APi Group Corp (NYSE: APG) — A Fire-Safety Compounder Priced for a Plan It Hasn’t Finished Yet

Independent equity research. Report date: 2026-08-04 (initiation of coverage). All figures reconciled to SEC filings unless labeled otherwise. The numbered sections below carry no recommendation and no price target; the one labeled exception is Kimi’s Take, immediately below.

⚡ Kimi’s Take

This block is Kimi’s own subjective opinion and general information only — an independent analyst’s view, not investment advice. The analysis in the numbered sections below carries no recommendation and no price target; this section, and only this section, takes a position.

Verdict: HOLD / no-chase — the transformation is real and largely priced; accumulate on weakness toward the mid-$30s. Medium conviction. Directional zone: base fair value sits around ~$47 (FY28 path: ~15.25% adjusted EBITDA margin on the 10/16/60+ plan, ~$1.54B EBITDA, a held 15.5x EV/EBITDA); the stock turns interesting below ~$35 (≈13x FY26E EBITDA, the quality-industrial bracket); the bear zone is ~$32 if the Specialty project cycle rolls over into the brutal 2H-26 comps and margin expansion stalls near 14%; the bull zone is ~$66 if $10B+/16%+/60%+ lands in 2028 and the market re-rates APG toward the EME/FIX contractor complex (18–20x+). At $41.54 — 16.9% off the April high, above all three EMAs after a beat-and-raise — you are paying for the 2028 plan delivered on schedule, with no re-rating.

The pivot is genuinely working, and the evidence for it is unusually clean: ISM (inspection/service/monitoring) revenue share went from ~40% (2021) to ~54% (2024); the $3.1B Chubb acquisition — bought at ~14.5x pre-synergy EBITDA, financed to 4.25x pro forma leverage — was deleveraged to 1.6x on schedule, its value-capture target raised from $20M to $125M, its restructuring program closed out 6/30/25 with zero impairments; the “13/60/80” margin and FCF legs were delivered (13.2% FY25, 80.3% conversion); 2026 guidance has been raised four times; backlog is a record >$5B. But at 16.8x FY26E EBITDA, three independent reverse-solves converge on the same answer: the price already underwrites the plan — ~10% EBITDA CAGR for four-plus years at a held multiple; ~15–16% EPS CAGR even with a de-rating to 20x in 2030; and the 8.5x→17x post-Chubb re-rating, the easy money on this story, already happened. The growth-quality wrinkle sharpens the caution: Q2’s +10.1% organic was Specialty-led (+22% project work, data centers ~10–12% of revenue, project margins ~10 points below service) while Safety — the recurring moat the multiple pays for — grew +4.7%. Current growth is cycle-led and laps brutal 2H comps; meanwhile the 16%-by-2028 margin target needs ~75bps/yr of expansion from exactly the mix that is diluting it. The insider tape agrees with the arithmetic: co-chair Franklin sold ~$206M in two discretionary blocks at $40.88–$42.08 — today’s price — and no officer has bought open-market since 2022. Add the quality-of-earnings flags (GAAP EPS negative through 2026 on the Series A preferred appreciation dividend — ~19M shares of real dilution over two years, ending with the 12/31/26 auto-conversion; and “systems and business enablement,” $96M in FY25, a recurring cost dressed as one-time) and the stock is a good company at a full price, not a mispricing.

Framing: a long-run winner in stalled-momentum repair — neither a falling knife nor a one-way street. APG is +181% over five years vs SPY +88%, with a −49.6% max drawdown (2022) and a fresh −22% correction off the April 2026 ATH; 3M return is −8.9% vs SPY +6.8%, the Momentum factor loading is zeroed (L1) in every nested model, ~50% of variance is idiosyncratic, and the golden-cross configuration is intact with price back above all EMAs after the Q2 beat-and-raise. The value-leaning buyer is no longer fighting the tape on the long side — but momentum money has not re-engaged either, and the relative-strength lag vs SPY is the live tension.

Conviction: medium. Flips bullish if Safety/ISM organic re-accelerates to high-single-digits while Specialty holds growth against the 2H comps — that proves the recurring-service engine, not just the cycle, and earns the re-rate. Flips bearish if Specialty organic rolls toward zero/negative with consolidated margin expansion stalling below ~14% — that kills the 2028 plan the price underwrites.

Tag: a fire-safety compounder priced for a plan it hasn’t finished yet.

📈 Stock Price Action — Five-Year Event Map

Arc (FACT — AZI adjusted prices, pulled 2026-08-04). Five years ago APG traded at $14.76 (2021-08-04). It fell to an intraday five-year low of $8.73 on 2022-10-13 after levering up for Chubb into the 2022 rate shock, then compounded through four years of deleveraging, margin delivery, and guidance raises to an all-time intraday high of $49.99 on 2026-04-30 (closing peak $49.40 on 4/27). A −22.2% correction into late July reversed on the Q2-26 beat-and-raise; the stock closed 2026-08-04 at $41.54 — 16.9% off the high, +24% above the 52-week low of $33.40 (2025-09-25). Trailing returns: 1Y +16.7%, 3Y +121.6%, 5Y +181.4% (SPY +23.6%/+79.2%/+88.0%).

Cycle narrative.

1. The 2022 collapse (−45%, trough $8.73 on 10/13/22) was leverage meeting macro: APi closed the $3.1B Chubb acquisition on 1/3/22 at ~4.25x pro forma net leverage — weeks before the fastest rate-hiking cycle in decades. The stock was a high-beta levered roll-up in a bear market; the Q2-22 print (−7.2% on 8/4/22) confirmed there was no shelter in the results. This is also where the only meaningful insider buying of the entire five-year window occurred: director Anthony Malkin bought ~$2M at $14.47–$15.50 (pre-split; ≈$9.6–10.3 split-adjusted) between June and September 2022. (Source: 8-K filed 2022-01-03; AZI CSV; Form 4 corpus.)
2. The 2022–23 recovery (+98%) was the deleveraging trade: $475M of debt repaid in 2023 alone, an extra $175M TLB paydown in December 2023, and a declared path below 2.5x net leverage — delivered ahead of schedule. The market began paying for the margin plan before it fully arrived. (Source: 8-K 2023-12-19; FY23 results; AZI CSV.)
3. Early 2024 (+29%) re-rated the stock on the record Q4/FY23 print (3/1/24, +6.6% day): revenue $6.9B, adjusted EBITDA margin 11.3% and climbing, Chubb synergies tracking above plan. (Source: FY23 earnings release; AZI CSV.)
4. The spring-2025 whipsaw was macro, not company: −6.7% on 4/3/25 (tariff shock), +9.6% on 4/9/25 (pause rally), then the Q1-25 print (5/1/25, +6.6% day) delivered record revenue and a guidance raise that started the next leg. (Source: AZI CSV; Q1-25 release.)
5. September 2025 to April 2026 (+48% to the ATH) was the full transformation re-rating: 13/60/80 declared achieved, the 10/16/60+ targets framework in place, four straight beats, FY25 adjusted EBITDA margin of 13.2%, and the CertaSite/Onyx/WTech M&A cadence accelerating. The stock peaked at $49.99 intraday on 4/30/26 — the morning of the Q1 print. (Source: FY25 10-K; Q4-25 release; AZI CSV.)
6. The April–July 2026 correction (−22%) is the expectation-risk episode the current debate is built on: Q1-26 beat (record revenue +25.6% reported) but the stock fell −6.0% on the day because adjusted gross margin ticked down 40bps on project mix after the run to all-time highs — the market was already paying for perfection. The slide then ground on as Industrials/Infrastructure factors sagged (Infrastructure −9.6% over the trailing 252 days) while SPY rallied +12.5%. (Source: Q1-26 release; AZI CSV; FactorsToday factor returns.)
7. The latest leg (+8% in four sessions) is the Q2-26 beat-and-raise: revenue $2,254M (+13.3%, +10.1% organic), adjusted EPS $0.44 vs ~$0.41 consensus, FY26 revenue guidance raised a fourth time, record >$5B backlog. The stock gapped +6.2%, faded to +2.0% on the day, then added +4.3% on 8/4 — a repair rally, not yet a re-engaged momentum trend. (Source: 8-K EX-99.1 filed 2026-07-30; AZI CSV.)

Price moves are FACT; attributed drivers are INTERPRETATION unless a dated event is cited. This block is factual price history — no price target, no recommendation.

1. Executive Summary

APi Group is the largest fire-and-life-safety services company in the world, assembled from a SPAC (J2 Acquisition, founders Franklin/Ashken/Lillie) into a ~$7.9B-revenue platform across two segments: Safety Services (FY25 revenue $5,456M, 69% of the total, segment margin 16.8%) — fire suppression, detection and alarms, electronic security, and elevator/escalator services, sold inspection-first and increasingly recurring — and Specialty Services ($2,460M, 31%, margin 10.7%) — infrastructure/utility, specialty contracting (including HVAC), and fabrication/distribution, a project business currently riding the data-center and infrastructure capex wave. Demand in the core franchise is statutorily mandated: fire codes require annual inspection of nearly all installed life-safety systems, which makes inspection revenue non-discretionary and recurring; management has pushed inspection/service/monitoring (“ISM”) from ~40% of revenue in 2021 to ~54% in 2024, with a 60%+ long-term target. The transformational event was the January 2022 acquisition of Chubb Fire & Security from Carrier for $3.1B — a full price (~14.5x pre-synergy EBITDA) that has since been vindicated: value capture raised from $20M to $125M, 55 loss-making branches cut to fewer than five, restructuring closed out mid-2025, zero impairments, and group adjusted EBITDA margin up from 10.3% (2021) to 13.2% (FY25), delivering the “13/60/80” margin and cash-conversion targets on schedule. (Source: FY25 10-K, filed 2026-02-25; 2025 Investor Day deck.)

