All Articles
6-Lens AnalysisApril 25, 202614 min read

Comfort Systems (FIX): $12.45B Backlog and Record 25.5% Margin vs. $43.6M Insider Selling Cluster

Comfort Systems just reported a $9.37 EPS quarter against $6.77 expectations, with backlog at $12.45B (+80.8% YoY) and gross margin at a company-record 25.5%. The market has rewarded the AI-infrastructure reclassification with a 276% trailing-year return. Six lenses found that the operational story is genuinely real, the balance sheet is fortress, and customer concentration is concrete. They also found a $43.6M discretionary insider selling cluster in the eight days following the print — and the same management running the buybacks at $489 last year is now selling personal shares at $1,475.

The Quarter in Numbers

Backlog
$12.45B

Q1 2026, +80.8% YoY

Q4 EPS
$9.37

vs. $6.77 estimate (+38% beat)

Gross Margin
25.5%

Q4 2025, company record

Tech Revenue
45%

up from 33% in 2024

Free Cash Flow
$1.0B

FY2025, ~100% of net income

Net Cash
$725M+

Q3 2025 (no revolver)

Trailing P/E
~44x

vs. historical 12-18x

Insider Sales
$43.6M

8 days post-print, 100% discretionary

Want the full 6-lens analysis with signal assessments and model debates?

Opus + Sonnet ensemble. 6 lenses. 8 signals. 5 debates. Full evidence citations.

View FIX Analysis
The Central Question
Comfort Systems is a mechanical and electrical contractor that historically traded at 12-18x earnings. It now trades at ~44x. The reclassification from boring contractor to AI-infrastructure beneficiary is operationally grounded — the numbers are real. The question our committee asked: is the equity pricing a duration of demand that even insiders are not personally willing to bet on?

What 6 Lenses Found

Revenue Durability
CONDITIONAL
Gravy Gauge

$12.45B backlog reflects 1-2.5yr-old hyperscaler decisions per CFO. Two-customer concentration in modular is the operational risk.

Unit Economics
PROVEN
Atomic Auditor

310bps gross margin expansion in 2025 to a record 25.5%. Both segments contributing. Operating leverage real.

Expectations Priced
DEMANDING
Myth Meter

~44x trailing P/E for a contractor. Even on FY2026E, forward P/E is 32-35x. Multiple-compression risk is material.

Narrative-Reality Gap
ALIGNED
Myth Meter

Unlike most AI-narrative trades, FIX's numbers ARE the narrative. The gap is in duration, not current quarter.

Funding Fragility
MINIMAL
Stress Scanner

Net cash $725M+, no revolver, $1.1B credit facility through 2030, all covenants compliant. Fortress balance sheet.

Capital Deployment
DISCIPLINED
Stress Scanner

Bought $200M of stock at avg $489 — now $1,280+. Two dividend hikes in 2025. M&A pace measured.

Governance Alignment
MIXED
Insider Investigator

$43.6M insider sales in 8 days, 100% discretionary. CEO + 4 directors. COO is a net buyer (the contrary signal).

Tail Risk Severity
MATERIAL
Black Swan Beacon

40-55% drawdown achievable from multiple compression alone, even with continued growth. Cisco 2000-2003 analog.

The Reclassification Is Operationally Grounded

Most AI-narrative trades are detached from current operations: speculative chip plays, early-stage AI software, second-derivative cooling stocks. Comfort Systems is the rare case where the fundamentals fully justify the reclassification. Technology customers (mostly data centers) grew from 33% of 2024 revenue to 45% of 2025 revenue. Same-store revenue grew 33-35% per quarter through 2025. Q4 2025 EPS of $9.37 versus an expected $6.77 was a 38% beat magnitude. Full-year EPS doubled.

The CFO Bill George articulated the most important durability anchor on the Q4 call: "By the time we are booking backlog and especially by the time we are booking revenue, we are really working on things that came up at one to two and a half years ago. So for these gigantic projects, we will see whatever commitments [hyperscalers] are making now, we will see that in 2027, 2028 in our revenue." Current $12.45B backlog therefore embeds 2024-2025 hyperscaler CapEx decisions. The 2026 hyperscaler CapEx wave will appear in 2027-2028 revenue.

Late-Cycle Visibility
Backlog of $12.45B is 1.4x annual revenue at current run rate, with the 2026 hyperscaler CapEx wave not yet booked. This is unusual forward visibility for a contractor.

