Sunbelt Rentals Holdings SUNB

Reference table

The file advertises an 11.15% free-cash-flow yield, 8.97x price-to-free-cash-flow and 11.9x enterprise value to free cash flow. All of it rests on fiscal-2026 capital expenditure of $352M — against implied depreciation of roughly $2,316M, a capex-to-depreciation ratio of 0.15. For an equipment-rental company, whose entire asset base is a depreciating fleet that must be continuously replaced, spending 15 cents of capital for every dollar of depreciation while growing is not possible. Rental-fleet purchases are simply not in the capitalExpenditure line. Correct for it and free cash flow is nearer $1,480M, the yield is roughly 4.8%, and enterprise value to free cash flow is about 28x rather than 11.9x.

Once the artifact is removed, what remains is a decelerating, margin-compressing cyclical. Revenue growth: +21.4% (FY23) → +12.3% (FY24) → −0.6% (FY25) → +3.4% (FY26). Net income: $1,617.7M → $1,598.4M → $1,510.5M → $1,325.0M — three consecutive annual declines totalling 18.1% while revenue rose 15.4%. Operating margin: 24.4% (FY24) → 23.7% (FY25) → 19.6% (FY26), a 480 basis-point compression.

Against that: leverage is manageable at 2.25x, the buyback is real ($1,432M in FY26, shares down 4.4%), interest expense fell 27.5% year-over-year in the April quarter, and 7.8x forward EBITDA is not expensive. But with four analysts covering it, seg_prod and seg_geo both entirely empty, an empty peer list, no 200-day average, only two earnings prints on record and zero knowledge-base coverage, the required discount is large. Base case $76 against $75.07. Watch.

1. Data integrity — the free-cash-flow artifact governs this dive

This file contains the single most consequential vendor error in this batch, because it materially overstates the metric most investors would use to judge the stock. Set out in full:

REJECTED — every free-cash-flow metric in the file

cf_a reports for fiscal 2026 (year ended 2026-04-30):

• netCashProvidedByOperatingActivities: $3,784M
• capitalExpenditure: −$352M
• freeCashFlow: $3,432M

And km_ttm / ratios_ttm derive from it:

• freeCashFlowYieldTTM: 11.15%
• priceToFreeCashFlowRatioTTM: 8.97x
• evToFreeCashFlowTTM: 11.92x
• capexToRevenueTTM: 3.16%
• capexToDepreciationTTM: 0.1524

The last figure is the proof. A capex-to-depreciation ratio of 0.15 means the company spent fifteen cents of capital for every dollar of depreciation charged. For an equipment-rental company that is impossible while the business is growing — the depreciating asset is the revenue-producing asset, and a fleet not replaced is a fleet shrinking toward zero.

Independent confirmation from the income statement: FY26 EBITDA of $4,497M less EBIT of $2,181M implies depreciation and amortisation of roughly $2,316M — 20.8% of revenue, exactly what one expects from a rental fleet. Against that, $352M of recorded capital expenditure is 15.2%. The rental-fleet purchases are not in the capitalExpenditure field.

Our corrected estimate, labelled as our computation: in a low-growth year, replacement fleet capital expenditure approximates depreciation. Taking operating cash flow of $3,784M less roughly $2,300M of true fleet-plus-property capital expenditure gives free cash flow of approximately $1,480M, a free-cash-flow yield of roughly 4.8% on the $30.77B market capitalisation and enterprise value to free cash flow of roughly 28x. We use these corrected figures throughout and we discard the vendor's entirely.

This single correction is the difference between a stock that looks obviously cheap and one that looks fairly valued, and it is the reason this dive exists in the form it does.

REJECTED — the entire gross-profit series

grossProfit is internally incoherent across the file:

The cost-of-revenue definition changed mid-series, and then changed again. A margin sequence of 87.6%, 37.3%, 10.5% across three consecutive quarters is not an operating event. All gross-profit and gross-margin figures are rejected and none appears in this dive. Operating income, which is consistent throughout, is used instead.

REJECTED — the Q3 FY26 EBITDA figure

inc_q for 2026-01-31 reports ebitda: 605,000,000 against operatingIncome: 492,000,000 — an implied depreciation charge of only $113M, against $636M in the preceding quarter (Q2: $1,340.7M EBITDA less $704.4M operating income) and $528M in the following quarter. For a company with $22,268M of total assets, $113M of quarterly depreciation is impossible. Rejected. The quarterly EBITDA sum is therefore understated at $4,200.9M; we use the annual figure of $4,497M, which is consistent with the FY25 and FY24 annual figures ($5,000.7M and $4,885.7M).