The current print is strong and the momentum is real. Q2-2026 (reported 7/30/26): record revenue $2,254M (+13.3%, +10.1% organic), adjusted EBITDA $311M (+14.3%, 13.8% margin), adjusted EPS $0.44 (+12.8%, ~6.5% above consensus), record backlog above $5B, and the fourth FY26 guidance raise of the year — revenue now $8,875–9,025M and adjusted EBITDA $1,205–1,245M. Leverage is 2.2x (company basis) after absorbing >$1B of 2026 acquisitions (CertaSite, Onyx-Fire, WTech); there is no debt maturity before 2029. The balance sheet is sound, the cash economics are honest (FY25 GAAP operating cash flow $759M corroborates the adjusted figures), and management has a multi-year record of doing exactly what it said it would do. (Source: Q2-26 10-Q and 8-K EX-99.1, both filed 2026-07-30.)

The honest offsets, each developed in the body: (i) growth quality — Q2’s consolidated +10.1% organic was led by Specialty projects (+22.0% organic, data centers ~10–12% of revenue) while the recurring Safety franchise grew +4.7%; project margins run ~10 points below ISM, so the fastest growth is the lowest-quality growth, and it laps brutal 2H-26 comps; (ii) the margin glidepath — the 16%-by-2028 target requires ~75bps/yr of expansion (Q1 delivered +70bps, Q2 only +10bps) from a mix currently tilting the wrong way; (iii) per-share quality — GAAP EPS is negative through 2026 because the Series A preferred appreciation dividend (a SPAC-legacy claim held by co-chair Franklin’s Mariposa entities) charged $590M against FY25 common earnings and issued 15.2M shares in January 2026 (~3.5% of the count); the mechanism ends with the 12/31/26 conversion to 6.0M shares; (iv) insider behavior — founders have sold ~$1.1B over five years, almost all programmatic, but Franklin’s two 2026 blocks were discretionary (non-10b5-1) at $40.88–$42.08, and no officer has bought open-market since 2022; and (v) valuation — at $41.54 the stock trades at 16.8x FY26E adjusted EBITDA and ~24x derived FY26E adjusted EPS, top-decile of its own history on price/sales, with three reverse-solves converging on “the 2028 plan, delivered on schedule, with no re-rating.” The Street’s mean target ($52.57, +27%) underwrites base-to-bull; the insider tape sits at today’s price. (Source: Form 4 corpus; valuation work-up; MarketBeat consensus 2026-08-04.)

2. Business Overview

What APi is (FACT). APi Group Corp is a ~$7.9B-revenue (FY25), ~29,000-employee services company operating from 600+ locations in 20+ countries. It reports two segments, recast in May 2025 (HVAC moved from Safety to Specialty, retroactive to the FY24 10-K):

(Source: FY25 10-K Note 22, MD&A pp. 40–41.) Safety Services spans fire suppression, fire detection and alarms, electronic security, and elevator/escalator services, organized as two operating segments — North America Life Safety and International Life Safety (largely the 2022 Chubb acquisition). Specialty Services FY25 sub-mix (10-K Note 7): Infrastructure & Utility $1,024M, Specialty Contracting $1,094M, Fabrication & Distribution $342M. Geography (FY25, invoiced-country basis): US $5,180M (65%), France $691M (9% — Chubb’s largest ex-US market), all other $2,040M (26%); roughly 35% of revenue is foreign.

The revenue model: inspection-first, land-and-expand (FACT + INTERPRETATION). The strategic core is the ISM annuity — inspection, service, and monitoring contracts that are statutorily or insurance-mandated and renew behaviorally. ISM was ~40% of consolidated revenue in 2021 and ~54% in 2024 (2025 investor day deck); the long-term target is 60%+. Management estimates each $1 of inspection work generates $3–4 of incremental annual service work (deck p. 19 — a leadership estimate, not an audited ratio), and reported 19 consecutive quarters of double-digit organic inspection revenue growth as of May 2025. Contracts themselves are short — terms “ranging from days to five years, with the majority having durations of less than six months,” per the 10-K — so the recurrence is behavioral, not contractual. Critically, APi has never disclosed a quantitative retention or renewal rate; the single most important moat metric is asserted (“consistent renewal rates”) but never measured in any filing or presentation this corpus covers. That is a diligence gap, not disproof — but it means the annuity claim rests on organic-growth prints and margin evidence rather than a disclosed cohort metric.

The two Safety operating segments have opposite service intensities (2024 investor-day mixes, FACT): Safety NA is ~46% service / ~54% project, with a ~$7K average ticket, 87% US / 13% Canada, and offerings weighted to fire suppression (72%), detection/alarms (15%), elevator (8%, annualizing the Elevated acquisition), and electronic security (5%). Safety International (Chubb) is ~63% service / ~37% project, spread across Northern Europe (50%), Central/Southern Europe (30%), and APAC (20%), with offerings weighted to detection/alarms (40%), electronic security (30%), and monitoring/connected services (20%) — 14 monitoring centers and 50M+ connected devices on a deliberately product-agnostic model.

The customer and labor base (FACT). No customer exceeds 5% of FY25 revenue; the book runs from Fortune-500 portfolios (served through the National Service Group’s single-point-of-contact model) to single-site facilities. Roughly 50% of employees are covered by collective bargaining agreements (US) or works councils (international); OSHA recordable rate is 0.87 versus an industry ~2.2, and the company captively self-insures most workers’ comp/GL/auto risk ($125M accrued) through its Sprocket captive. Multiemployer pension exposure exists (some plans “endangered/critical” per the 10-K risk factors) but the Chubb-assumed UK defined-benefit plans were annuity-bought-in during 2023, with wind-up targeted for 2027 and a net pension asset of $129M at 12/31/25 — not a material overhang.

Backlog and visibility (FACT). Remaining performance obligations (in-process project contracts only, excluding short-cycle service) were $3,605M at 12/31/25 (~74% recognizable within 12 months), rising to $3,772M at 3/31/26 and $4,066M at 6/30/26 (81% within 12 months). The CEO’s “record backlog exceeding $5 billion” at Q2-26 uses a broader definition including Specialty pipeline. The 10-K itself warns that backlog is cancellable and customers often are not committed to specific volumes — backlog is a visibility aid, not a receivable.

Chubb: the deal that made the company (FACT). Announced July 2021, closed January 3, 2022: Chubb Fire & Security from Carrier for $3.1B EV ($2.9B cash plus ~$200M assumed liabilities) — ~14.5x LTM adjusted EBITDA of ~$213M pre-synergies (a ~9.6%-margin business; the deal deck marketed ~13.3x including synergies). Financed with $800M of 5.5% Series B perpetual convertible preferred ($600M Blackstone / $200M Viking), a $1.1B incremental term loan, $300M of 4.750% notes, a September 2021 equity offering (19.75M shares at $20.25, ~$400M), and cash — pro forma net leverage 4.25x with a stated goal of returning below 3.0x “expeditiously.” Execution since: the value-capture target was raised $20M → $40M → $100M → $125M and is described as on track; loss-making branches went from 55 (2021) to fewer than five (2024) with ~60 closures/consolidations; France — the case study — went from ~$550M revenue and single-digit margins with 33 loss-making branches (−$17M) to >$700M revenue at mid-teens-plus margins with zero loss-makers and ~225 fewer heads. The restructuring program formally ended June 30, 2025 ($4M residual FY25 cost, $13M liability remaining), cumulative transaction costs were $45M, final goodwill assigned was $1,367M (all to Safety Services), and there has been no goodwill or intangible impairment on Chubb to date. The 5.5% Series B preferred was retired in February 2024 via conversion to ~49.2M (pre-split) shares and an immediate $600M repurchase of half — cleaning out the expensive paper at ~$24.60 split-adjusted against a ~$16.26 conversion price.

Verdict (INTERPRETATION). This is a genuinely well-constructed services platform: a statutorily-anchored inspection annuity at the core, a national-account wrapper locals cannot match, a proven integration machine, and a project business bolted on for cyclical torque. The structure’s known weaknesses are equally clear: the annuity claim lacks a disclosed retention metric; roughly a third of revenue is a no-moat project business; and the balance sheet is a museum of past acquisitions (goodwill + intangibles $5.4B = 153% of equity at 6/30/26).

3. Industry Dynamics

Fire & life safety services: a regulated annuity industry (FACT). Demand is created by code, not by cycles. NFPA and International Code Council standards — adopted into state and local law — require testing, inspection, maintenance, and retrofit of fire suppression and alarm systems; per the 10-K, “nearly all facilities that have existing life safety systems are required by law to have that system inspected on at least an annual basis.” Inspection is therefore non-discretionary operating spend, and code tightening is a structural tailwind. The company-estimated TAM is ~$160B+ (global fire protection ~$60B+ including ~$25B North America; global electronic security ~$90B+; US elevator/escalator services ~$10B+ of a $15B+ vertical-transportation market) — soft-sourced (“company estimates based on various market studies”), usable for framing, not precision.

