Comfort Systems USA (FIX) FY2025 Earnings Quality Report

Grade: F — Inventory Growth Flag Amid Data Center Boom

Framework: Schilit *Financial Shenanigans* + Beneish M-Score + forensic accounting principles

Data: SEC EDGAR 10-K (Filed 2026-02-19) + Yahoo Finance

Auditor: Deloitte & Touche LLP — Clean opinion

One-line verdict: Comfort Systems USA delivered an extraordinary 2025 — revenue +29.5% to $9.10B, operating income +75% to $1.31B, net income +95% to $1.02B, record backlog of $11.94B. The company has delivered 27 consecutive years of positive free cash flow, extended a new $1.10B revolving credit facility in August 2025 (up from $850M), and operates with minimal debt (total debt $479M, $100M drawn on the revolver). So why an F? Three mechanical red flags from the screening engine: (1) accounts receivable has outpaced revenue growth for two consecutive years (A2), (2) inventory grew 41.9% while cost of sales grew 24.4% — and gross margin rose — which triggers the Schilit inventory-build "fraud signal" pattern (B1), and (3) goodwill plus intangibles of $1.52B equals 62% of equity, slightly over the 50% threshold (D1). The real story is not fraud — it is the mechanical rules tripping on a construction contractor mid-explosion. Revenue doubled in two years on data-center demand, and working capital has to expand to fund that scale. The Beneish M-Score of -2.35 sits just 13 basis points below the -2.22 fraud threshold, which is a meaningful reminder that rapid growth + margin expansion + working-capital buildup is exactly the pattern Beneish was built to flag. Deloitte signed a clean opinion but investors should read the 10-K's one-customer concentration disclosure: "one customer represented approximately 12.8% of our consolidated revenue."

The Numbers: Data Center Tailwind in Full Force

Comfort Systems USA is a national specialty contractor providing mechanical and electrical services — installation of HVAC, plumbing, piping, electrical and controls systems in commercial, institutional and industrial facilities. The MD&A is blunt about what is driving growth: "We experienced an unprecedented demand environment in 2025… we currently expect that supportive conditions for our industry, especially for our industrial and technology customers, are likely to continue in 2026."

Revenue has more than doubled in three years (from $4.14B to $9.10B), gross margin has expanded 620 basis points (17.9% → 24.1%), and net income has quadrupled. These are exactly the kind of growth rates that trigger the Schilit framework's "too good to be true" flags — not because they are necessarily manipulated, but because the pattern is historically associated with manipulation.

Cash Flow: Strong and Accelerating

The MD&A provides this cash flow summary:

Free cash flow grew from $550.7M (2023) to $1,035.1M (2025) — an 88% increase over two years. The MD&A explains the 2025 OCF growth: "The $337.3 million increase in cash provided by operating activities was primarily driven by higher earnings before non-cash expenses such as amortization of intangible assets in the current year and an $877.9 million benefit from changes in billings in excess of costs and estimated earnings and deferred revenue driven by the timing of customer billings and payments. These increases were partially offset by a $778.9 million decrease in accounts payable and other current liabilities driven by the size and timing of payments."

The $877.9M "billings in excess of costs" benefit is important — it means FIX has collected cash from customers ahead of recognizing the related revenue. That is healthy advance billing behavior for a contractor with strong customer position.

The MD&A also discloses: "We have generated positive net free cash flow for the last 27 calendar years, much of which occurred during challenging economic and industry conditions." That track record is unusual for a specialty contractor.

The 18-Point Screening

Revenue Quality

A2 is a red flag. AR grew from $1.32B (2023) to $1.86B (2024) to $2.58B (2025) — 41% and 39% YoY growth in two successive years, ahead of revenue growth of 35% and 29%. DSO also extended from 96 days to 103 days. On a pure numbers basis, this is the Schilit pattern — but note A3 passed because CFFO grew faster than revenue (thanks to the $877.9M billings-in-excess benefit). The story is consistent with rapid project growth in large data center construction, not with revenue recognition manipulation.