The Two-Hyperscaler Customer Disclosure

When asked on the Q4 call why modular capacity is going from 3M to 4M square feet, Bill George answered: "It is primarily us taking steps to meet more of the demand from our two largest customers. They would buy more if they could... The floor space right now is planned for those two large hyperscaler customers." This is the clearest concentration disclosure FIX has ever provided.

Modular was 18% of 2025 revenue and growing faster than the rest of the business. The two customers are not named but are likely from the cohort of Microsoft, Google, Meta, Amazon, or Oracle. The capacity expansion is partially committed (one large building purchase closing in February 2026, another in North Carolina under negotiation).

Concentration in the Highest-Margin Segment
If one of those two customers materially scales back commitments — due to model efficiency gains, regulatory action, or capital discipline — modular utilization could drop sharply with disproportionate margin impact. Modular margins are structurally above the corporate average.

Margin Expansion Is Broad — and Partially Cyclical

The 310-basis-point gross margin expansion in 2025 (21.0% to 24.1%) is not concentrated in one segment. Mechanical climbed from 22.4% to 24.9% YoY in Q4; electrical from 23.9% to 26.9%. Operating leverage is real: SG&A as a percentage of revenue dropped from 10.4% to 9.7% despite a $153M absolute SG&A increase. Free cash flow conversion of approximately 100% of net income is rare for a contractor.

Our Atomic Auditor lens debate resolved a key question: how much of the expansion is structural versus cyclical? The committee separated the components. Modular mix shift is structural (~50-100 bps). Operating leverage on SG&A is structural. But labor-scarcity-driven pricing power is partially cyclical (~100-200 bps that would compress in normalization). Net structural margin floor is likely 22-23% — still a step-up from historical 18-20% — with the 25%+ being the cyclical peak.

Margin ComponentTypeMagnitude
Modular mix shiftStructural+50-100 bps
Labor-scarcity pricing powerCyclical+100-200 bps
SG&A operating leverageStructural+70 bps

The $43.6M Insider Selling Cluster

Within eight days of the Q4 2025 earnings release on February 19, 2026, ten insiders sold $43.6M in stock. Every one of those sales — twenty Form 4 filings — was discretionary. Zero were pre-planned via 10b5-1 trading arrangements. The Item 408 disclosure in the Q1 2026 10-Q confirms: no 10b5-1 plans were adopted or terminated during the quarter.

Discretionary sales carry more signal than 10b5-1 sales because the seller decides timing with full knowledge of recent operational results. The cluster spans the CEO, three directors, and several officers, executing within an 8-day window at prices ranging from $1,369 to $1,475.

InsiderRoleSale ValueAvg Price
Brian E. LaneCEO$13.6M$1,453
Franklin MyersDirector$12.7M$1,472
Darcy AndersonDirector$6.2M$1,369
William SandbrookDirector$3.6M$1,443
Julie ShaeffCAO$3.2M$1,382
Constance SkidmoreDirector$1.4M$1,425
The COO Was a Net Buyer
COO and President Trent McKenna acquired more shares than tax-withholding disposed (+707 shares net). CFO Bill George was effectively neutral. The split between board-level selling and operational-leadership accumulation is informative: McKenna runs operations; if operational reality is the strongest indicator, his behavior is the most directly informed signal — and he is not selling.

Where the Models Disagreed

1

Are 25%+ Gross Margins Structural or Cyclical?

Opus argued margins are partially cyclical and would normalize 200-300 bps lower in a balanced labor market. Sonnet emphasized the modular mix shift and segment-wide breadth as structural anchors.

Resolution: Both partially correct. Net structural floor is likely 22-23% (still a step-up from historical 18-20%); the 25%+ is the cyclical peak. PROVEN classification reflects today's economics with explicit awareness that some expansion is cyclical.

2

Is the Cisco 2000-2003 Analog Apt or Misleading?

Catastrophist (Black Swan Beacon) emphasized the Cisco analog as directly applicable: revenue grew 4x while stock fell 80% as the multiple compressed from 200x to 25x. Optimist argued FIX's gross margin shift is genuinely structural, limiting the analog's magnitude.

Resolution: Cisco analog is overstated in magnitude but directionally correct. A 30-40% drawdown from multiple compression alone — even with continued operational excellence — is a realistic outcome at current ~44x trailing P/E.

3

Is the Insider Cluster a Pattern or a One-Time Window?

Opus weighted the cluster as a strong negative signal of valuation concern. Sonnet noted that a single cluster is consistent with multiple readings — tax planning, RSU vesting timing, or estate planning — not pure valuation view.