Structure: hyper-fragmented below a two-player top (FACT). The ENR Top 600 (October 2025, 2024 revenue) ranks APi #1 in fire protection and sprinkler contracting at $3,509M (+8%), EMCOR #2 at $1,620M, then J.F. Ahern at $286M and a cliff to sub-$150M regional players. APi is more than 2x the #2, and APi + EMCOR are the only >$1B fire-specialty platforms; below them the trade is atomized across thousands of local shops. (Caveat: ENR ranks contractors’ fire-segment revenue; JCI’s fire service revenue sits inside its building-solutions segments and is not ranked.) The 10-K says against itself that “there are relatively few barriers to entry in many of the industries in which we operate” — the barriers that do exist are licensing/certification (state contractor and individual technician licenses, NICET-type certifications), insurance and safety prequalification for national accounts, and working-capital/bonding capacity for large projects. Entry is easy; density is hard.

Route density is the economics lever (INTERPRETATION anchored in FACT). Inspection and service are metro route businesses: a technician’s day is a sequence of scheduled stops, so branch density drives utilization, response time, and margin. The Chubb turnaround is the financial proof — taking International from 55 loss-making branches to fewer than five with ~60 closures roughly doubled underlying branch profitability, and Safety NA segment margin has walked from ~16% (2022) toward ~19% (2025F) with a 21%+ 2028 target. If density were not the mechanism, consolidation would not have produced that margin path.

Named competition (FACT). Johnson Controls (Tyco/Simplex/Grinnell fire plus security service — the other global scale player, but fire is one product line in an HVAC-led, OEM-tied portfolio; per the author’s prior Johnson Controls analysis (2026-06-13), JCI itself frames fire/security margin “catch-up” as the opportunity); EMCOR (fire segment $1.6B within a construction roll-up); Pye-Barker (Altas-backed, self-described largest fully integrated fire/life-safety/security consolidator), Sciens Building Solutions, Impact Fire/AI Fire (Audax) — PE roll-ups still assembling APi’s footprint; Cintas Fire (extinguisher-route niche); Siemens Smart Infrastructure, Bosch Building Technologies, and Securitas in Europe; Convergint in security integration. Carrier exited fire/security entirely in 2024–25 (Global Access Solutions to Honeywell; industrial fire to Sentinel Capital; Kidde divested) — a structural positive for service-focused incumbents, removing a price-disciplined-ish OEM competitor from the consolidation market and from national accounts.

Specialty contracting: mediocre structure, hot cycle (FACT + INTERPRETATION). Specialty serves critical infrastructure, high-tech manufacturing, healthcare, and energy end-markets, North America only. It is a participant in the data-center boom, not a pure play — contrast FIX (45–56% of revenue data centers) and EME (data-center RPO +60% to $4.46B). Management sized the exposure on the Q2-26 call: data centers ~10–12% of 2026E revenue and ~15–20% of the Q2 backlog increase, with data-center fire projects inflating from $7–8M to approaching $20M each. Per prior analysis of Arcosa, Comfort Systems, and EMCOR (2026-06/07), the sector’s pricing power is real but cyclically amplified — an AI/data-center/IIJA demand shock ($550B of IIJA funding, which runs through 2026) colliding with a hard skilled-labor supply constraint — and the value-chain margin hierarchy runs self-perform specialty trades (EMCOR ~9% operating margin, ~31% ROIC) above asset-light design above lump-sum EPC. APi’s Specialty segment at a 10.7% margin sits in the upper tier of a structurally mediocre industry. Industry verdict (INTERPRETATION): the fire-services core is a structurally attractive niche — mandated demand, fragmented supply, density economics, small-ticket/high-consequence pricing — that rewards the #1 consolidator; the specialty-contracting third is a structurally average business enjoying an above-average cycle, with IIJA expiry after 2026 a dated tail-risk.

4. Competitive Position

Moat verdict: narrow-but-real in Safety; none in Specialty (INTERPRETATION, argued below).

1. Route density / local scale — the strongest claim. As developed in the Industry Dynamics section, inspection/service economics are metro-level, and the evidence that APi’s density is real is financial: the Chubb branch-optimization program converted 55 loss-making branches into fewer than five while doubling branch profitability; Safety segment margin has expanded ~90bps/yr (15.9% FY24 → 16.8% FY25, recast); and France went from single-digit to mid-teens-plus margins on flat-to-down headcount. A competitor cannot replicate this with capital alone — density is accumulated contract by contract, which is precisely why the inspection-first M&A strategy (acquire route density, then cross-sell service) is coherent.

2. Compliance switching costs — real, moderate, and unquantified. The inspection vendor holds the customer’s system records, deficiency history, and the working relationship with the authority having jurisdiction (AHJ). Switching is contractually easy (most contracts <6 months, cancellable MSAs) but practically costly: a new vendor inherits compliance risk on systems it doesn’t know, for a service that is — in management’s phrase — a small percentage of facility spend but “mission-critical, high cost-to-fail.” That is the classic low-ticket/high-consequence pricing pocket, and the observed behavior (19 consecutive quarters of double-digit organic inspection growth through May 2025; “consistent renewal rates”) is consistent with sticky demand. But APi has never disclosed a retention or renewal rate — the metric that would prove the moat is the one number management does not print. The moat claim here is behavioral and inferential, and this report treats it as such.

3. National-account coverage — a real differentiator against everyone except JCI. The National Service Group offers multi-site portfolios single-point-of-contact service across 600+ locations and 20+ countries — a capability Pye-Barker and Sciens are still assembling and locals structurally cannot offer. The land-and-expand evidence is the Davis-Ulmer pattern (alarm revenue growth via cross-sell into acquired customer bases) and the $3–4 service-per-inspection-dollar claim.

4. Technician labor — necessary, not differentiating. 29,000 employees, ~50% union/CBA-covered. Union affiliation helps prequalification for large projects and constrains flexibility (strikes, multiemployer pension exposure); versus FIX/EME’s non-union craft moats, APi’s labor position is neutral-to-positive, not superior. Management’s Q2-26 rhetoric — treating field labor as “precious gems” — is the right posture in a scarce-labor market, but it is a posture every scaled competitor shares.

5. What is NOT a moat. “AI field productivity tools,” “technology suite,” brand, and “data” are not moats here: the products installed are third-party (the product-agnostic service model is the pitch), nothing proprietary blocks a competitor, and no switching cost attaches to the software layer. Rejected per the no-fluff standard.

6. Specialty Services: essentially no moat. Beyond craft-labor access and execution reputation, this is the same fragmented, low-barrier industry the FIX/EME reports describe, currently enjoying cycle-amplified pricing. APi’s defense is mix discipline — “disciplined customer and project selection” — which has shown up as Specialty gross margin expansion (15.9% → 19.7% adjusted gross margin 2022→2024). That is good management of a mediocre asset, not a franchise.

Head-to-head (INTERPRETATION). APi is the only scaled pure-play, service-first fire platform: 69% of revenue in life safety with an inspection-first go-to-market. JCI is larger in absolute fire revenue but OEM-tied and HVAC-led; EMCOR’s fire is a segment of a construction roll-up; the PE roll-ups lack the international footprint and monitoring centers. The ENR #1 ranking at >2x the #2 substantiates the “industry leader in life safety” claim. The moat is real enough to compound on — narrow enough that execution, not franchise, determines whether the 2028 margin plan lands.

5. Growth History and Forward Opportunities

History (FACT)

Long-run, roughly two-thirds of APi’s growth is acquired: 2011–2019 organic revenue CAGR was ~7% against a ~14% reported CAGR (~18% for 2020–24). The recent decomposition (organic vs M&A+FX, per quarterly releases and the financials work-up):

Two observations. First, the FY24 organic decline was chosen — walking away from low-margin project work — and the margin trajectory (gross margin 23.8% → 31.4% over four years) shows the trade worked. Second, the quality of the current +10% organic is inverted relative to the story: Q2-26’s growth came from Specialty projects (+22.0%) while Safety grew +4.7%, and within Safety, International was flat in Q2 (returning to growth late in the quarter, per management). Becker attributed the Specialty growth to share gains and volume, not price — genuinely good news about competitive position, and simultaneously a statement that the fastest-growing revenue is project work carrying margins ~10 points below ISM.

