industry spend, but are also forgoing an array of inspection, maintenance, and testing revenue streams that can be either recurring or re-occurring in nature.
The GFC provides a relevant example. EMCOR and Comfort Systems are two of the largest U.S. building systems contractors. At both, new construction revenue fell hardest in the downturn, and both came out of it with meaningfully more of their revenue in existing buildings and service:
New construction fell hardest in the downturn; both contractors came out of it with more service
Share of revenue by type of work, 2007 and 2015; and change in new construction revenue in the downturn
−35%new construction revenue, 2008–10
−22%new construction revenue, 2008–09
Neither company discovered service in 2008. EMCOR had been pitching facilities services for years as “an additional buffer against economic downturns,” and Comfort Systems already generated roughly 40%+ of revenue from maintenance, repair, and replacement. The downturn proved the point, and both leaned in afterward. EMCOR put its two largest acquisitions of the period into service businesses (USM in 2011 and RSI in 2013, ~$715 million combined), citing USM’s “excellent revenue visibility.” Comfort Systems made growing its maintenance base a priority, telling investors in 2009 that every $1.00 of maintenance work brings $2.50+ of repair and replacement work.
Not all existing-building work held up, however. EMCOR’s renovation and retrofit revenue fell ~32% from 2008 to 2010, nearly as far as new construction, as discretionary retrofit projects were deferred like new builds. The work that held up for these companies was the inspection, maintenance, and repair offerings that keep existing systems running. Today, every data center, substation, and plant built adds to the installed base that requires these services from electrical contractors.
Evidence of repeatable demand
Conventional wisdom has long held that project-based revenue is less investable than services revenue. While this may hold for truly episodic projects, not all new construction revenue is created equal.
Programmatic new builds – for example, via MSAs with hyperscalers (or the developers and general contractors building for one) – are successive projects for the same customer, released under an agreement that sets the commercial terms once. Each building is then issued as its own scope against that agreement. These arrangements carry considerations and benefits for contractors. On one hand, they are likely to increase concentration with a small handful of customer accounts. On the other, they improve revenue quality and predictability.
Credit rating agencies provide a lens into how MSA revenue is viewed from a financing perspective. In upgrading Dycom to BB+ in 2025, S&P credited a record $8.1 billion backlog for “good visibility into the company’s revenue,” and separately counted maintenance – more than half of revenue – as “a multi-year recurring revenue base.” MSA visibility is not recurrence; customers still have to release the work, and most agreements are terminable for convenience. Pike, a utility T&D contractor, did about 85% of its revenue under MSAs when taken private in 2014 and was still single-B at Moody’s in 2022. Programmatic demand nonetheless compares favorably to one-off project revenue, with the potential to create a more financeable business for leveraged transactions.
That said, concentration has limits. Buyers tend to compress multiples or shift value into earnouts when too much revenue sits with a single customer or a small handful of accounts, and MSAs may carry change-of-control provisions that give the customer the right to consent, reprice, or terminate when the business is sold. The most valuable version of this profile pairs programmatic work with a customer base broad enough that no single relationship has an outsized impact on the business.
M&A Opportunity and Activity
The electrical contracting services industry remains highly fragmented – per FMI Consulting, ~70% of market spending is captured by firms with less than $50 million in revenue. Given the breadth of specialized services that exist across the electrical contracting landscape, there are compelling buy-and-build opportunities that offer not just the benefits of density and scale, but significant room to cross-sell an expanding menu of services across customers.
