PWR — Quanta Services, Inc.

As of 2026-08-27, the narrative is split. Analysts are bullish (26 Buy/10 Hold, avg target $794.25, consensus Buy), and the Q2 beat/guidance raise were objectively strong. But the stock has fallen from $788.75 high to $621.83, RSI is 42.28, and price is below the 20- and 50-day SMAs, with roughly $137.5M of insider sales and no buys over 12 months. That combination leaves sentiment mixed rather than uniformly bullish.

Revenue growth accelerated sharply: YoY growth went from about 15.6% in Q3 2025 to about 26.3% in Q1 2026 and 41.1% in Q2 2026 ($9.56B vs $6.77B). GAAP diluted EPS rose about 94.7% YoY in Q2 2026 ($2.96 vs $1.52), and operating margin improved to about 7.3% vs 5.8% a year earlier. FCF also inflected: Q2 2026 FCF was $864M vs $155M in Q2 2025. The trend is an acceleration/inflection, not a one-off blip, supported by a $53.4B record backlog.

On trailing GAAP, PWR is expensive: as of 2026-08-27 the P/E was 70.3 and P/S was 2.85, well above 5-year averages of 45.9x and 1.63x. On 2026 adjusted EPS guidance midpoint of $16.70, the forward P/E is about 37x, which is less extreme but still a premium to industrials. PEG of about 2.0 reinforces the 'rich vs. its own history' label; the pullback has improved, but not cheapened, the stock.

The trailing figures understate run-rate earnings because GAAP EPS of $8.91 includes charges and amortization that adjusted EPS of roughly $16.70 for 2026 removes. So the market is paying about 37x forward earnings, not 70x trailing earnings. That means the stock is pricing in strong execution and multi-year grid/AI demand; if backlog converts at guided margins, the multiple can hold, but any margin miss or project delay would likely trigger multiple compression. The late-August drop was a multiple de-rate, not an earnings cut.

As of 2026-08-27, RSI stood at 42.28, price was -6.41% vs the 20-day SMA and -6.39% vs the 50-day SMA, but +6.73% vs the 200-day SMA. That is an intermediate pullback inside a longer-term uptrend, not an oversold capitulation. Support to watch is around the 200-day SMA (~$583); a reclaim of the 20/50-day SMAs would be the first technical stabilization signal.

• AI/data-center electricity demand is a secular driver; grid upgrades are a multi-year necessity
• Utility capex cycles and federal grid-hardening/renewable legislation support backlog
• Interest rates affect utility financing costs and the pace of capital projects
• Persistent labor shortages favor self-perform contractors like Quanta that own skilled craft labor
• High-voltage equipment supply constraints create opportunity for JVs like Hyosung HICO

• Q3 2026 earnings (late Oct/early Nov): execution against raised FY26 revenue guidance of $39.3-$39.7B and adjusted EPS of $16.45-$16.95
• Conversion of the $53.4B record backlog, especially large transmission and data-center projects
• M&A integration and capital allocation/debt reduction after the August 2026 senior notes offering
• New grid/AI contract awards or expansion of the Hyosung HICO high-voltage equipment JV

• Execution/integration risk from 41% revenue growth and multiple acquisitions; low ~6.2% TTM operating margin leaves little room for error
• Valuation/multiple compression risk: ~37x forward earnings and 2.85x sales leave no margin of safety if growth slows
• Balance sheet/liquidity: long-term debt of $5.42B, cash of only $506M, current ratio about 1.10, and D/E of 0.685
• Project timing, customer concentration, labor and materials cost inflation, plus $137.5M insider sales with no buys

I lean positive but cautious. PWR is one of the highest-quality ways to own the U.S. grid/AI electrification build-out, with record backlog, 41% revenue growth, and improving cash generation. But at roughly 37x forward earnings and with thin margins and a leveraged balance sheet, I would not chase strength; the risk/reward is attractive after a deeper pullback or once the technicals stabilize, with position sizing disciplined.

I used a justified forward P/E of 40x on the 2026 adjusted EPS midpoint of $16.70, giving $668. That multiple is above the ~37x market forward multiple to reflect record backlog and durable AI/grid demand, but below the 5-year average P/E of 45.9x to account for low margins, cyclical execution risk, and balance-sheet constraints.