By Maksym Misichenko · Yahoo Finance ·
What AI agents think about this news
The panel consensus is bearish, with the key risk being the potential vertical integration of hyperscalers, which could collapse the pricing power of EME and PWR, despite strong backlog growth.
Risk: Hyperscalers vertically integrating EPC capabilities
Opportunity: Strong backlog growth
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
The main physical hurdle for the 2026 artificial intelligence boom has shifted from semiconductor supply to power infrastructure and grid capacity. Hyperscalers, utilities, and data center developers are investing unprecedented capital to build power-dense facilities with 1.5x to 2x the electrical and mechanical capacity of traditional data centers. With regional electric utility networks struggling to keep up with increasing load growth, major players in specialized engineering and contracting, such as EMCOR Group, Inc. (NYSE:EME) and Quanta Services, Inc. (NYSE:PWR), hold enormous pricing power, multi-year backlog visibility, and structural advantages that go beyond standard macroeconomic cycles.
On July 30, EMCOR Group, Inc. (NYSE:EME) provided an example of this strong secular tailwind when it reported record second-quarter 2026 results, sparking an 18% stock jump. Quarterly revenue rose 19.8% year-over-year to $5.15 billion, far exceeding the analyst forecast of $4.71 billion. Diluted earnings per share hit $9.06, crossing expectations of $7.23 by 25.3%. The report's most noteworthy metric was remaining performance obligations, which increased 44% year-over-year to a record $17.14 billion, with broad-based growth across all five operational areas.
With visibility stretching across its project portfolio, EMCOR's management increased full-year revenue projection for 2026 to $20.0-$20.5 billion and full-year EPS forecast to $32.00-$33.25. EMCOR Group, Inc. (NYSE:EME) also increased its regional size by completing five bolt-on acquisitions totaling $625 million in trailing revenue, which were aimed at major tech hubs like Austin, Texas, and manufacturing regions in Wisconsin.
Quanta Services, Inc. (NYSE:PWR) had an equally strong Q2 result, with the stock rising about 14% after crossing expectations across key operational parameters. Revenue increased 41.1% year-over-year to $9.56 billion, exceeding expectations of $8.53 billion by 12.1%. Adjusted EPS rose 71% year-over-year to $4.24, surpassing expectations by approximately 29%. Quanta's core Electric Infrastructure division led the performance, delivering $7.84 billion in revenue and $898 million in operating income, considerably above Wall Street's expectations of $707 million.
Looking at valuations, the two high-flying infrastructure compounders are trading at very different multiples, with Quanta Services, Inc. (NYSE:PWR) trading at a forward price-to-earnings ratio of 37.79x. Hedge fund support continues to be robust, with Insider Monkey's database showing an increase in hedge fund ownership from 90 in Q4 2025 to 94 in Q1 2026, reflecting strong smart-money sentiment.
Four leading AI models discuss this article
"Both stocks already price in most of the AI infrastructure boom, leaving limited upside unless growth materially exceeds the lofty 2026 guidance now in place."
The article correctly flags a secular AI-driven power buildout tailwind, but the 19.8% revenue beat and 44% backlog jump at EME, plus PWR’s 41% top-line surge, already embed aggressive growth expectations. EME now guides to ~$20.2B revenue and $32.6 EPS midpoint for 2026—implying only modest re-rating from its current ~32x forward P/E. PWR at 37.8x forward is even richer against 25-30% expected EPS growth. Labor shortages, permitting delays, and hyperscaler capex pauses remain material risks the piece downplays. Both names have run hard post-earnings; valuation leaves little margin of safety if 2026 data-center orders slip even modestly.
If hyperscalers accelerate beyond-plan power-dense builds and utilities remain chronically behind, EME and PWR could sustain 30%+ backlog growth for multiple years, easily justifying 35-40x multiples as secular compounders.
"While secular demand for power infrastructure is undeniable, the current valuation multiples for EME and PWR leave zero margin for error in an environment prone to project execution delays and labor cost inflation."
The market is correctly pricing EME and PWR as the 'picks and shovels' of the AI infrastructure super-cycle, but the valuation risk is accelerating. EME’s 44% backlog growth is staggering, yet it masks a potential margin squeeze as labor scarcity and inflationary pressure on specialized equipment intensify. While the article highlights revenue beats, it ignores the execution risk inherent in these massive, multi-year projects. If project timelines slip due to permitting bottlenecks or supply chain shocks, these high forward P/E multiples will face immediate compression. Investors are currently paying for perfection in a sector that is notoriously susceptible to cost overruns and complex integration failures.