REJECTED — two annual EPS figures

• FY2023: eps: 2.92 against net income of $1,617.7M on 446.5M diluted shares = $3.62. A 24% discrepancy. Rejected.
• FY2025: eps: 3.66 against net income of $1,510.5M on 436.95M diluted shares = $3.46. A 6% discrepancy. Rejected.
• FY2022 ($2.79 against a computed $2.785), FY2024 ($3.68 against $3.638) and FY2026 ($3.17 against a computed $3.170 — exact) all reconcile. We use computed EPS where the reported figure fails.

FLAGGED — annual interest expense does not match the quarterly sum

inc_a FY26 reports interestExpense: 387,000,000. The four quarters sum to $454.7M ($129.2M + $131.5M + $98.0M + $96.0M) — a 15% discrepancy. We use the quarterly series, which is internally consistent and shows the deleveraging trend clearly, and we flag the annual figure.

FLAGGED — two incompatible operating-margin definitions in one record

ratios_ttm reports ebitMarginTTM: 0.15738 and operatingProfitMarginTTM: 0.20649 for the same trailing period — a 491 basis-point difference between two measures that should be identical or near-identical. Both are flagged; we compute operating margin directly from inc_a (FY26 operating income $2,181M ÷ revenue $11,154M = 19.55%).

FLAGGED — insufficient price history for standard risk measures

• tech.sma200: null
• tech.ret_6m: null
• tech.ret_12m: null
• profile.beta: 1.649 — computed on roughly five months of data since the 2026-03-02 listing, and therefore statistically unreliable.
• The fifty-two-week high differs between sources ($86.68 in quote, $86.06 in tech) — a minor inconsistency reflecting the short history.

No primary-trend read, no medium-term relative-performance read, and no reliable beta is available.

SEVERE GAP — seg_prod, seg_geo AND peers are all empty

• seg_prod: [ ] — no product or business-unit revenue, despite the profile explicitly naming three core units: North American general tool, North American specialty, and the United Kingdom division.
• seg_geo: [ ] — no geographic revenue at all, despite a named UK business.
• peers: [ ] — the only ticker in this batch with an empty peer list. No relative-value anchor of any kind is available from the vendor.

The consequences are direct. The specialty-versus-general-tool mix and the UK exposure are the two variables that most determine an equipment-rental company's margin and cyclicality, and neither is disclosed anywhere in this file. We cannot attribute the 480 basis points of operating-margin compression to mix, to geography, to rate or to utilisation. No claim in this dive rests on a segment split, and the absence is a direct input into the size of the discount we apply.

GAP — only two earnings dates exist, and there is no beat history

earn_cal contains three rows, one pending:

• 2026-03-12 (Q3 FY26): adjusted EPS $0.78 against $0.811 estimated — a MISS; revenue $2,637M against $2,651.5M — also short.
• 2026-06-23 (Q4 FY26): adjusted EPS $0.74 against $0.745 — in line; revenue $2,754M against $2,637.2M — a 4.4% beat.
• 2026-09-08 (Q1 FY27): pending, $1.04 estimated on $2,979.7M.

One miss and one in-line result is not a track record. Note also that adjusted EPS exceeds GAAP EPS in both quarters ($0.78 against $0.69; $0.74 against $0.55), so the adjusted series is meaningfully higher, and with only two observations the size and composition of the adjustment cannot be assessed.

CLEAN — the core revenue and share-count reconciliation

The four FY26 quarters sum to $11,154.2M, matching inc_a's $11,154.0M. km_ttm.revenuePerShareTTM of $26.714 × 417.5M shares = $11,153M. Confirmed. ratios_ttm.netIncomePerShareTTM of $3.153 × 417.5M = $1,316M against a reported $1,325M — a 0.7% difference. Confirmed. km_ttm.enterpriseValueTTM of $40,905.8M equals market capitalisation $30,774.8M plus net debt $10,131M exactly — a rare clean tie. Note a minor share-count spread: the quote's market capitalisation at $75.07 implies 409.9M shares against the 418.0M diluted count; we use 418.0M and flag the 2% difference.

2. The business — and what we are not told about it

Per the profile: an equipment-leasing firm supplying tools, machinery and engineered solutions to construction, industrial and specialty customers, organised into three core units — North American general tool services, North American specialty services, and a United Kingdom division. 25,428 employees. Chief executive Brendan Christopher Horgan. Headquartered in Fort Mill, South Carolina. The entity was established on 2025-02-12 and listed on 2026-03-02, so the pre-listing financials are carve-out accounts.

