The Shovel Sellers of AI

Where Capital Compounds Best: Comparing Quanta Services, Comfort Systems USA, Astera Labs, and Vertiv.

The AI gold rush isn't just about chips and software. The real money might be in the picks and shovels: the power systems, cooling infrastructure, and connectivity that make AI data centers possible. Four companies sit at different points in this value chain, each claiming exposure to the AI infrastructure boom. But which one has the moat that matters? And more importantly, which can sustain 15%+ revenue growth for five years while generating exceptional returns on capital?

Let's dissect Quanta Services (PWR), Comfort Systems USA (FIX), Astera Labs (ALAB), and Vertiv (VRT) through the lens that separates winners from pretenders: ROIC, competitive positioning, and what the market is actually paying for future growth.

The ROIC Scoreboard: Where Capital Goes to Multiply

Return on Invested Capital (ROIC) tells us how efficiently a company turns capital into profits. But the real insight comes from ROIIC—Return on Incremental Invested Capital. This reveals whether a company's moat is widening or eroding. A rising ROIC over time signals that each new dollar invested generates even more profit than the last.

Comfort Systems USA (FIX): The ROIC Champion

• ROIC: ~30.8% (5-year average)

• ROIIC: Rising (evidenced by ROIC expansion)

• The Story: FIX isn't just efficient—it's getting more efficient. With a 5-year ROIC consistently above 30%, FIX stands out as the capital allocation master. Even more telling: their ROIC has been expanding, hitting record territory as they scale. This suggests their moat is widening, not just holding steady.

• 2024 Performance: Q4 2024 same-store revenue growth of 22%, EPS up 60% year-over-year, and EBITDA margins of 12.7%. The company is turning down work due to capacity constraints—a sign of pricing power and selectivity.

Vertiv (VRT): The Margin Expander

• ROIC: ~16-18% (improving)

• Operating Margin: 21.5% (Q4 2024, up from 17.7% prior year)

• The Story: Vertiv has undergone a dramatic transformation. From a struggling infrastructure player to a margin machine, VRT's ROIC has climbed from mid-single digits to high-teens. Q4 2024 adjusted operating margin of 21.5% represents a 380bp expansion year-over-year—extraordinary for a hardware business.

• 2024 Performance: Revenue up 17% to $8B, but the real story is profit expansion. Operating profit surged 60% in Q4, demonstrating powerful operating leverage as scale kicks in.

Astera Labs (ALAB): The Growth Story With Question Marks

• ROIC: ~5.1% (as of Q1 2025)

• Gross Margin: 74.9% (semiconductor-class margins)

• Operating Margin: ~7% (still ramping)

• The Story: ALAB has fabulous gross margins but weak ROIC—a classic high-growth startup profile. They're investing heavily (negative free cash flow) to capture market share. The question: Will scale translate to ROIC expansion, or will competition compress margins before they achieve critical mass?

• 2024 Performance: Revenue exploded 242% to $396M, with Q2 2025 showing another 150%+ growth. But ROIC remains anemic, suggesting the company is prioritizing land-grab over profit optimization.

Quanta Services (PWR): The Execution Machine

• ROIC: ~7.8% (TTM), but mixed signals

• Revenue Growth: 17.5% (Q3 2025)

• Backlog: Record $39.2B

• The Story: PWR has scale and diversification, but ROIC is unimpressive for a company trading at 60x P/E. The low ROIC reflects the capital-intensive nature of infrastructure contracting. However, their record backlog and consistent execution suggest they're capturing AI infrastructure spending at scale.

• 2024 Performance: Full-year 2024 revenue of $23.7B (up 13%), with 2025 guidance calling for continued double-digit growth. Adjusted EPS growth of 25% shows operating leverage kicking in.

Winner: Comfort Systems USA. FIX's 30%+ ROIC dwarfs the competition, and the fact that it's expanding means their moat is strengthening. This is the profile of a business that can compound capital at exceptional rates.

Moat Analysis: Who Has Pricing Power That Lasts?

Comfort Systems USA: The Specialized Integrator

Moat Type: Niche expertise + local relationships + brand

FIX has built an enviable position as the go-to HVAC and electrical contractor for complex data center projects. Their moat stems from:

• Technical specialization: Data centers require precision thermal management. FIX has mastered liquid cooling systems and high-density power distribution—skills that can't be commoditized overnight.

• Installed base advantage: Once you're in a hyperscaler's vendor network, switching costs are high. FIX works with all major cloud providers.

• Capacity constraints driving pricing: FIX is literally turning away work. Their Q3 2025 backlog hit $9.4B (up 65% YoY), and management noted they're being selective on projects—pricing power in action.