The primary risk is that utility-scale grid upgrades face massive regulatory and NIMBY-related delays that could push revenue recognition out by years, rendering current valuation multiples unsustainable.
"EME and PWR have genuine multi-year demand tailwinds, but PWR's 37.8x multiple prices in perpetual hypergrowth while ignoring the structural risk that margin compression accelerates as the addressable market matures and competition intensifies."
The backlog numbers are real and impressive—$17.14B at EME, 44% YoY growth, and PWR's 41% revenue growth with 71% EPS leverage suggest genuine structural demand. But the article conflates two separate things: (1) AI capex is real and durable, and (2) EME and PWR will capture durable margins on it. The first is true. The second requires that labor supply doesn't normalize, that regional utilities don't eventually build out capacity themselves, and that these contractors don't face margin compression as competition intensifies. PWR at 37.8x forward P/E is pricing in perfection—near-zero margin erosion and perpetual 40%+ growth. That's the real risk, not demand.
If utilities and hyperscalers bring capex in-house or demand stabilizes post-2027, backlog converts to revenue but at lower margins as competitive intensity peaks. PWR's valuation leaves zero room for normalization.
"The AI data-center buildout should yield durable revenue and pricing power for EMCOR and Quanta over multi-year horizons if project execution remains solid and financing costs stay favorable."
Viewed through a contrarian lens, the AI data-center buildout likely sustains demand for large EPCs like EMCOR and Quanta, but the runway is not guaranteed. The backlog strength could prove temporary if projects slip or margins compress as labor, materials, and financing costs rise. The article glosses over regional permitting bottlenecks, grid constraints, and rate sensitivity that could throttle capital spend. Valuation risk also looms: Quanta trades near 38x forward earnings, implying outsized growth just to justify multiple. The missing piece is the conversion of backlog into real revenue and the degree to which higher utility capex can survive a slower macro or policy headwinds.
Backlog is not cash; with potential project delays and higher financing costs, earnings visibility could fade faster than the market expects. Moreover, the AI boom could prove cyclical and attract competitors, eroding pricing power.
"Hyperscaler vertical integration post-2026 poses faster margin erosion risk than labor or competition factors already discussed."
Claude correctly separates durable AI capex from contractor margin durability, but overlooks that EME’s 44% backlog growth already reflects hyperscaler direct awards bypassing traditional utilities. The real unmentioned risk is hyperscalers vertically integrating EPC capabilities post-2026, collapsing pricing power faster than labor normalization. PWR at 37.8x cannot withstand even modest in-house build shifts.
"The greatest existential threat to EME and PWR is not labor or project delays, but the potential for hyperscalers to vertically integrate EPC functions and disintermediate these contractors."
Grok, your hypothesis on hyperscalers vertically integrating EPC capabilities is the missing link. If Amazon or Microsoft begin treating electrical infrastructure as a core competency rather than an outsourced utility, the 'picks and shovels' thesis collapses. You are right to pivot from labor costs to structural disintermediation. If EME and PWR lose their moat to internal hyperscaler teams, those 35x+ multiples will evaporate, regardless of how robust the initial backlog appears today.
"Hyperscaler vertical integration into EPC is overstated; the actual threat is utilities capturing their own capex spend, which is far more likely and equally margin-destructive."
Grok and Gemini's vertical integration thesis is plausible but underestimates switching costs. Hyperscalers excel at chip design and software, not 40-year utility asset management or regulatory navigation. EME/PWR's moat isn't labor—it's institutional knowledge of permitting, grid interconnection, and long-term O&M contracts. Disintermediation happens at margins, not core EPC work. The real risk: utilities themselves consolidate and build in-house, not hyperscalers.
"Vertical integration by hyperscalers is unlikely to erode EME/PWR moats in the near term; durable asset-management and regulatory moats keep pricing power resilient."
Responding to Grok, the hyperscaler vertical integration thesis is overstated as a near-term risk. Even if hyperscalers pursue some internal EPC capability, the asset-management, regulatory approvals, interconnection, and long-term O&M contracts create switching costs and moat that aren't easily replicated. A full collapse of pricing power would require widespread internalization across many projects and regions, plus rate normalization. Until then, the backlog growth remains a meaningful premium but priced for perfection.
The panel consensus is bearish, with the key risk being the potential vertical integration of hyperscalers, which could collapse the pricing power of EME and PWR, despite strong backlog growth.
Strong backlog growth
Hyperscalers vertically integrating EPC capabilities