Expense Quality

B1 is a red flag, and the engine calls it a "fraud signal." The test: if inventory growth > COGS growth + 15% AND gross margin is rising, it is exactly the pattern associated with under-reporting COGS by capitalizing too much into inventory. Comfort Systems' inventory grew from $59M to $84M (+42.4%) while COGS grew 24.4% — and gross margin expanded 310bp. This is mechanically what Schilit's Shenanigan #2 (recording bogus revenue) and #6 (shifting current expenses to a later period) look like.

The human interpretation, however, is different. For a specialty contractor, inventory primarily consists of materials purchased for specific projects in progress. If the backlog is $11.94B and growing, and large data center jobs require long-lead-time materials (chillers, electrical switchgear, cooling towers), the inventory buildup is legitimate pre-purchasing for known projects. The gross margin expansion is partially driven by mix — data center work carries better margins than general commercial construction. Nothing about the 10-K disclosures suggests actual manipulation. But the mechanical rule fires, and the rule exists for a reason.

Cash Flow Quality

Clean on all cash flow tests. This is important — if the inventory flag were actually a manipulation signal, we would expect CFFO to be dragging behind net income. Instead CFFO/NI is 1.16, FCF/NI is 1.01, and accruals are negative (-2.5%), all consistent with real earnings quality.

Balance Sheet Health

D1 is a red flag. Goodwill of $1.03B plus intangibles of $0.49B totals $1.52B against equity of $2.45B — 62%. Just over the 50% threshold. FIX has grown through regular bolt-on acquisitions of regional HVAC and mechanical contractors, which naturally accumulates goodwill.

D2 is exceptional — debt/EBITDA of just 0.3x, essentially net-cash. This is a critical context: the D1 flag is not leveraged-buyout-level goodwill, it is the cumulative result of two decades of small acquisitions on a small equity base.

Acquisition Risk

Beneish M-Score: -2.35 (Just Below Threshold)

M-Score of -2.35 is just 13 basis points below the -2.22 threshold. This is one of the closest readings in our coverage. The drivers pushing FIX close to the threshold: high sales growth (SGI), rising gross margin variable (GMI), and the DSRI (days sales in receivables) variable — all consistent with the A2 and B1 flags.

Combined with the "fraud signal" on B1, this is the one company in this batch where the screening framework is literally saying "this looks statistically like the pattern of companies that have manipulated earnings — even though we cannot point to any specific manipulation." The disclosure quality in the 10-K and the clean Deloitte opinion argue the statistical flag is a false positive driven by legitimate growth. But the flag is the flag.

Altman Z-Score: 4.04 (Safe)

Z-Score of 4.04 comfortably in the safe zone. FIX's essentially net-cash balance sheet and strong EBIT generation keep distress risk low.

Key Risks from the 10-K

1. Backlog Conversion Risk

Item 1A discloses the fundamental contractor risk: "Our backlog as of December 31, 2025 was $11.94 billion. The predictive value of backlog information is limited to indications of general revenue direction over the near term, and we cannot guarantee that the revenue projected from our backlog will be realized or, if realized, will be profitable. Projects may remain in our backlog for an extended period of time, or project cancellations or scope adjustments may occur with respect to contracts reflected in our backlog."

2. Customer Concentration — 12.8% From a Single Customer

The 10-K discloses directly: "A limited number of customers have in the past and may in the future account for a significant portion of our revenue. For example, in 2025, one customer represented approximately 12.8% of our consolidated revenue. Although we have long-standing relationships with many of our significant customers and believe that our portfolio of customers is reasonably diverse, one or more of our significant customers may reduce, fail to renew, or terminate their contracts with us in the future."

12.8% from one customer on a $9.1B revenue base is ~$1.16B of revenue from a single relationship. The 10-K does not name the customer, but for a specialty contractor pivoting to data center work, the probability it is a hyperscaler is high. Hyperscaler capex cycles can be abrupt — one delay or re-prioritization could materially impact 2026.