Forward vectors (FACT for announced items; INTERPRETATION for weighting)

1. The 10/16/60+ framework (May 2025 investor day): $10B+ revenue by 2028 (4–6% organic plus ~$250M/yr of bolt-ons at ~6x EBITDA, plus optional platform deals); 16%+ adjusted EBITDA margin by 2028 (from 13.2% FY25 — a bridge of organic scale, mix, pricing, branch/field optimization, and procurement); 60%+ ISM revenue; $3.0B+ cumulative adjusted FCF through 2028. Segment-level 2028 targets: Safety $7.3B+ revenue at $1,400M+ earnings (+300bps); Specialty $2.7B+ at $375M+ (+300bps, of which 200bps+ from named SG&A/fleet/project-execution initiatives). Track record on the prior framework (“13/60/80”): the 13% margin and 80% FCF-conversion legs were delivered (13.2%, 80.3% FY25); the 60% ISM leg was not — mix sits at ~55% (2025F) versus the 60% target, which is why the new framework makes it explicit (“60+”) and long-dated.
2. ISM mix shift — the core algorithm: inspection growing double-digit organically, every inspection dollar pulling $3–4 of service, mix shift mechanically margin-accretive. This is the leg with the worst recent evidence: Safety organic +4.7% in Q2-26 and International flat.
3. Pricing — an explicit lever (“strategic pricing initiatives, sustainable pass-through of inflationary costs”); price books and inspection retention named as go-to-market systems; management described tariff and input-cost pass-through as near-real-time on the ~53–54% ISM base (Q1-26 call).
4. Elevator/escalator — Elevated ($578M, June 2024) plus Oracle/Premier/EMR bolt-ons positioned as a $1B+ opportunity in a fragmented, regulatory-driven $10B+ US services market; OEM-alternative positioning. The Q2-26 quarter included the first elevator bolt-on of 2026.
5. Specialty cyclical torque — backlog >$1.5B at May 2025 (part of the >$5B consolidated record), data-center/infrastructure exposure real but minority (~10–12% of 2026E revenue); management itself guides Specialty to decelerate on tougher 2H comps. The risk is symmetric: the same hyperscaler-capex exposure that dominates the FIX/EME bear cases, plus IIJA funding expiry after 2026.
6. M&A cadence — 14 bolt-ons for $233M in FY25; $816M cash in H1-26 alone (CertaSite $271M, Onyx-Fire $522M, seven small $47M) plus WTech ~$394M closed 7/1/26. Criteria: non-discretionary recurring services, service mix at or above baseline, low capex/high FCF conversion. Becker guided ~$250M of 2026 bolt-ons at the Q1 call — H1 spend already tripled it; the pace is accelerating, not steady-state.

Growth verdict (INTERPRETATION). The forward algorithm — mid-single-digit organic Safety compounding, cyclical Specialty torque, bolt-on layering, and margin self-help — is credible and backed by a delivered track record. The fragility is compositional: the 2028 plan needs the recurring engine to carry growth, while 2026’s actual growth is coming from the cyclical engine. If that inversion persists through the 2H-26 comp wall, revenue beats will coexist with margin misses — the exact pattern the Q1-26 −6% reaction (a beat sold on mix) previewed.

6. Financial Quality

The five-year spine (FACT — XBRL companyfacts, tied to 10-Ks; $M; share/per-share post-split)

(FY22 +66.4% revenue = Chubb, closed 1/3/22; FY23–24 GAAP EPS restated in the FY25 10-K. 3-for-2 split Q2-2025; history restated.)

The spine is the strongest evidence in the report for the “real transformation” claim: gross margin up 760bps in four years, operating income up 4x, OCF up 4.2x since FY21, capex ~1.2% of revenue, adjusted FCF conversion climbing to 80.3%. This is operating leverage from pricing discipline and mix shift onto a statutorily-anchored revenue base — not financial engineering.

Segment margins (FACT, recast basis)

Safety’s march toward the 2028 target (~21% NA / segment +300bps) is intact; Specialty’s dip to 10.7% FY25 (project starts, mix, materials) partially recovered in Q2-26. Corporate/eliminations ran −$139M FY25. SG&A ex-amortization crept from 21.1% to 21.6% of revenue FY24→FY25, driven by the “systems and business enablement” program and M&A SG&A — a cost creep to watch against the margin glidepath.

The GAAP ↔ adjusted gap — the defining QoE issue (FACT + INTERPRETATION)

Three distinct wedges separate GAAP from adjusted earnings, and they differ in legitimacy:

1. Amortization (~$240M/yr) is the structural gap — versus FY25 GAAP operating income of $554M, roughly 44% of adjusted operating profit is amortization add-back. For a serial acquirer this is a permanent feature: customer relationships ($2,114M gross) amortize over ~15 years, and every new deal adds a layer. Treating it as non-economic is defensible for a services business whose customer assets renew behaviorally; but it means GAAP operating margin (7.0% FY25) versus adjusted EBITDA margin (13.2%) will never converge while the M&A machine runs.
2. “Systems and business enablement” ($96M FY25, $52M H1-26) replaced “business process transformation” the year the latter ended. A multi-year ERP/IT program with a cash cost ($118M in FY25) is opex in substance; adding it back flatters adjusted EBITDA by ~1.2pts of margin. This is the one add-back this report flags as genuinely aggressive (INTERPRETATION).
3. The Series A preferred appreciation dividend — the per-share wedge (FACT). GAAP diluted EPS is negative in FY23–25 ((0.46)/(0.56)/(0.69) post-split) despite positive, growing net income, because the Series A Preferred — 4.0M shares held by co-chair Martin Franklin’s Mariposa Capital — is a participating security paying an annual stock dividend equal to 20% of share-price appreciation over a base, deducted in the two-class EPS calculation: $270M (FY23), $95M (FY24, plus a $372M one-time Series B conversion charge), and $590M (FY25), settled with 15,212,810 shares issued January 2026 (~3.5% of the share count). Adjusted EPS adds the dividend back and counts only the ~6M if-converted shares — so adjusted EPS excludes a real, recurring economic transfer to the preferred holder and is flattering until the instrument converts to 6.0M common shares on December 31, 2026 (a final 2026 dividend may also be due — magnitude unknown, OPEN QUESTION). GAAP EPS is the honest per-share measure through 2026; adjusted EBITDA is honest; adjusted EPS sits in between. On an economic-earnings basis the company genuinely earns ~$480M (FY25 NI $302M + after-tax amortization), growing ~20%/yr.

QoE verdict (INTERPRETATION). High-quality cash economics and honest EBITDA adjustments; per-share quality impaired until the preferred overhang clears at end-2026. GAAP OCF of $759M against $1,041M adjusted EBITDA (73% pre-capex conversion) corroborates the adjusted story; SBC ($44M FY25, ~0.6% of revenue) is modest and not added back. The flags are the enablement add-back and the preferred dividend — both known, both dated to end.

Balance sheet and cash (FACT)

Debt at YE25: 2021 Term Loan $2,157M (SOFR+175 after two repricings, due 1/3/29, no remaining quarterly amortization), 4.125% notes $337M (7/29), 4.750% notes $277M (10/29), $750M revolver undrawn; no maturity before January 2029; ~$1.1B of swaps fixing ~3.5%. Ratings BB / Ba2 (stable). 2026 activity: $500M of 5.750% Senior Notes due 2034 (May), TLB extended to 2033, revolver upsized to $1.0B with $280M drawn; first-lien net leverage 1.3x versus a 3.75x springing covenant. The company-quotes “net leverage” of 2.2x at 6/30/26 versus this report’s 2.4x TTM arithmetic — different bases (covenant pro forma adjustments), both disclosed; the direction (re-levering from 1.6x at YE25 on >$1B of M&A) is what matters. Chubb-assumed UK pensions are annuity-bought-in with wind-up targeted 2027 and a net pension asset of $129M — no material legacy liability. Goodwill impairment history: FY20 $193M (predecessor era); none since; a $12M charge on the 2023 divestiture is the only impairment in the window. Annual test October 1. The H1 working-capital pattern is a consistent seasonal use (H1-26: −$265M, OCF $168M, adjusted FCF conversion 67.7% versus the FY guide of 115% of adjusted net income — implying a very large H2 swing; seasonal precedent exists but the gap is a watch item).

Financial-quality verdict (INTERPRETATION). Real and improving economics, corroborated by GAAP cash flow; a balance sheet that is entirely a function of M&A execution (GW+intangibles 153% of equity); and per-share optics that are distorted — in opposite directions for GAAP and adjusted — until the Series A conversion on 12/31/26 makes 2027 the first clean year.

7. Capital Allocation

M&A scorecard (FACT)

Discipline verdict (INTERPRETATION). Multiples paid are full — 14.5x for Chubb pre-synergy; ~60% goodwill on the 2026 deals implies low-double-digit EV/EBITDA estimates (undisclosed; flagged as estimates). The discipline is evidenced by follow-through, not entry prices: (a) Chubb deleveraged 4.25x → 1.6x on schedule; (b) the margin/synergy thesis was delivered and the value-capture target raised 6x; © zero goodwill impairments on retained acquisitions; (d) financing mixed preferred/equity/term debt and avoided permanent balance-sheet damage; (e) the expensive 5.5% Series B preferred was retired via conversion plus a $600M repurchase at ~$24.60 split-adjusted — half today’s price. The counterpoint: CEO Becker framed 2026 bolt-on deployment as “~$250 million at attractive multiples” on the Q1-26 call; H1 actual spend of $816M (>$1.2B annualized with WTech) at the richest goodwill shares yet is a material overshoot of that framing. The pipeline is winning against the stated discipline.