We are told the three units exist and we are told nothing about their size. seg_prod and seg_geo are both empty (§1). In equipment rental, the specialty business (climate control, power, pumps, trench safety, remediation) typically carries materially higher margins and lower cyclicality than general tool, and a UK business carries separate currency, cycle and cost dynamics. The mix between them is the single most important structural fact about this company, and it is absent.

3. The multi-year decline — the fact the metrics obscure

Three things, in order.

First, growth decelerated in a straight line and then stalled. +21.4%, +12.3%, −0.6%, +3.4%. Over the four years revenue rose 15.4% in total — a 3.7% annual rate — with all of the gain arriving in the first year.

Second, net income fell in each of the last three years — $1,617.7M to $1,325.0M, a cumulative 18.1% decline — while revenue rose 15.4%. That divergence is the entire story of this business over the period, and no free-cash-flow ratio in this file surfaces it.

Third, the cause is visible in the operating margin: 26.1% (FY23) to 19.6% (FY26), a 650 basis-point fall from peak and 480 from FY24. On $11,154M of revenue, 480 basis points is roughly $535M of operating income — more than a third of the entire FY26 net income. Whether it is rental-rate pressure, utilisation, fleet-cost inflation, mix or the UK is unanswerable from this file (§1).

The one hopeful detail is at the very end. Quarterly revenue, year over year:

+8.9% in the final quarter after +0.7% and +2.7% is a genuine re-acceleration, and it is why the verdict is Watch rather than Avoid. But note the counterpoint from the same quarter: net income fell to $226M from $311.3M a year earlier (−27.4%) and EPS to $0.55 from $0.71 — revenue accelerating while earnings fell 27%. The margin problem got worse, not better, in the quarter the top line improved.

4. Balance sheet, deleveraging and the buyback

• Net debt / EBITDA: 2.25x on FY26 EBITDA of $4,497M (our computation). km_ttm reports 2.41x using the quarterly sum contaminated by the corrupt Q3 (§1). Both are comfortable for the asset class.
• Interest coverage 5.07x, debt service coverage 3.71x. Adequate.
• Current ratio 0.90x, working capital negative $244M, cash ratio 0.012x — the company holds $29M of cash against $22.3B of assets. Normal for a revolver-funded rental business, but it means there is no liquidity buffer and the model depends entirely on continuous cash generation and facility access.
• Debt to equity 1.37x, debt to assets 45.6%.
• The deleveraging is visible in the interest line and it is real: fourth-quarter interest expense fell from $132.4M (April 2025) to $96.0M (April 2026) — a 27.5% reduction, worth roughly $145M annualised on a business earning $1,325M. Across the year the quarterly progression was $129.2M, $131.5M, $98.0M, $96.0M.
• The buyback is substantial: $1,432M in FY26 — 4.7% of the current market capitalisation — against $427.4M (FY25) and $108.3M (FY24). Diluted shares fell from 437.1M to 418.0M, −4.4%.
• Dividend $0.75 annually, a 1.00% yield, 35.2% payout ratio.
• Goodwill $3,476M is 15.6% of assets and 11.3% of market capitalisation — modest, and tangible book value per share is positive at $8.61.
• Returns: return on equity 17.5%, return on invested capital 8.2%, return on assets 5.9%. The 8.2% return on invested capital is the honest measure and it is unremarkable for a capital-intensive business — roughly at or slightly above a plausible cost of capital.

Read: the balance sheet is the strongest part of this story. 2.25x leverage with 5.1x coverage, falling interest expense, a 4.7% annual buyback and a covered dividend. It is also the only part of the story that is improving.

5. Knowledge base — zero claims, and a regional name collision

A search of 51,928 knowledge-base entries for "SUNB", "Sunbelt" and "Ashtead" returned two string matches, both from Odd Lots, and neither refers to this company:

Odd Lots (2023-01-30, conviction 65, neutral) — categories rental housing / regional housing: "New supply concentrated in Sunbelt (Austin, Texas, Florida, Phoenix); coastal superstar cities (NY, SF, Boston) keep underbuilding, staying expensive and tight."

Odd Lots (2024-03-22, conviction 60, bullish) — categories office / commercial real estate: "Not all office is bad — new clean/green office even in NYC is doing exceptionally well, as is Sunbelt office benefiting from strong demographic shifts."

Both refer to the US Sunbelt REGION — the southern band of states — in the context of housing supply and office real estate. Neither mentions Sunbelt Rentals Holdings Inc. This is a pure name collision of exactly the type flagged in round one, and the correct count is zero.

kb_claim_count: 0. kb_breadth: 0.