Is it expanding? Yes. Same-store revenue growth of 33% in Q3 2025 + margin expansion (gross margin 24.8%) indicates they're capturing more value per project. The data center electrical segment is becoming a larger share of revenue, and these projects command premium pricing.

Vertiv: The Incumbent with Ecosystem Lock-In

Moat Type: Installed base + integrated solutions + scale economies

Vertiv's moat comes from being deeply embedded in existing data center infrastructure:

• Rack-level integration: VRT doesn't just sell power or cooling—they provide integrated rack solutions that tie power, cooling, and monitoring together. This creates stickiness.

• Service revenue: 20%+ of revenue comes from services and lifecycle management. Once installed, these systems need ongoing maintenance—high-margin recurring revenue.

• Partnership with NVIDIA: VRT collaborated with NVIDIA and CoreWeave on GB300 NVL72 deployments, positioning them as the default infrastructure provider for cutting-edge AI systems.

Is it expanding? Absolutely. Orders up 60% YoY in Q3 2025, with sequential acceleration. Their acquisition of Great Lakes (rack manufacturing) signals a move to capture more of the white space value chain. Operating margins expanding from 17% to 21% in one year shows they're flexing pricing muscle.

Quanta Services: Scale + Diversification = Stability, Not Dominance

Moat Type: Scale economies + project execution + regional presence

PWR's moat is real but less differentiated:

• Scale advantages: They can mobilize massive workforces quickly and handle billion-dollar projects. But scale in contracting is less defensible than in tech—labor and equipment are widely available.

• Diversification: Exposure to electric grid, renewables, and data centers spreads risk but also dilutes focus. Their underground utility segment underperformed in 2024.

• Backlog as a buffer: $39B in backlog provides revenue visibility, but backlog isn't moat—it's just deferred revenue.

Is it expanding? Mixed. Electric segment is booming (data centers, grid modernization), but the underground infrastructure segment faces margin pressure. Overall, the moat is holding but not widening materially.

Astera Labs: The Innovator's Dilemma

Moat Type: Technology lead... for now

ALAB's moat is the most fragile:

• First-mover in PCIe/CXL connectivity: They've shipped billions of units of PCIe retimers and are positioned in the Scorpio fabric switch market. But this is a technology moat, not a structural one.

• Hyperscaler lock-in: ALAB is embedded in AI server designs at major cloud providers. Switching costs exist but aren't insurmountable.

• Competitive threats: Broadcom, Marvell, and even custom silicon from hyperscalers pose existential risks. The recent ESUN initiative (backed by OpenAI, NVIDIA, AMD, Broadcom) could sideline ALAB's switch business.

Is it expanding? Unclear. Revenue is exploding (242% in 2024, 150%+ in early 2025), but market share dynamics are shifting. ALAB has innovation velocity but lacks the installed base moat that protects incumbents. Operating margins of 7% vs. 18% industry average suggest they're buying growth, not earning it through defensibility.

Winner: Vertiv. VRT has the strongest combination of installed base, ecosystem lock-in, and pricing power. FIX is a close second, but their moat is more dependent on specialized labor (replicable over time). ALAB's technology moat is weakest—disruptable by well-funded competitors.

Revenue Growth: Who Can Sustain 15%+ for 5 Years?

The test: Can these companies maintain 15%+ annual revenue growth through 2030?

Astera Labs (ALAB): Explosive... Until It Isn't

• 2024 Growth: 242%

• 2025 Projection: ~50-60% (based on current trajectory)

• 5-Year Sustainability: Low

ALAB's growth is spectacular but from a small base ($396M in 2024). Consensus sees $511M in 2025 and $703M in 2026—a deceleration from 242% to 30-40% CAGR. By 2028-2030, they'll be lapping tough comps. Add in competitive threats (ESUN, Broadcom), and 15%+ growth beyond 2027 seems unlikely.

Verdict: Blazing hot for 2-3 years, then cooling to mid-teens or lower.

Comfort Systems USA (FIX): Riding the Data Center Wave

• 2024 Growth: 35%

• 2025 Projection: High single-digit to low double-digit (guidance suggests tougher comps)

• 5-Year Sustainability: Medium-High

FIX is in the sweet spot. Data center spending on thermal and electrical infrastructure is structural, not cyclical. Management guided for "high single-digit percentage growth in same-store revenue" for 2025—conservative after a blockbuster 2024. But with $9.4B in backlog (up 65% YoY) and ongoing AI infrastructure build-out, low-teens growth through 2028 seems achievable.