Capital returns (FACT)

• Dividends: none, ever, none planned — “We intend to retain future earnings for reinvestment” (FY25 10-K).
• Buybacks: 2022 $44M (bought the post-Chubb trough at ~$11.7 split-adjusted — good timing, small size); 2024 $600M (effectively the Blackstone/Viking Series B buyout, not open-market); Q1-25 $75M; Q2-26 $66M (~1.59M shares at ~$41.6 — the first purchases under the $1B 2025 SRP, which sat essentially untapped through YE25). The authorization-to-usage gap ($1B authorized, $66M used) says management prefers M&A to its own stock at $41+.
• Share count: weighted 353M (FY23) → 402M (FY24) → 416M (FY25); 432.2M outstanding 7/23/26 — net dilution of ~23% over two years despite buybacks, driven by the Series B conversion (+~73.8M post-split shares, half repurchased), the 2024 equity offering (+28.5M), Series A appreciation-dividend shares (+3.8M Jan-25, +15.2M Jan-26), and SBC/ESPP (~2–3M/yr). The Series A converts to a final 6.0M shares on 12/31/26, ending the mechanism; total SPAC-legacy preferred cost to common holders 2023–25: ~$955M of P&L charges and ~19M shares (FACT, summed from filings; INTERPRETATION that this was a material wealth transfer to the co-chair’s entities — disclosed, contractual, and expiring).

Insider behavior (FACT — 190 Form 4s parsed, Jul 2021–Jun 2026; deduped)

• Open-market buys: essentially none. CEO Becker bought 112 shares in 2021 ($2.5k). The only meaningful buyer in 60 months is director Anthony Malkin: ~$2M at $14.47–15.50 (pre-split) during the 2022 trough, plus ~$270K at ~$39.0–39.6 in late 2025.
• Sales ~$1.1B: Viking Global ~$382M (full exit across 2024); Blackstone out via the Series B conversion and secondaries; founders Ashken ~$120M and Lillie ~$101M on quarterly 10b5-1 cadence since 2023; co-chair Franklin ~$352M and accelerating — including two discretionary, non-10b5-1 Rule 144 blocks in 2026: 3.0M shares @ $40.88 (3/19/26, $122.6M) and 2.0M @ $42.08 (6/11/26, $84.2M). Franklin retains 11.5% beneficial ownership (50.4M shares, including the Series A) plus sole voting power over another ~4.7% under a proxy agreement.
• Signal read (INTERPRETATION): the founders’ programmatic selling is diversification after a 4x run; but the two discretionary blocks — the only non-planned insider sales in the corpus — were executed at essentially today’s price by the person with the best information on the 2028 plan. That does not contradict the bull case; it declines to endorse it at $41–42.

Compensation design (FACT — DEF 14A filed 2026-04-03)

CEO Becker FY25 total comp $10.45M (salary $1.425M flat three years; stock $6.41M; STI $2.53M). STI is a single metric — annual adjusted EBITDA (threshold $932.9M / target $992.4M / max $1,052.0M; actual-for-comp $1,017.4M → 141.9% payout). LTI is 60% PSUs on 3-year cumulative adjusted EBITDA dollars plus 40% time-based RSUs. There is no ROIC, FCF, organic-growth, or relative-TSR metric anywhere in the incentive design. For a serial acquirer, cumulative EBITDA dollars can be bought; the design rewards scale and margin execution but does not police capital efficiency — a structural flaw now that the stated forward priority is continued M&A at record prices. Mitigants: FX/M&A adjustments in the STI definition, founder ownership (~17% combined), the delivered deleveraging, and say-on-pay above 95% for three straight years (97.2% in 2025) — shareholders are not objecting.

Capital-allocation verdict (INTERPRETATION). Good, with two structural caveats. The Chubb cycle — buy, lever, integrate, delever, retire the expensive paper — is one of the cleaner sponsor-era transformations on record, and buyback timing has been good when used. The caveats: the Series A appreciation dividend (expiring 12/31/26) transferred ~$955M of P&L and ~19M shares to the co-chair’s entities; and 100%-EBITDA incentives under-price capital efficiency for a company deploying $1.2B+/yr at its richest multiples yet while its $1B buyback sits idle. Deployment priority is unambiguously M&A first.

8. Changes and Headwinds — Last Two Years

Major changes (FACT, dated):

1. Segment recast (5/2/25): HVAC moved from Safety to Specialty, retroactive — improving reported Safety margins and burdening Specialty; all segment history in this report is on the recast basis.
2. 13/60/80 achieved; 10/16/60+ launched (5/21/25, with a 3-for-2 split): the 13% margin (13.2% FY25) and 80% FCF-conversion (80.3%) legs delivered; the 60% ISM leg slipped into the new long-dated framework (~55% at FY25).
3. CFO transition: Kevin Krumm resigned 11/22/24; David Jackola interim, confirmed EVP & CFO 3/31/25. Smooth, internal, no restatement attached.
4. M&A re-acceleration: from $186M (2025) to >$1.2B annualized in 2026 (CertaSite 2/2/26, Onyx-Fire 6/8/26, WTech 7/1/26) — including the first European platform bolt-on since Chubb (WTech) and the first Canadian platform (Onyx).
5. Balance-sheet re-levering and term-out: net leverage 1.6x (YE25) → 2.2x (6/30/26, company basis); $500M 5.750% 2034 notes; TLB extended to 2033; revolver upsized to $1.0B. No maturity before 2029.
6. Series A preferred endgame: the January 2026 appreciation dividend issued 15.2M shares (~3.5% dilution, $590M FY25 charge); auto-conversion to 6.0M shares on 12/31/26 ends the mechanism, making 2027 the first clean GAAP EPS year.
7. Sponsor overhang cleared; founder distribution continues: Blackstone and Viking fully exited (2024); Franklin sold ~$207M in H1-26 alone, including the two discretionary blocks.

Headwinds (FACT + INTERPRETATION):

1. The 2H-26 comp wall. Specialty organic +22.0% in Q2 laps a H2-25 that itself grew double-digit; management guides deceleration. Safety’s +4.7% and International’s flat Q2 mean the recurring engine must accelerate just as the cyclical engine decelerates — the central operating question of the next two prints.
2. Margin-mix inversion. Data-center and infrastructure projects (~10–12% of 2026E revenue, ~15–20% of Q2 backlog growth) carry margins ~10 points below ISM. The 16%-by-2028 target requires ~75bps/yr of expansion; Q1-26 delivered +70bps, Q2-26 only +10bps — the mix headwind is visible in the run-rate.
3. Project durations lengthening — from ~6–9 months to ~9–12 months (management, Q2 call): delays revenue recognition and stretches working capital against the 115% FCF-conversion guide (H1 actual: 67.7%).
4. Labor. ~50% CBA/works-council coverage, some multiemployer plans “endangered/critical,” and a skilled-labor shortage that is simultaneously the industry’s pricing umbrella and its capacity constraint.
5. IIJA funding expiry after 2026 — a dated tail-risk for the infrastructure/utility sub-segment ($1.0B FY25 revenue).
6. Macro/factor exposure: beta ~1.2–1.25 with Industrials/Infrastructure factor tilts that have been out of favor (Infrastructure −9.6% over 252d); the stock’s 12-month gain has been idiosyncratic, i.e., print-dependent.

9. Risk Analysis

Catastrophic-loss assessment (INTERPRETATION). A total loss is remote: the core demand is statutorily mandated, no customer exceeds 5% of revenue, leverage is moderate with no maturity before 2029, and the model is capital-light (capex ~1.2% of revenue) and cash-generative through the FY20 COVID year and the FY22 leverage stress alike. The realistic disaster is not ruin but a goodwill-impairment-plus-de-rating combination: a Specialty downturn that stalls margins near 14%, forces a re-test of the $3.6B goodwill balance (Oct 1 test), and compresses the multiple from ~17x toward the PRIM/DOV bracket (11–15x) — the bear-zone math in the Valuation section (~−23%) is the honest size of that scenario, not a tail.

Risk verdict. The risk that matters most is not any single line above but their correlation: the Specialty rollover (1), the margin stall (2), and the conversion miss (4) are the same event seen from three angles — a 2H-26 in which cycle-led growth decelerates before the recurring engine re-accelerates. That event is observable quarterly in the segment organic prints, which is how the falsification tests in the What Must Be True section are built.

10. Valuation Discussion

Capital structure at $41.54 (2026-08-04 close): 432.16M shares (10-Q cover, 7/23/26) → market cap $17.95B; net debt $2,672M at 6/30/26 → EV $20.62B; pro forma for WTech (~$394M cash, closed 7/1/26) EV ~$21.0B. Adjusted diluted WASO ~440M (Q2-26) for per-share math; the Series A conversion (6.0M shares, 12/31/26) is already in the adjusted count via if-converted.

Multiples (FACT)

Own-history context (FACT with distortions flagged). GAAP P/E is useless through 2026 (EPS negative FY23–25 on the preferred dividend; AZI’s trailing 54.2x prints at the 30.7th percentile because the denominator is depressed — both the level and the percentile are noise). P/B 5.18x (94.1st percentile) is inflated by M&A goodwill (tangible book is negative). P/S 2.15x (94.5th percentile) and the AZI composite (73.1st percentile) put the stock in the top quartile of its own history on every clean metric. The honest own-history frame: at the post-Chubb trough (2022, ~$9.6–10.3 split-adjusted — where director Malkin bought) APG traded at ~8.5x EV/adjusted EBITDA. The journey from 8.5x to ~17x is the re-rating; today’s price embeds the transformation as delivered and continuing. 2027 will be the first clean GAAP EPS year, which mechanically removes the negative-EPS screening artifact — a small, dated, unpriced tailwind.