Resolution: Both readings partially correct. Recurrence over 2026 quarters would resolve toward valuation-concern interpretation. Q2 2026 post-print activity is the key signal to watch.

What the Lenses Agree On

Operations Are Real

Backlog +80.8% YoY, Q4 gross margin 25.5% (record), FCF $1B, balance sheet net cash $725M+ — Gravy Gauge, Atomic Auditor, and Stress Scanner all confirm execution at exceptional levels.

Revenue Is Anchored Through 2027

CFO's late-cycle visibility quote means current backlog already embeds the 2024-2025 hyperscaler wave. Even a 2026 CapEx slowdown does not reach FIX revenue until 2027-2028.

Two-Customer Concentration Is the Operational Risk

CFO's explicit disclosure makes concentration concrete. Gravy Gauge flagged it; Black Swan Beacon escalated it; Insider Investigator's timing resonates with the same concern.

Valuation, Not Operations, Is the Risk

~44x for a contractor. Insider selling cluster validates the valuation concern from inside. Black Swan Beacon detects ASSUMPTION_FRAGILITY across the committee on duration.

What to Watch

CRITICALAggregate hyperscaler CapEx growth (MSFT+GOOG+AMZN+META+ORCL)

Customer-end CapEx is the leading indicator for FIX backlog 1-2.5 years out. Aggregate quarterly CapEx guide-down greater than 10% would trigger valuation reset analysis.

CRITICALSequential backlog change

Currently $12.45B (+4.3% sequential). Sequential decline greater than 5% would be the first concrete signal that the cycle has turned.

HIGHDiscretionary insider selling cluster (Q2 2026 post-print)

A single cluster is ambiguous. A recurring cluster after the next earnings print would convert the signal from soft to clear.

HIGHGross margin (consolidated)

Currently 24.1% FY, 25.5% Q4. Sustained drop below 22% for two consecutive quarters would signal cyclical normalization.

HIGHCost-to-cost reversal disclosures

Auditor flagged Revenue from Contracts with Customers as the Critical Audit Matter. Any single-quarter reversal greater than $25M from prior estimates would trigger Fugazi Filter review.

Committee Posture: Proceed with Caution

This is a duration-risk situation, not a broken-business situation. Operational execution is exceptional, the balance sheet is fortress, and the data center exposure is genuine. Position sizing and entry discipline matter more than directional view. The same management running disciplined buybacks at $489 last year is now selling personal shares at $1,475 — and that is a coherent read, not a contradictory one.

Path to More Favorable Assessment

  • • Hyperscaler CapEx guides remain greater than 25% growth through 2027
  • • New modular capacity proves saturated by additional customers
  • • Insider selling clusters do not recur in 2026
  • • Multiple compresses below 30x via earnings growth (not stock decline)

Path to Less Favorable Assessment

  • • Aggregate hyperscaler CapEx guide-down greater than 10%
  • • Sequential backlog decline in any quarter
  • • Gross margin breach below 22% for two consecutive quarters
  • • Cost-to-cost reversal disclosure greater than $25M
  • • New 10b5-1 plans adopted by C-suite covering greater than 5% of holdings

This analysis is for educational purposes only — it is not a recommendation to buy or sell any security.

Public Sources Used
  • Comfort Systems USA Annual Report (10-K) FY2025
  • Comfort Systems USA Quarterly Report (10-Q) Q1 2026
  • Comfort Systems USA 10-Q Q3, Q2, Q1 2025
  • Comfort Systems USA Current Reports (8-K) (10 filings, 2025-2026)
  • Comfort Systems USA Proxy Statement (DEF 14A) 2026
  • Q4 2025 Earnings Call Transcript (Feb 20, 2026)
  • Q3, Q2, Q1 2025 Earnings Call Transcripts
  • Form 4 Insider Transaction Filings (20 filings, Feb-Apr 2026)
  • Form 144 Proposed Sale Notices (10 filings, Feb-Mar 2026)
  • Google Trends Data (data center cooling, HVAC contractor)
  • Congressional Trading STOCK Act filings (Quiver Quantitative)

Full Analysis with Signal Breakdowns

Explore the complete 6-lens assessment including debate transcripts, evidence citations, and monitoring triggers.

View FIX Analysis

This report was generated by the Runchey Research AI Ensemble using primary SEC data and reviewed by Matthew Runchey for accuracy.

This analysis is for educational purposes only and does not constitute investment advice. See our Editorial Integrity & Disclosure Policy and Terms of Service.