Verdict: 12-18% CAGR through 2028 is plausible, then likely moderation to 8-12% as the market matures.

Vertiv (VRT): The Dark Horse

• 2024 Growth: 17%

• 2025 Projection: 25% (guidance: $9.2-9.3B vs $8B in 2024)

• 5-Year Sustainability: High

VRT is accelerating. Orders up 60% YoY in Q3 2025, with organic orders up 30% for the TTM period. Management raised 2025 guidance to $9.2-9.3B (25% growth) despite FX headwinds. With power and cooling being the #1 constraint in AI data centers, VRT sits at the bottleneck. Their integrated solutions (Smartrun prefab platforms) and Great Lakes acquisition expand TAM into white space infrastructure.

Verdict: 20%+ growth for 2025-2027, moderating to 12-15% through 2030. Most likely to sustain 15%+ for the full 5 years.

Quanta Services (PWR): Steady Eddie

• 2024 Growth: 13%

• 2025 Projection: 12-15% (guidance: $26.6-27.1B vs $23.7B in 2024)

• 5-Year Sustainability: Medium

PWR is growing, but not explosively. Their diversification (grid, renewables, underground utilities) provides stability but dilutes the AI data center purity play. The electric segment is strong, but underground margins are under pressure. With record backlog ($39B), revenue visibility is high, but incremental margin capture is the question.

Verdict: 10-13% CAGR through 2028 is realistic, with risk of slowing to high single-digits post-2028.

Winner: Vertiv. VRT has the momentum, the backlog, and the strategic positioning to sustain 20%+ near-term and 15%+ long-term. FIX is second. ALAB will decelerate fastest.

Reverse DCF: What's Baked Into Today's Price?

A reverse DCF asks: "What growth rate is implied by the current stock price?" This reveals whether expectations are sane or delusional.

Assumptions (Standardized):

• WACC: 10%

• Terminal growth rate: 3%

• Current market caps (as of Oct 2025): PWR ~$62B, FIX ~$30B, VRT ~$75B, ALAB ~$37B

Quanta Services (PWR)

• Current EV: ~$67B

• TTM Revenue: ~$27B

• TTM Operating Income: ~$1.5B (5.6% margin)

• Implied Growth: Market is pricing in ~12-13% revenue growth for 5 years + margin expansion to 7-8%

Analysis: Reasonable but unexciting. PWR trades at 2.4x EV/Sales and needs to hit its guidance consistently while expanding margins. The valuation assumes AI infrastructure demand persists and margins improve from today's 5.6% to mid-single-digits. Achievable but leaves little room for error.

Fair Value Range: Current price ($455) looks about right, implying 8-10% annual returns if they execute.

Comfort Systems USA (FIX)

• Current EV: ~$31B

• TTM Revenue: ~$8.3B

• TTM Operating Income: ~$770M (9.3% margin)

• Implied Growth: Market is pricing ~15-18% revenue growth for 5 years + margin maintenance/slight expansion

Analysis: Aggressive but supportable. FIX trades at 3.7x EV/Sales (premium to peers) and ~40x P/E. The market is betting they can sustain mid-teens revenue growth while protecting industry-leading ROIC. Given their backlog growth (65% YoY) and pricing power, this isn't insane—but any revenue miss will hurt.

Fair Value Range: Stock at ~$680 (as of Q3 2025 reports) prices in near-perfection. Upside requires exceeding 18% growth or expanding margins beyond 13-14%.

Vertiv (VRT)

• Current EV: ~$76B

• TTM Revenue: ~$9B

• TTM Operating Income: ~$1.5B (16.7% margin)

• Implied Growth: Market is pricing ~22-25% revenue growth for 3 years, then 12-15% for years 4-5 + margin expansion to 20%+

Analysis: Extremely aggressive but momentum-supported. VRT trades at 8.4x EV/Sales—a SaaS-like multiple for a hardware company. The valuation assumes they sustain accelerating growth (60% order growth continues) and hit 20%+ operating margins (up from 16.7% TTM). This is plausible given Q4 2024's 21.5% margin, but any demand slowdown will crater the stock.

Fair Value Range: Stock at ~$193 (Oct 2025) is pricing 25%+ growth for 3 years. If they deliver, upside to $250-280 exists. If growth moderates to 15%, stock is overvalued by 30-40%.