Comp set (FACT — yfinance, pulled 2026-08-05 UTC ≈ 2026-08-04 close; TTM EV/EBITDA, forward P/E consensus)

(yfinance multiples are TTM with its own EV conventions — flagged for reconciliation; FactorsToday’s factor-similar peers — DOV 0.968, GTES 0.967, URI 0.967, CR 0.966, PH 0.957 — corroborate the diversified-industrial comp framing over a pure fire-safety one.)

Where APG sits (INTERPRETATION). At 16.8x FY26E EV/EBITDA, APG trades below the entire hot data-center contractor complex (EME 18.0x TTM and up; FIX 32x; PWR 36x), below JCI (23.6x) and PH, above plain industrials (DOV 15.6x, GTES 12.7x) and well above PRIM (11.4x — what a pure project contractor gets without the data-center narrative). Read two ways: (a) the discount to EME/FIX is a growth-mix discount — the market pays 32x for pure data-center torque and ~17x for a 55%-recurring blend growing ~10% organically; (b) versus DOV/GTES/JCI-on-margin, APG’s premium to plain industrials rests entirely on the 16%-by-2028 promise. The recurring-service mix argues APG should carry a quality premium over project contractors in a normal cycle; it does not today because the cycle is what the market is paying for, and APG’s cyclical exposure is its smaller, lower-margin half. That is internally consistent pricing — not an obvious mispricing in either direction.

Embedded expectations — three reverse-solves (INTERPRETATION)

1. EBITDA CAGR for a 10% IRR at a held multiple (exit YE-2030, 4.4 yrs, net debt ~$3.0B): at a held 16.8x, required 2030 EBITDA ≈ $1,804M → 10.2% CAGR from FY26E $1,225M; if the multiple compresses to 15x, 13.3% CAGR. Benchmarks: FY22→26E adjusted EBITDA CAGR was 16.1% (M&A-assisted); the 10/16/60+ path implies $1,600M+ by 2028 = 14.3% CAGR for 2026→2028. At a held multiple, the market is underwriting roughly the management plan — not heroic, not conservative; it is the plan.
2. EPS CAGR with de-rating (exit at 20x 2030 EPS, 9–10% discount): required 2030 adjusted EPS $3.03–3.16 → ~15–16% adjusted EPS CAGR from ~$1.74. Historical adjusted EPS CAGR FY21–25 was 21% (M&A + margin repair); consensus forward growth ~13–15%. The more demanding solve — it assumes the margin glidepath plus buyback/M&A accretion both land even with a 4-point multiple give-back.
3. Implied perpetual growth from FCF yield (Gordon on EV): adjusted FCF/EV of 4.2% implies ~4.8–5.8% perpetual FCF growth at a 9–10% discount rate — the market is paying for ~5% forever. For statutorily mandated recurring revenue compounding at or above nominal GDP plus bolt-ons, that is not aggressive; the load-bearing assumption is permanence of the growth algorithm, not its near-term level.

Convergence (INTERPRETATION). All three solves land on the same sentence: the price ≈ the 10/16/60+ plan delivered on schedule with no re-rating. At a held 16.8x, plan delivery is worth ~12.6% IRR to 2028 — adequate, not compelling. What is NOT priced: a re-rate toward the EME/FIX bracket (each +1x on 2028E EBITDA ≈ +$3.5/share); ISM re-acceleration (the 60% leg is stuck at ~55%); and the 2027 clean-GAAP-EPS mechanical. What IS priced and therefore fragile: ~75bps/yr of margin expansion for four straight years while the fastest-growing revenue line carries margins ~10 points below the corporate average.

Scenario zones through FY2028 (INTERPRETATION — illustrative zones, NOT price targets)

Common mechanics: exit YE-2028 (2.4 yrs); net debt held ~$3.0B (FCF funds ~$0.5B/yr of bolt-ons); ~440M shares (SRP offsets SBC plus the 6M preferred conversion); equity value = multiple × FY28E adjusted EBITDA − net debt.

• Bear: Specialty rolls over against the 2H-26 comps; organic fades to ~2–4% with M&A carrying the rest; mix stalls consolidated margin at ~14% (Safety holds ~17%, Specialty ~11–12%); the market stops paying a quality multiple for cycle-led growth → de-rate to ~13x (between PRIM and DOV). Load-bearing: that +22% Specialty organic is a cyclical peak. Counter-evidence: >$5B record backlog with 81% of RPO recognizable in 12 months makes a near-term rollover hard to time.
• Base: FY26 guide delivers; ~7%/yr revenue (mid-single organic + bolt-ons); ~75bps/yr margin expansion to ~15.25%; multiple holds ~15–16x. Load-bearing: the glidepath survives the 2H-26 comp wall. Note the base zone (~$47) sits below the Street mean PT ($52.57) — the Street is underwriting base-to-bull.
• Bull: 10/16/60+ delivered clean; ISM re-acceleration validates the recurring premium; data-center Specialty work becomes programmatic; 2027’s first clean GAAP EPS year broadens the buyer base; re-rate to 18–20x (still below FIX/PWR, at or below JCI/EME). Load-bearing: ISM mix actually reaches 60%+ — the leg with the worst recent evidence (stuck ~55%; Safety organic +4.7%).

Sensitivity (FACT, arithmetic): each 1x of exit multiple ≈ ±$3/share; each 50bps of FY28 margin ≈ ±$1.1/share.

11. Variant Perception

Consensus (FACT — MarketBeat, 2026-08-04). Moderate Buy — 1 Strong Buy / 6 Buy / 2 Hold / 0 Sell (9 analysts); mean 12-month PT $52.57 (high $55 Truist, low $47 Baird) → ~27% implied upside. Named marks: Citi Buy $54, Barclays OW $54, RBC Outperform $53, UBS $52, JPM Overweight, CJS Strong Buy, William Blair Outperform. Short interest 9.30M shares = 2.64% of float, 3.0 days to cover (2026-06-15 settlement), up 27% month-over-month — rising but not crowded in either direction.

The consensus belief (INTERPRETATION). “The transformation is working; APG is a quality compounder with a visible margin glidepath and a record backlog; 16.8x is fair for 55%-recurring revenue growing 10% organically.” The mean PT of $52.57 sits between this report’s base (~$47) and bull (~$66) zones — i.e., the Street prices the plan delivered and a partial re-rate.

The strongest disagreement (FACT/INTERPRETATION) — and its tape-level expression. The debate is not demand; it is (a) whether 16% margins are reachable while the fastest-growing revenue is the lowest-margin, and (b) how much of current organic growth is cyclical project pull versus durable service compounding. Its most interesting expression is not the (small) short interest but the insider tape: co-chair Franklin sold ~$206M in two discretionary blocks at $40.88 (3/19/26) and $42.08 (6/11/26) — essentially today’s price — while the Street models $52+. No officer has bought open-market since 2022; the only meaningful insider buying in five years was Malkin at the 2022 trough and ~$39 in late 2025. Insiders with the best information on the 2028 plan are monetizing where the sell-side sees 27% upside. Both can be right (estate planning versus fundamentals) — but the tape does not endorse the consensus upside.

Bull case (as underwritten by Truist $55 / Citi $54 / Barclays $54). Organic durability (10.1% in Q2 versus a mid-single-digit long-term algorithm); margin credibility from self-help (pricing, project selection, procurement) rather than mix; record >$5B backlog with better margin quality than a year ago; leverage of 2.2x with >$1B of M&A absorbed and maturities termed out to 2033/2034; an inspection-first M&A machine (140 acquisitions since 2005) compounding route density; and a 2027 clean-GAAP-EPS catalyst. If ISM re-accelerates and the cycle holds through the comps, the base zone is the floor and the bull zone is the destination.

Bear case (as expressed by the rising short and the April–July tape). The bar is high — Q1-26 beat and fell 6%; 16%-by-2028 requires ~75bps/yr against a mix headwind the bulls are simultaneously citing as the growth driver; Specialty +22% laps brutal 2H comps; data-center exposure (~10–12% of revenue) is the cyclical piece to haircut; leverage went 1.8x → 2.2x in a quarter; goodwill plus intangibles are 153% of equity; GAAP EPS is negative through 2026; and the founders are selling at $41. In that world the multiple migrates toward the project-contractor bracket and the bear zone (~$32) is the destination.

Where this report lands (INTERPRETATION, no position — the position lives only in Kimi’s Take). The disagreement is genuine and resolvable by data within two quarters: the 2H-26 segment organic prints answer whether the recurring engine can carry growth as the cyclical engine decelerates. The weight of evidence favors the transformation being real (delivered targets, clean deleveraging, cash-corroborated earnings) and the price being the problem (the plan already underwritten; the insider tape flat; the momentum stall unresolved). Both sides are arguing about the same number — the durability of Specialty-led growth into the comps — and both will know much more by the Q3-26 print.