This dive is fundamentals- and quant-driven only — an accepted and precedented outcome, and never a reason to manufacture conviction. Combined with four analysts, an empty peer list, empty segment and geographic disclosure, and no 200-day average, Sunbelt is tied with Medline as the least-triangulated name in this batch. There is no expert view, no consensus depth, no relative-value anchor and no primary-trend indicator. That is not evidence against the company; it is a direct and substantial input into the discount we require.

6. Valuation — corrected, and what is priced at $75.07

At $75.07 (market cap $30.77B, net debt $10,131M, EV ~$40.91B):

Note the coverage collapse: FY29 and FY30 EPS rest on one analyst each, and revenue on four. The base case is built on FY27E, where coverage is 9 on revenue and 7 on EPS — still thin, but the best available.

Note also the FY30 revenue step to $15,140M (+12.3%) after three years of 6-7% — on four contributors. We do not use it.

6a. What today's price assumes (the inversion)

Reverse-engineering $75.07 into falsifiable claims. Consensus-derived arithmetic, labelled as such.

• Revenue grows ~6.3% in FY27 and ~6.4% in FY28, reaching $12,615M. Consensus (9 analysts). That is roughly double the +3.4% just delivered, and requires the April quarter's +8.9% re-acceleration to persist. Falsifiable 2026-09-08, where consensus revenue of $2,979.7M implies +6.4% against the July 2025 quarter.
• EBITDA margin RECOVERS from 40.3% (FY26 actual, $4,497M on $11,154M) to 44.4% (FY27E, $5,262.4M on $11,858.0M). This is by far the most fragile assumption in the price. It requires roughly 410 basis points of EBITDA-margin expansion in a single year, from a business whose operating margin has fallen 650 basis points from peak over three years and fell again in the most recent quarter. Nothing in the reported trend supports it.
• Adjusted EPS reaches $4.007 in FY27, up 8.0% from the ~$3.71 FY26 estimate. Consensus (7 analysts). Roughly a quarter of that comes from the buyback and the interest-expense reduction rather than from operations.
• The buyback continues near the FY26 pace of $1,432M (4.7% of market capitalisation). It ran $108.3M in FY24 and $427.4M in FY25, so the FY26 rate is one year of behaviour, not an established run-rate, and its continuation depends on the corrected free cash flow of ~$1,480M rather than the reported $3,432M — a distinction that matters enormously for whether it is affordable.
• Net leverage stays near 2.25x and interest expense continues falling from $96M per quarter. On track, and the most reliable of the assumptions.
• The market keeps paying 7.8x forward EBITDA for a business with three years of falling earnings. Not obviously demanding — but the multiple has to be paid on the forecast EBITDA, which depends on the margin recovery above.

6b. The return bridge (why the multiple moves)

Expected return over an eighteen-to-twenty-four-month horizon: EBITDA growth (+17% cumulative to FY28 on consensus) + deleveraging transfer + multiple drift + shareholder yield (1.00% dividend plus ~4.7% buyback).

Our base case assumes multiple COMPRESSION: we apply 8.0x FY27E EBITDA against the 7.8x the market pays for FY27E today — nominally near-flat, but the important point is that we do not assume any expansion, and we haircut nothing from the consensus EBITDA figure while noting in §6a that the 410 basis-point margin recovery it embeds is unsupported. The de-rating in our base case is therefore implicit: if consensus EBITDA proves too high, 8.0x on a lower number is what produces the bear case.

The deleveraging transfer is a genuine and quantifiable leg here. At an enterprise value of $40.91B with $10.13B of net debt, equity is 75.2% of the capital structure. Corrected free cash flow of ~$1,480M, if applied to debt, transfers roughly $1,480M — about $3.54 per share, or 4.7% — to equity annually at a constant multiple. But the company is spending $1,432M of it on buybacks instead, which is an equivalent-value transfer through a different route (shares down 4.4% in FY26). Either way it is real, and it is the single largest component of the expected return.

So the base-case return of roughly +1% decomposes as: ~8% of consensus EPS growth, plus ~5.7% of combined dividend and buyback yield, offset by a ~12% implicit de-rate arising from our scepticism about the embedded margin recovery and by a starting price that already reflects the consensus case. State it plainly: on the consensus numbers this stock is worth more than $75; our base case is at $75 because we do not believe the FY27 margin assumption.

The bull case at $95 requires 9.5x FY27E EBITDA and the consensus margin recovery to land — two things at once. That is the fragile leg.