3. Cost-Overrun Risk on Fixed-Price Contracts

Item 1A: "Because we bear the risk of cost overruns in most of our contracts, we may experience reduced profits or, in some cases, losses under these contracts if costs increase above our estimates." The 10-K describes the input variables: "prices, including commodity prices and inflation; availability of labor, including the costs of providing labor, equipment, and materials; and other factors outside our control."

4. Cyclical End Markets

"Economic downturns in the markets in which we operate may materially and adversely affect our business because our business is dependent on levels of construction activity… The industries and markets in which we operate have been and will continue to be vulnerable to macroeconomic downturns because they are cyclical in nature. When there is a reduction in demand, it often leads to greater price competition as well as decreased revenue and profit."

5. Inflation and Interest Rate Exposure

The 10-K acknowledges: "we have exposure to changes in interest rates under our revolving credit facility, and as interest rates increase, our debt service obligations on our variable rate indebtedness will increase even though the amount borrowed remains the same." Interest expense detail: FY2025 total interest expense was just $9.0M — minimal today, but the exposure is real if rates rise and the revolver is more heavily drawn.

6. Surety Bonding Capacity

The MD&A: "approximately 10% to 20% of our business has required bonds. While we currently have strong surety relationships to support our bonding needs, future market conditions or changes in the sureties' assessments of our operating and financial risk could cause our sureties to decline to issue bonds for our work. If that were to occur… such an interruption would likely cause our revenue and profits to decline in the near term."

7. Competition From In-House Providers

"We also expect increased competition from in-house service providers because some of our customers have employees who perform service work similar to the services we provide." Large industrial customers building out multi-site data center campuses could vertically integrate MEP contracting.

8. Acquisition Integration

"Our recent and future acquisitions may not be successful… Acquisitions may expose us to additional business risks different than those we have traditionally experienced. We also may encounter difficulties integrating acquired businesses."

Summary

Comfort Systems USA just posted the strongest year in its history. Revenue of $9.10B is more than double 2022. Gross margin is up 620bp. Net income is up nearly 4x from 2022. Backlog at $11.94B is a record. 27 consecutive years of positive free cash flow. The business is firing on every cylinder.

Three red flags — all pattern-based, none clearly substantive:

Plus a very close M-Score at -2.35 — only 13 basis points from the fraud threshold, pushed close by the same growth + margin + receivables dynamics that trigger A2 and B1.

Things working: 27 years of positive FCF, Debt/EBITDA of 0.3x (essentially net-cash), Altman Z-Score 4.04, CFFO/NI 1.16 (cash-backed earnings), clean Deloitte opinion with multiple critical audit matters discussed, $877.9M billings-in-excess benefit showing customers are paying ahead of work completion.

Things to watch: 12.8% single-customer concentration (almost certainly hyperscaler), dependency on data center capex continuation, M-Score proximity to threshold, and the mechanical Schilit-pattern flags that will persist as long as FIX grows rapidly while carrying working capital for long-lead-time materials.

Bottom line: The F grade from this screening framework does not mean accounting manipulation is happening — it means Comfort Systems' 2025 numbers statistically match the pattern of companies that have been manipulated. Every piece of narrative disclosure in the 10-K argues this is legitimate explosive growth in a data center construction boom, and the cash flow verification (CFFO > NI, FCF > NI, accruals negative) supports that. Investors should take the mechanical flags as a reason to read the 10-K carefully, understand the customer concentration, and form an independent view on whether the hyperscaler capex tailwind continues into 2026 and beyond. The underlying business looks real; the screening framework is just doing its job.

**Disclaimer**: This report is based on Comfort Systems USA's FY2025 10-K (SEC EDGAR) and public financial data. It uses forensic accounting screening frameworks (Schilit's *Financial Shenanigans*, Beneish M-Score, Altman Z-Score) for red flag detection. This is NOT investment advice. Screening for red flags does not constitute a buy or sell recommendation. Past financial performance does not predict future results. Always do your own research and consult a qualified financial advisor.

**About EarningsGrade**: We screen earnings reports to help investors identify financial red flags. Our approach: "Screen out, not screen in." A passing grade means no red flags were detected — it does not mean the stock is a good investment.