A long tail of small firms captures most electrical contracting spend
Share of U.S. electrical contracting market spending by firm revenue; each square = 1% of spendWhile consolidation remains in its early innings, several sponsor-backed platforms have emerged in recent years. The following selected platforms provide an illustrative look at the landscape of sponsor-owned assets in the space:
Selected sponsor-backed electrical contracting services platforms
| Company | Sponsor | Year | Description | Selected add-ons |
|---|---|---|---|---|
 |  | 2026 | West Coast electrical installation, repair, and maintenance for health care, biopharma, and data centers | Scholes Electric (2026) |
 |  | 2025 | Multi-trade technical services for data center, infrastructure, and industrial customers | Stewart Electric (2025)*, Weifield Group (2025)*, Prism Electric (2026) |
 |  | 2025 | Electrical testing, maintenance, commissioning, and repair for utility, industrial, and data center customers | Power Quality Engineering (2026) |
 |  | 2025 | Construction, maintenance, and storm response for electric utilities | Not disclosed |
 |  | 2025 | Transmission and substation construction services for U.S. utilities and EPC contractors | Not disclosed |
 |  | 2025 | Electrical design, prefabrication, and installation for health care, manufacturing, education, and data centers | Hawkeye Electric (2026) |
 |  | 2025 | Northeast electrical services for utility, solar, commercial and industrial, and EV charging customers | NuWave Energy Solutions (2025), Horton Electrical Services (2026) |
 |  | 2025 | Commercial and industrial electrical construction and maintenance in the Southeast | Bright Future Electric (2021)*, Goss Electric (2024)*, Langford Electric (2025) |
 |  | 2024 | National electrical, technology, and specialty systems installation, and facilities services | Orlando Diefenderfer Electrical Contractors (2025), Bruce & Merrilees Electric (2025) |
 | | 2024 | Industrial electrical, mechanical, and automation services for manufacturing, power, and data centers | J&J Electric of Indiana (2025), Ruder Electric (2026) |
 |  | 2024 | Southeast electrical, mechanical, technology, and fire protection services | Regency Electric (2025), Vintage Electric (2026), IEM Electric (2026) |
 |  | 2024 | Commercial and industrial electrical contracting in Idaho, Utah, and the Mountain West | Rocky Mountain Electric (2024) |
 |  | 2023 | Electrical services and operations and maintenance for industrial, utility, and renewables customers | Not disclosed |
 |  | 2021 | Industrial automation and electrical services for e-commerce, data center, solar, and manufacturing | XS Telecom (2022), JSET Automated Technologies (2024), D&L Electric (2024) |
 |  | 2021† | Traffic signal, streetlight, intelligent transportation, and utility electrical infrastructure | Bear Electrical Solutions (2024), James D. Hinson Electrical Contracting (2025), MSL Electric (2026) |
Public strategics have already shown a willingness to pay for scale. Three of the largest electrical contractor acquisitions since mid-2024 were each priced at roughly 9x–11x EBITDA:
Public strategics have paid roughly 9x–11x EBITDA for scaled electrical contractors
| Acquirer | Target | Closed | EV | Adj. EBITDA | EV / EBITDA |
|---|---|---|---|---|---|
| Cupertino ElectricData center and solar, national | Jul 2024 | $1.54B1 | $155–175M2024E | 8.8–9.9x |
| Miller ElectricData center, Southeast | Feb 2025 | $865M | ~$80M2024E | ~10.8x |
| Power SolutionsData center, Mid-Atlantic | Dec 2025 | $1.95B | ~$201M2Trailing | ~9.7x3 |
- 1. Plus up to $200 million earnout.
- 2. Not disclosed; implied by Dycom’s stated multiple ($1.95B / 9.7x), trailing four quarters.
- 3. As stated by Dycom.
As early-2020s-vintage sponsor assets begin to move into their exit window, there exists a growing universe of relevant strategic and large-cap sponsor buyers for electrical contracting services assets:
Potential buyer pools for scaled electrical contracting platforms
| Category | Representative potential acquirers |
|---|---|
| Public infrastructure strategics |   |
| Large-cap sponsors with infrastructure or maintenance experience |    |
Connect With Us
If you’re thinking through the process of selling or acquiring an electrical contracting services business, we’d be keen to support your process. Contact me directly at brenden@cortlandt.ai with questions or to find time to connect.