6c. Variant perception (where we differ, what would surprise)

• Our primary variant perception is about the data, not the company, and it is the reason this dive exists. The market is being shown an 11.15% free-cash-flow yield and an 8.97x price-to-free-cash-flow ratio that are artifacts of rental-fleet capital expenditure missing from the capex line (capex-to-depreciation of 0.15, §1). Corrected, the yield is ~4.8% and enterprise value to free cash flow is ~28x rather than 11.9x. Any screen or model using the vendor's field will rank this stock far too attractively. Watchable: whether the 2026-09-08 filing discloses rental-fleet capital expenditure separately — its appearance would resolve the discrepancy in one direction or the other.
• We differ from consensus on the FY27 margin assumption. The street models EBITDA margin rising from 40.3% to 44.4% — roughly 410 basis points in a year — from a business whose operating margin has fallen 650 basis points over three years and fell again in the most recent quarter, where revenue grew 8.9% while net income fell 27.4%. We think the margin assumption is the weakest link in the consensus and that FY27 EBITDA is more likely to land near $4.9-5.0B than $5.26B. Negative surprise, watchable 2026-09-08: operating margin below ~18%.
• We may be wrong on the cycle, and the April quarter is the reason. Revenue re-accelerated to +8.9% after +0.7% and +2.7%, which in a construction-cycle business is exactly what an inflection looks like. If two more quarters confirm it, the margin recovery becomes plausible on volume leverage alone and the bull case is live. Positive surprise, watchable 2026-09-08: revenue above ~$3.05B (+9%) with operating margin above 20%.
• The under-priced structural risk is the disclosure vacuum itself. A business with a named United Kingdom division reporting no geographic revenue whatsoever, a specialty-versus-general-tool structure reporting no segment revenue whatsoever, an empty peer list, four analysts, two earnings prints and no 200-day average cannot be underwritten to the standard this house applies. Watchable: the appearance of segment disclosure, which would itself justify a narrower discount.
• Where we have NO edge: on the non-residential construction cycle, which is the external variable that determines everything here and which this dive does not forecast.

Synthos fair values

• Bear ~$57 — 6.5x FY27E EBITDA ($5,262.4M) = EV $34.20B less net debt $10,131M = $24.07B ÷ 418.0M shares = $57.58. Assumes the margin recovery fails, the construction cycle rolls over, EBITDA lands nearer $4.9B, and the market applies a trough multiple to a 2.25x-levered fleet. −24%.
• Base ~$76 — 8.0x FY27E EBITDA ($5,262.4M) = EV $42.10B less $10,131M = $31.97B ÷ 418.0M = $76.48; cross-checked at 19x FY27E adjusted EPS ($4.007) = $76.13. Assumes modest revenue growth, partial margin stabilisation, continued buyback and deleveraging. +1%.
• Bull ~$95 — 9.5x FY27E EBITDA = EV $49.99B less $10,131M = $39.86B ÷ 418.0M = $95.36, just below the $96 street high. Assumes the April quarter's re-acceleration is a genuine cycle inflection, the consensus margin recovery lands, and the buyback continues. +27%.

Base sits ~1% above spot with a −24% bear and a +27% bull. Symmetric around the current price, with the entire distribution hinging on a margin assumption the recent trend does not support.

7. Technicals — five months of history is not enough to read

• Price $75.07, +4.4% on the day, −12.8% from the fifty-two-week high of $86.06 (the quote separately shows $86.68 — a minor inconsistency, noted); fifty-two-week low $63.09, so the stock sits 19.0% above its low.
• 1.6% BELOW a 50-day average of $76.32.
• There is NO 200-day average — tech.sma200 is null — and no six-month or twelve-month return. The company listed on 2026-03-02, roughly five months ago.
• RSI 57.8, MACD slightly negative at −0.43.
• The only measurable relative performance is three-month: −2.4% against +5.1% for the S&P — a 7.5-point deficit.
• The reported beta of 1.649 is computed on ~5 months of data and is statistically unreliable.
• Read: there is not enough price history to draw a technical conclusion, and no risk framework that depends on beta or a primary trend can be applied. The 50-day at $76.32 is the only reference level available. That argues for waiting rather than for sizing small.

8. Insider activity

The file records only equity awards and tax-withholding events, no discretionary transactions:

• Lynne Fuller-Andrews (EVP & General Counsel) — A-Award of 16,822 shares at $0, transaction dated 2026-02-27 (pre-listing), filed 2026-07-14 on a Form 4/A.
• Barbara Clark (SVP & Chief Accounting Officer) — F-InKind disposition of 2,819 shares at $86.06 on 2026-06-19, leaving 33,522. F-InKind is shares withheld to settle tax on vesting, not a sale decision. Note the price: $86.06 is the fifty-two-week high, which is a vesting-date coincidence rather than a signal.
• Alexander W. Pease (Chief Financial Officer) — J-Other disposition of 4,032 deferred stock units at $0 on 2026-07-10.