12. Fact vs. Interpretation

13. Open Questions

1. Retention/renewal rate — never disclosed. The single most important moat metric (inspection/service contract retention) is asserted (“consistent renewal rates”) but has never been quantified in any filing, deck, or available transcript. Direct question for IR; resolution would either harden or puncture the moat claim.
2. The FY26 FCF-conversion swing. Guide: 115% of adjusted net income. H1-26 actual: 67.7%. Seasonality is the normal pattern, but the implied H2 swing is very large, project durations are lengthening (9–12 months from 6–9), and the conversion denominator changed (from adjusted EBITDA in 2025 to adjusted net income in 2026), complicating comparability. Q3-26 cash flow is the first real test.
3. No verbatim Q2-2026 transcript yet. The Fool typically lags ~1 month; Seeking Alpha is paywalled; GuruFocus blocked. This report used the MarketBeat call highlights plus the EX-99.1 release; re-check early September 2026 — call nuance on Safety organic re-acceleration and data-center margins could move the growth-quality debate.
4. CertaSite / Onyx-Fire / WTech multiples undisclosed. Goodwill shares of ~57–62% imply low-double-digit EV/EBITDA estimates — richer than the “~6x bolt-on” framing in the 10/16/60+ algorithm. If 2026’s $1.2B was deployed at materially richer multiples, the M&A leg of the plan carries more execution risk than modeled.
5. The unexplained 2026-06-10 −5.3% session. No company filing or news identified (the Onyx close was 6/8, and the stock was +3.5% on 6/9). Attributed to market/flow, unconfirmed — as is the 2025-02-19 +7.8% up-day, whose driver was never identified.
6. Magnitude of the final 2026 Series A appreciation dividend. 20% of 2026 price appreciation × 211.8M shares, stock-settled, payable around the 12/31/26 conversion — could be another multi-hundred-million-dollar charge and further share issuance before the mechanism ends. Treatment to be confirmed on the Q3/Q4-26 calls.
7. Data-center exposure granularity. Management’s ~10–12% of 2026E revenue is a “rough estimate”; the split between Safety data-center projects and Specialty is not disclosed. Given that this exposure drives both the bull (programmatic demand) and bear (cyclical peak) cases, its true size matters more than its current precision.
8. Safety NA vs International split. The 10-K aggregates the two Safety operating segments; International’s flat Q2-26 organic cannot be tracked in filings between prints. Investor-day 2024 mixes are the last granularity.
9. Net leverage definition. The 10-Q discloses first-lien covenant leverage (1.3x); the FY25 release quoted “net leverage” 1.6x; this report’s TTM arithmetic gives 2.4x. All defensible on different bases — but the company should be asked to standardize the headline figure.

14. What Must Be True

Bull case — what must be true: (i) the recurring engine re-accelerates — Safety/ISM organic growth returns to high-single-digits by mid-2027, with International Safety back to growth, proving the inspection annuity compounds independent of the project cycle; (ii) Specialty holds growth against the 2H-26 comps — data-center and infrastructure awards become programmatic multi-year work rather than a 2026 spike, keeping consolidated organic ≥7%; (iii) the margin glidepath delivers ~75bps/yr through the mix headwind, reaching ~16% by 2028 with the $10B+ revenue leg intact; (iv) 2027’s first clean GAAP EPS year (post-preferred) broadens the buyer base and the market re-rates APG toward the EME/JCI bracket (18–20x+). Falsification test: two consecutive quarters through mid-2027 in which Safety organic stays ≤~5% while Specialty organic decelerates below ~10% and consolidated adjusted EBITDA margin expansion runs under ~50bps/yr annualized — that combination proves growth was cycle-led, the glidepath is hostage to mix, and the bull zone (~$66) is dead.

Bear case — what must be true: (i) Specialty’s +22% organic was a cyclical peak — organic rolls toward zero or negative through the 2H-26 comps as data-center awards slow and IIJA funding expires, overwhelming mid-single-digit Safety growth; (ii) consolidated margin expansion stalls below ~14% as lower-margin project mix persists and the enablement/SG&A creep continues; (iii) the 115% FCF-conversion guide is missed, exposing the lengthening project-duration working-capital drag; (iv) the market re-classifies APG from “recurring-service compounder” to “levered project contractor with a goodwill-heavy balance sheet,” de-rating the multiple toward ~13x. Falsification test: the >$5B backlog converts on schedule — Specialty organic holds positive double-digit growth through Q4-26 and consolidated margin expansion re-accelerates to the ~75bps/yr glidepath and FY26 FCF conversion lands ≥110% — in which case the demand break the bear case requires did not happen, backlog quality was real, and the bear zone (~$32) is unreachable absent a macro shock.

15. Source Appendix

Full source detail — filings with accession numbers, transcripts, data tools, and news URLs — is in Appendix B below. Primary load-bearing sources: FY2025 10-K (filed 2026-02-25, accession 0001628280-26-011620); Q2-2026 10-Q (filed 2026-07-30, accession 0001628280-26-051007); Q2-2026 earnings 8-K EX-99.1 (filed 2026-07-30, accession 0001628280-26-050806); DEF 14A (filed 2026-04-03, accession 0001628280-26-023528); Chubb close 8-K (2022-01-03, accession 0001193125-22-000650); the 2025 Investor Day deck (2025-05-21); Motley Fool verbatim Q4-25 and Q1-26 call transcripts and MarketBeat Q2-26 call highlights (no free verbatim Q2-26 transcript exists yet — see Open Questions); the 60-month EDGAR corpus (405 filings indexed, 190 Form 4 XMLs parsed); AZI price/valuation data and FactorsToday factor data pulled 2026-08-04; and the author’s prior analysis of peer companies (JCI/FIX/EME/ACM).

End of memo. The analysis above carries no recommendation and no price target; the single labeled exception is Kimi’s Take at the top.

APPENDIX A — Standard Diligence Questionnaire

Supplemental to the research memo (2026-08-04). Every question of the standard diligence questionnaire answered for APG, with Fact / Interpretation / Assumption labels where it matters. Sector analogs substituted where a question does not map to a services business.

General

What thoughtful questions have other investors asked about this company?

From the Q1/Q4-2025 and Q2-2026 earnings calls (verbatim Motley Fool transcripts and MarketBeat call highlights) and the sell-side framing in the news corpus:

1. How big is the data-center exposure, really, and how durable is it? Management’s answer: ~10–12% of 2026E revenue, ~15–20% of the Q2-26 backlog increase (CFO Jackola, “rough estimate”), with data-center fire projects inflating from $7–8M to approaching $20M each. The follow-up investors keep pressing: is this programmatic multi-year demand or a 2026 spike? Management’s posture is deliberately disciplined — not “pushing all the chips” on data centers, selective project acceptance.
2. Is the 16%-by-2028 margin target reachable while project mix dilutes? Project margins run ~10 points below ISM; Q2-26 delivered only +10bps of margin expansion against the ~75bps/yr the glidepath needs. Management’s answer: self-help (pricing, procurement, branch/field optimization, project selection) rather than mix; Q1’s +70bps as the truer run-rate.
3. Will Specialty decelerate on 2H comps — and can Safety re-accelerate to offset? Management pre-emptively guided deceleration on the Q1 call; Q2 delivered Specialty +22% organic versus Safety +4.7% with International flat. This is the central operating debate of the next two prints.
4. Backlog quality and duration. Analysts probed whether the record >$5B backlog is margin-accretive (answer: better margin mix than a year ago) and noted durations extending to ~9–12 months from 6–9 — a working-capital/recognition watch item.
5. Capital deployment priority — why is the $1B buyback authorization essentially unused while M&A runs at 3x the guided pace (~$250M guided; >$1.2B annualized actual)? Management’s framing: M&A first at “attractive multiples,” buybacks opportunistic.
6. The Series A preferred endgame — the size of the final 2026 appreciation dividend and the mechanics of the 12/31/26 conversion (a recurring question since the $590M FY25 charge).

Cyclicality & Earnings Nature

Are earnings at a cyclical high or low? Mixed by segment (INTERPRETATION). The Safety Services core (69% of revenue) is statutorily driven and mid-cycle at worst — inspection demand does not cycle with GDP. Specialty Services (31%) is at or near a cyclical high: +22.0% organic in Q2-26 on the data-center/infrastructure capex wave, with management itself guiding deceleration against 2H comps. Consolidated adjusted EBITDA margin (13.8% Q2-26) is at an all-time company high but below the 16% 2028 target — so reported earnings are at a cyclical high for this company, with the mix question being how much of the increment is cycle-led.

Driven by the external environment or internal actions? Both, separable (INTERPRETATION). Internal: gross margin +760bps over four years from pricing discipline, project selection, and Chubb branch optimization — demonstrably self-help. External: the current growth rate (10%+ organic) leans on the AI/data-center/IIJA demand shock and the skilled-labor shortage that gives the whole specialty-contracting sector pricing power (per prior analysis of Comfort Systems, EMCOR, and Arcosa). FY24 proved the internal levers work against the cycle — organic revenue declined 0.9% (deliberate low-margin shed) while margins expanded.

How stable are revenues? High stability in the ~55% ISM base (FACT: code-mandated annual inspections; “consistent renewal rates”; no customer >5% of revenue). Moderate stability in the project book (RPO $4,066M at 6/30/26, 81% recognizable within 12 months; backlog cancellable per the 10-K’s own risk factor). Historical revenue has never declined year-over-year in the Chubb era; the FY24 +1.3% was the low-water print.

Outlook for products/services? (FACT + INTERPRETATION.) Secular tailwinds: code tightening, insurance requirements, aging installed base, monitoring/connected services (50M+ connected devices, 14 centers), elevator services ($10B+ fragmented US market, regulatory-driven). The product layer is third-party — APi is deliberately product-agnostic, which avoids OEM obsolescence risk but caps product-level differentiation.