Astera Labs (ALAB)

• Current EV: ~$37B

• TTM Revenue: ~$605M

• TTM Operating Income: ~$42M (7% margin)

• Implied Growth: Market is pricing ~60-70% revenue growth for 2-3 years, then 30%+ for years 4-5 + margin expansion to 25-30%

Analysis: Insane. ALAB trades at 61x EV/Sales with ROIC of 5%. The valuation assumes they become a $3-4B revenue company by 2030 with operating margins rivaling NVIDIA. This requires winning the switch market (Scorpio), maintaining retimer dominance (Aries), and fending off Broadcom/Marvell—simultaneously. Possible? Yes. Probable? No.

Fair Value Range: Stock at ~$160 (Oct 2025) prices in hyper-growth + margin miracles. If they hit consensus (revenue slowing to 30-40% by 2027), fair value is $80-100. If they execute perfectly, $200+ is possible, but risk/reward is asymmetric to the downside.

Winner (Best Risk/Reward): Quanta Services. PWR is priced for modest growth and margin improvement, achievable targets. VRT and ALAB are priced for perfection. FIX is expensive but justifiable given ROIC.

The Verdict: Where Should You Deploy Capital?

Best ROIC, Widening Moat: Comfort Systems USA (FIX)

If you believe in quality compounders, FIX is the answer. 30%+ ROIC, expanding margins, and a backlog that's up 65% YoY make this a capital allocation machine. The valuation is rich (~40x P/E), but you're paying for a business that can compound at 15%+ annually for a decade.

Trade: FIX is a buy on any pullback below $600. Current price ($680) is fair but offers limited margin of safety.

Best Growth, Highest Risk: Vertiv (VRT)

VRT is the momentum play. If AI data center spending continues accelerating, VRT's integrated power/cooling solutions will print money. Operating margins expanding from 17% to 21%+ in one year is extraordinary. But the valuation (8.4x EV/Sales) assumes zero disappointment.

Trade: VRT is a buy for aggressive growth investors willing to accept volatility. Price target: $250+ if they deliver on 25% growth, but downside to $140-150 if growth slows.

Best Value, Steady Returns: Quanta Services (PWR)

PWR is the "boring" winner. They're everywhere AI infrastructure is being built, have scale economies, and a $39B backlog. The stock is fairly priced for 12-13% annual returns—not sexy, but reliable.

Trade: PWR is a hold at current levels (~$455). Add on dips below $400.

Best Technology, Worst Valuation: Astera Labs (ALAB)

ALAB is the wildcard. If they dominate connectivity silicon for AI clusters, the stock could 3-4x. But at 61x EV/Sales with a 5% ROIC and rising competitive threats, you're paying for a dream. ESUN and Broadcom could make ALAB's switch business obsolete before it scales.

Trade: ALAB is a pass unless you're a high-conviction tech bull. Fair value is $80-100 unless they prove out the switch/CXL market.

Final Rankings

Best Overall (Next 5 Years):

1. Vertiv (VRT) — If AI infrastructure demand is real, VRT is the pick-and-shovel play with pricing power.

2. Comfort Systems USA (FIX) — Quality compounder with widening moat and exceptional ROIC.

3. Quanta Services (PWR) — Solid execution, diversified exposure, fair valuation.

4. Astera Labs (ALAB) — Great technology, terrible valuation, existential competitive risks.

Best ROIC Profile:

1. FIX (30%+)

2. VRT (16-18%, expanding)

3. PWR (~8%, stable)

4. ALAB (5%, ramping but unproven)

Strongest Moat:

1. VRT (installed base + ecosystem)

2. FIX (specialized expertise + pricing power)

3. PWR (scale + diversification)

4. ALAB (technology lead, easily disrupted)

Most Likely to Sustain 15%+ Growth:

1. VRT (20%+ for 3+ years, 15%+ through 2030)

2. FIX (15-18% through 2027, 10-12% thereafter)

3. PWR (12-13%, moderating to 8-10%)

4. ALAB (60%+ near-term, slowing to 20-30% by 2027)

Bottom Line: If you want to own one AI infrastructure stock for the next five years, buy Vertiv. If you want the safest compounder, buy Comfort Systems USA. If you want a diversified play at fair value, buy Quanta Services. And if you're feeling lucky, speculate on Astera Labs—but size it appropriately for a binary bet.

The AI gold rush is real. Just make sure you're backing the miners who actually know where the gold is buried. Right now all four sit on my watchlist, which you can access in real-time in our community.

Would you like to stay ahead of opportunities like this? Join our community where we share real-time trade alerts and deep-dive analyses of businesses with true competitive advantages. Don't just trade the market - invest in excellence.

Want to receive our trade alerts and detailed analysis in real-time? Join our community of value investors who understand that pricing power is the ultimate competitive advantage. Receive our trade alerts on your phone? Download the app here