Read: no directional signal whatsoever. There are no open-market purchases and no open-market sales anywhere in the file. For a company five months public, that is unremarkable — but it also means there is no insider evidence to weigh in either direction, and there is no visibility on any former-parent or sponsor position, because none appears in the data.

9. Moat and competitive position

Equipment rental at national scale has a genuine, well-understood structural advantage: branch density, fleet breadth and utilisation. A contractor needs a specific machine in a specific place tomorrow; the operator with the densest network and the widest fleet wins the booking and, critically, achieves higher utilisation on the same asset. Scale also confers purchasing power on the fleet itself. Sunbelt, at $11.2B of revenue and 25,428 employees, is one of a small number of participants with that density.

The evidence the advantage exists: a 40.3% EBITDA margin (FY26) is not achievable in a fragmented, price-taking business.

The evidence it is under pressure: operating margin fell 650 basis points from the FY23 peak and fell again in the most recent quarter while revenue accelerated 8.9%. A business with intact pricing power does not lose 27% of its quarterly earnings in a quarter when volume grows 9%. The most likely candidates are rental-rate compression, fleet-cost inflation, or a mix shift toward lower-margin general tool — and the empty segment disclosure (§1) prevents distinguishing them.

Moat: real in structure, currently not visible in the trend. That combination is precisely when a low multiple is a value trap rather than a bargain, and it is why the corrected free-cash-flow figure matters so much: at 11.2% the stock would be cheap enough to accept the uncertainty; at 4.8% it is not.

10. Capital allocation

• Buybacks $1,432M (FY26) — 4.7% of market capitalisation, against $427.4M (FY25) and $108.3M (FY24). Diluted shares 437.1M → 418.0M, −4.4%. Real and substantial.
• Dividend $0.75 annually, 1.00% yield, 35.2% payout. Modest and covered.
• Debt reduction of $192M in FY26 ($10,352.2M → $10,160M) — slow, but interest expense fell 27.5% year-over-year in the fourth quarter, implying refinancing benefit as well as principal repayment.
• Acquisitions $206M (FY26), against $147.4M (FY25), $873.7M (FY24) and $1,083.2M (FY23). Bolt-on activity has fallen roughly 80% from the FY23-24 pace — consistent with a business prioritising the balance sheet and the buyback over expansion, and also consistent with a cycle turning.
• Capital expenditure as reported $352M — but see §1; the true fleet spend is not in this line and is undisclosed. This is the single largest gap in the capital-allocation picture: we do not know what the company spends to maintain its own asset base.

The allocation judgement: returning 4.7% of market capitalisation via buyback while carrying 2.25x leverage and facing three years of declining earnings is defensible but aggressive, and its affordability depends entirely on which free-cash-flow figure is right. On the vendor's $3,432M it is comfortably funded; on our corrected ~$1,480M it consumes essentially all of it, leaving nothing for debt reduction or fleet growth. That is not a small difference, and it is the most important open question about this management team.

11. Verdict, kill-criteria and flip conditions

Watch.

The governing finding of this dive is a data one: the free-cash-flow metrics in this file — an 11.15% yield, 8.97x price-to-free-cash-flow, 11.92x enterprise value to free cash flow — are artifacts of rental-fleet capital expenditure missing from the capitalExpenditure line. A capex-to-depreciation ratio of 0.15 for a growing equipment-rental fleet is not possible. Corrected, free cash flow is roughly $1,480M, the yield is roughly 4.8%, and enterprise value to free cash flow is roughly 28x. Any screen ranking this stock on the reported figure is ranking it wrongly, and that correction alone moves it from apparently cheap to fairly valued.

What remains is a decelerating cyclical with a real balance-sheet story. Revenue growth +21.4% → +12.3% → −0.6% → +3.4%; net income down three consecutive years, −18.1% cumulative, while revenue rose 15.4%; operating margin down 650 basis points from peak, and down again in the most recent quarter while revenue accelerated 8.9%. Against that: 2.25x leverage with 5.1x coverage, a 27.5% year-over-year reduction in quarterly interest expense, a $1,432M buyback that retired 4.4% of the shares, and 7.8x forward EBITDA — genuinely undemanding if the numbers hold.

And the analytical support is the thinnest in this batch alongside Medline: four analysts in total; seg_prod, seg_geo and peers all entirely empty; no 200-day average, no six- or twelve-month return, an unreliable beta; two earnings prints on record — one miss and one in line; and zero knowledge-base claims, the two "Sunbelt" hits being references to the US region.

Base case $76 against a $75.07 spot. Position: none.

Pre-registered conditions to move from Watch to a starter position:

1. A disclosed rental-fleet capital-expenditure figure — the necessary condition. Until we know what it costs to maintain this asset base, no cash-flow-based valuation is possible.

2. Two consecutive quarters of stabilised operating margin at or above ~20% — halting the 650 basis-point slide.

3. Confirmation that the April quarter's +8.9% revenue re-acceleration extends — one more quarter above ~6% year-over-year.

4. A price below ~$62 — approaching the $63.09 fifty-two-week low, at roughly 7.0x FY27E EBITDA, where the base case would pay ~23% instead of ~1%.

5. Only then, and only at 1-2%, pending a usable 200-day average and beta.

Pre-registered KILL / do-not-buy criteria:

• A fourth consecutive year of declining net income (FY27 below $1,325M) — this would confirm that the margin compression is structural rather than cyclical and would invalidate the entire consensus EBITDA path.
• Operating margin below ~18% — a further 150 basis-point leg down from the FY26 level, at which point the 40%+ EBITDA margin that justifies the multiple is no longer credible.
• Net leverage rising above ~3.0x — with a 0.90x current ratio and $29M of cash, there is no buffer.
• The buyback being cut — it is roughly 4.7 points of the expected total return, and a cut would also be the clearest possible signal that the corrected free-cash-flow figure (rather than the reported one) is the operative constraint.
• Revenue growth returning below ~2% — the April quarter's re-acceleration proving to be a single-quarter comparison effect.

Pre-registered FLIP TO BUY:

• Disclosed fleet capital expenditure showing true free cash flow above ~$2.0B — this would mean our correction is too harsh, the real yield is ~6.5%+, and the stock is genuinely cheap. This is the single most valuable piece of information that could emerge and it is entirely within management's disclosure control.
• Two consecutive quarters of revenue growth above ~7% with operating margin above 20% — the combination that would confirm both a cycle inflection and margin stabilisation, and would validate the consensus FY27 EBITDA.
• A price below ~$62 with the balance sheet intact — the same business at 7.0x FY27E EBITDA.
• Segment and geographic disclosure appearing — a business with a UK division and a specialty unit publishing neither cannot be properly assessed; publication would materially narrow the opacity discount.

Where SUNB fits in the Synthos Framework Portfolio. The industrials / construction-cycle sleeve, as a levered cyclical fleet holding — target 1-2% at the right price, currently 0%. It is genuinely uncorrelated with the technology and healthcare names in this batch, tracking non-residential construction and industrial activity rather than compute capital expenditure or healthcare utilisation, so it would be a diversifier if it were buyable. Note that it is not a substitute for broad industrial exposure — it is a single, levered, capital-intensive fleet with an undisclosed geographic mix. Logged as a tracked Synthos call (Watch, no position; primary gate is a disclosed fleet-capex figure) as of 2026-08-04 at $75.07.

Single biggest risk: the reported free cash flow is roughly $2 billion too high, and the buyback that provides most of the expected return may not be affordable out of the real figure. On the vendor's $3,432M, the $1,432M buyback consumes 42% of free cash flow. On our corrected ~$1,480M it consumes essentially all of it — leaving nothing for debt reduction or fleet growth, in a business that has just seen net income fall for a third consecutive year.

Most fragile assumption in the price: that EBITDA margin recovers from 40.3% (FY26 actual) to 44.4% (FY27E consensus) — roughly 410 basis points in a single year. Operating margin has fallen 650 basis points from its FY23 peak, fell again in the most recent quarter, and fell while revenue accelerated 8.9%. Nothing in four years of reported results supports a 410 basis-point single-year recovery, and it is the assumption on which the entire consensus EBITDA path — and therefore the entire valuation — rests.

Provenance and disclosures

• Traceability: ZERO tagged knowledge-base claims. A search across 51,928 entries for "SUNB", "Sunbelt" and "Ashtead" returned two string matches, both from Odd Lots (2023-01-30 and 2024-03-22), and both refer to the US Sunbelt REGION in the context of rental-housing supply and Sunbelt office real estate. Neither mentions Sunbelt Rentals Holdings Inc. This is a pure name collision of the type flagged in round one, and the correct count is zero. kb_claim_count: 0, kb_breadth: 0. This dive is fundamentals- and quant-driven only — an accepted and precedented outcome, and never a reason to manufacture conviction.
• Data as-of: fundamentals 2026-04-30 (Q4 FY26; fiscal year ends 30 April) · estimates, prices, technicals and knowledge-base claims 2026-08-04. All market data drawn exclusively from scripts/deepdive/vti_data/SUNB_data.json. Company listed 2026-03-02; roughly five months of public trading history; entity established 2025-02-12, so pre-listing financials are carve-out accounts.
• PRIMARY DATA CORRECTION (§1) — every free-cash-flow metric in this file is rejected. cf_a FY26 reports capitalExpenditure of −$352M against implied depreciation and amortisation of roughly $2,316M (FY26 EBITDA $4,497M less EBIT $2,181M), giving capexToDepreciationTTM of 0.1524 — impossible for a growing equipment-rental fleet, whose depreciating asset is its revenue-producing asset. Rental-fleet purchases are not captured in the capitalExpenditure field. Consequently freeCashFlow ($3,432M), freeCashFlowYieldTTM (11.15%), priceToFreeCashFlowRatioTTM (8.97x), evToFreeCashFlowTTM (11.92x), freeCashFlowPerShareTTM ($8.22) and capexPerShareTTM ($0.843) are all overstated and discarded. Our corrected estimate, labelled as our computation: operating cash flow $3,784M less roughly $2,300M of true fleet-plus-property capital expenditure = free cash flow of approximately $1,480M, a ~4.8% yield, and enterprise value to free cash flow of roughly 28x.
• Further vendor data rejected: (a) the entire grossProfit series — implied margins run 88.2%, 88.5%, 88.9%, 88.0%, then 38.5% across annual periods, and 87.6%, 37.3%, 10.5% across three consecutive quarters; the cost-of-revenue definition changed mid-series. No gross-margin figure appears in this dive. (b) inc_q Q3 FY26 EBITDA of $605M against operating income of $492M — an implied $113M of quarterly depreciation against $636M and $528M in the adjacent quarters, on $22.3B of assets. Rejected; the annual $4,497M is used. (c) inc_a FY2023 EPS of $2.92 (computed: $3.62 on 446.5M shares) and FY2025 EPS of $3.66 (computed: $3.46) — both fail a net-income-over-shares check by 24% and 6%; FY2022, FY2024 and FY2026 reconcile.
• Vendor data flagged: inc_a FY26 interestExpense of $387M against a quarterly sum of $454.7M — a 15% discrepancy; the quarterly series is used. ratios_ttm reports two incompatible operating-margin measures for the same period — ebitMarginTTM 15.74% and operatingProfitMarginTTM 20.65%; operating margin is computed directly from inc_a (19.55%). The fifty-two-week high differs between quote ($86.68) and tech ($86.06). Share count spread: 418.0M diluted in inc_q against ~409.9M implied by the quote's market capitalisation; 418.0M is used.
• SEVERE data gaps stated, not papered over: seg_prod, seg_geo AND peers are ALL entirely empty — no business-unit revenue, no geographic revenue (despite a named United Kingdom division), and no peer set at all, the only such case in this batch. The specialty-versus-general-tool mix and the UK exposure are the two variables that most determine an equipment-rental company's margin and cyclicality, and neither is disclosed. The 650 basis-point operating-margin compression therefore cannot be attributed from this data, and no claim in this dive rests on a segment split. tech.sma200, tech.ret_6m and tech.ret_12m are all null; profile.beta of 1.649 is computed on ~5 months of data and is unreliable. earn_cal contains only two completed prints — 2026-03-12 (adjusted $0.78 against $0.811, a miss) and 2026-06-23 (adjusted $0.74 against $0.745, in line) — so there is no beat history, and adjusted EPS exceeded GAAP in both ($0.78 vs $0.69; $0.74 vs $0.55) with the adjustment's composition unassessable on two observations. FY29E and FY30E EPS rest on ONE analyst each; total coverage is four raters (2 Buy, 1 Hold, 1 Sell).
• Vendor data checked and found sound: the four FY26 quarters sum to $11,154.2M against the annual $11,154.0M; km_ttm.enterpriseValueTTM of $40,905.8M equals market capitalisation plus net debt exactly; netIncomePerShareTTM × shares reconciles to within 0.7%.
• Fair-value method: scenario multiples applied to consensus FY27E EBITDA ($5,262.4M, 9 analysts), less net debt of $10,131M, divided by 418.0M shares. Bear 6.5x, base 8.0x (cross-checked at 19x FY27E adjusted EPS of $4.007), bull 9.5x. We note explicitly in §6a that the consensus FY27E EBITDA embeds roughly 410 basis points of single-year margin expansion that four years of reported results do not support; the base case does not haircut it, so the risk sits in the bear scenario. Stated arithmetic only — no discounted-cash-flow model was built.
• Not investment advice. Independent research, educational and informational only, never personalised.
• Version: 2026-08-04-full.