How big will this market be — growing, shrinking, domestic or international? Company-estimated TAM ~$160B+ (global fire ~$60B+, electronic security ~$90B+, US elevator services ~$10B+) — soft-sourced, framing-grade (FACT that these are management estimates “based on various market studies”). The market is growing low-to-mid-single-digit structurally with code-tightening optionality, and it is global: APi is 65% US, 9% France, 26% other international (FACT, FY25 10-K). Consolidation, not market growth, is the company’s main share-gain engine — the fire-services trade below APi and EMCOR is atomized across thousands of local shops (ENR, Oct 2025).

Business Quality & Competitive Moat

Is the industry getting more or less competitive? Less at the top, unchanged at the bottom (INTERPRETATION on FACTS). Carrier exited fire/security entirely in 2024–25 (Global Access Solutions → Honeywell; industrial fire → Sentinel; Kidde divested), removing a scaled OEM competitor. PE-backed roll-ups (Pye-Barker, Sciens, Impact Fire/AI Fire) are consolidating the fragmented tail — which both raises bolt-on acquisition prices and validates APi’s strategy. JCI treats fire as one line in an HVAC-led portfolio. Net: APi’s relative position strengthened over the window.

How profitable is the business (ROIC, ROE)? (FACT + INTERPRETATION.) FY25: adjusted EBITDA margin 13.2%, GAAP operating margin 7.0%, GAAP net margin 3.8%, ROE ~9% (NI $302M on $3,408M equity). ROIC is the honest weak point: invested capital is goodwill-heavy (GW + intangibles $5.4B = 153% of equity at 6/30/26), so returns on gross invested capital are modest (roughly high-single-digits on adjusted operating profit over debt + equity) while returns on tangible capital are very high — the classic serial-acquirer services profile. There is no ROIC metric in the incentive plan, which is why this report flags it (see Capital Allocation below).

How profitable is the industry — how many competitors, what barriers to entry? Fire services: attractive niche economics (route density + mandated demand) but “relatively few barriers to entry” in the 10-K’s own words; the barriers that bind are licensing/certification, insurance/safety prequalification, bonding/working capital, and accumulated density. Two >$1B platforms (APi $3.5B, EMCOR $1.6B fire revenue), then a cliff to sub-$300M regionals (ENR). Specialty contracting: fragmented, low-barrier, mediocre structure enjoying a hot cycle — EMCOR ~9% operating margin/~31% ROIC marks the upper tier.

Can the business be easily understood? Yes (INTERPRETATION). Inspect what the law requires be inspected; service what you inspect; monitor what stays connected; and run a project business on the side. The accounting (preferred dividends, PPA amortization, add-back taxonomy) is materially harder than the business.

Can it be undermined by foreign low-cost labor? No (INTERPRETATION). Inspection and service are physically local, licensed, and code-governed; the labor moat threat runs the other way (domestic skilled-labor scarcity as a cost/constraint). Monitoring centers could theoretically migrate offshore; the 50M+ connected-device estate and AHJ relationships are local by nature.

Do brands matter? Little at the corporate level (INTERPRETATION). The acquired brand equity (Chubb in Europe, Davis-Ulmer regionally) carries local weight, but purchasing is driven by compliance relationships, national-account contracts, and response capability — not consumer-style brand preference. The product-agnostic model explicitly de-emphasizes product brands.

What is the nature of competition? Local route competition in service (density and response time), relationship/prequalification competition in national accounts, and price-plus-execution competition in projects. Not price-war prone in the inspection core (small-ticket, high cost-to-fail); more price-elastic in Specialty projects.

Customers’ switching costs? Real but moderate (INTERPRETATION). Contracts are short (majority <6 months) and MSAs cancellable — contractual switching costs are near zero. Practical switching costs: the incumbent holds system records, deficiency history, and the AHJ relationship; a switch imports compliance risk for a service that is a small fraction of facility spend. The missing disclosure: APi has never published a retention/renewal rate, so switching-cost strength is inferred from 19 consecutive quarters of double-digit organic inspection growth and “consistent renewal rates” language — behavioral evidence, not a cohort metric.

Financial Condition & Balance Sheet

Assets not fully recognized on the balance sheet? The assembled workforce and route density, and the installed-base/AHJ relationships, are unrecognized intangibles (INTERPRETATION). Conversely, the recognized intangibles ($1.7B net) and goodwill ($3.6B) are real only to the extent M&A execution continues to deliver — the balance sheet is 153% goodwill-plus-intangibles relative to equity (FACT, 6/30/26).

Off-balance-sheet liabilities? Nothing material identified (FACT, from the filings corpus): operating leases are on-balance-sheet under ASC 842; self-insurance retentions are accrued ($125M); multiemployer pension withdrawal exposure exists but is disclosed qualitatively (some plans “endangered/critical”); Chubb-assumed UK DB pensions were annuity-bought-in with wind-up targeted 2027 and a net pension asset of $129M; legacy Chubb claims liabilities came with a purchase-price adjustment. The Series A preferred (converts 12/31/26) is on the face of the equity section, not hidden.

How conservative is the accounting? Middle-of-the-road for a serial acquirer (INTERPRETATION). Conservative elements: no SBC add-back in adjusted EPS; GAAP OCF corroborates adjusted EBITDA (73% pre-capex conversion FY25); restructuring programs are dated and closed out; zero goodwill impairments. Aggressive elements: “systems and business enablement” ($96M FY25, $118M cash) added back as if one-time; adjusted EPS adds back the preferred appreciation dividend (a real economic transfer) until end-2026; the FY26 FCF-conversion guide changed denominators (adjusted EBITDA → adjusted net income), which flatters the optics of the 115% target.

How CapEx-hungry is the business? Not at all (FACT): capex $96M FY25 = ~1.2% of revenue (“<1.5% typically” per management); $106M TTM. The capital intensity is acquisition goodwill, not plant.

Capital Allocation & Management

How much FCF does the business generate, how does management use it, what is the philosophy? (FACT.) FY25: OCF $759M, capex $96M, FCF $663M; adjusted FCF $836M (80.3% of adjusted EBITDA). FY26 guide: ~115% conversion of adjusted net income (~$859M on derived numbers — ASSUMPTION on the derivation). Use: M&A first (FY25 $186M; H1-26 $816M plus WTech ~$394M), buybacks opportunistic ($66M H1-26 against a $1B authorization), no dividend ever, leverage held around 2x with a 2.5–3.0x long-term ceiling. Philosophy (INTERPRETATION): compounding through inspection-first consolidation at disciplined underwriting — though the 2026 deployment (>3x the guided ~$250M, at ~57–62% goodwill) tests the “disciplined” adjective.

Significant acquisitions recently? (FACT.) Chubb ($3.1B, closed 1/3/22, ~14.5x pre-synergy EBITDA — the transformation); Elevated Facility Services ($578M, 6/3/24 — elevator services); 2026: CertaSite ($271M, 2/2/26), Onyx-Fire ($522M, 6/8/26), WTech Fire Group (~$394M, 7/1/26) — >$1.2B deployed in six months, the richest prices yet (undisclosed multiples; ~57–62% goodwill implies low-double-digit EV/EBITDA estimates).

Buying back shares? (FACT.) Yes but modestly and well-timed: $44M at the 2022 trough (~$11.7 split-adjusted); $600M in 2024 (effectively the Blackstone/Viking Series B buyout at ~$24.60); $75M Q1-25; $66M Q2-26 at ~$41.6. The $1B 2025 SRP sat essentially untouched through YE25 — at $41+, management prefers acquisitions to its own stock.

Issuing large amounts of new shares to insiders? Not options-style dilution, but the Series A preferred appreciation dividend is functionally a stock transfer to the co-chair’s Mariposa entities: $270M/$95M/$590M charges FY23–25, settled with 3.8M (Jan-25) and 15.2M (Jan-26, ~3.5% of the count) shares; the instrument converts to a final 6.0M common shares on 12/31/26 (FACT). It is a disclosed, contractual SPAC-legacy claim — and it is ending. SBC is modest ($44M FY25, 0.6% of revenue, not added back). Share count nonetheless rose ~23% in two years (Series B conversion, 2024 equity offering, preferred dividends).

Compensation policy of directors/management? (FACT, DEF 14A 2026.) CEO Becker FY25 total $10.45M (salary $1.425M flat three years; stock $6.41M; STI $2.53M). STI: single metric — annual adjusted EBITDA (threshold/target/max $932.9M/$992.4M/$1,052.0M; FY25 actual-for-comp $1,017.4M → 141.9% payout). LTI: 60% PSUs on 3-year cumulative adjusted EBITDA dollars, 40% time RSUs. No ROIC, FCF, organic-growth, or relative-TSR metric anywhere (INTERPRETATION: a design flaw for a serial acquirer — cumulative EBITDA dollars can be bought). Say-on-pay 97.2%/98.5%/95.5% (2025/24/23) — shareholders are not objecting.

Motivations of management? (INTERPRETATION on FACTS.) Becker (1.2% owner) is compensated on EBITDA scale and margin — consistent with the delivered margin plan and the M&A cadence. The founders (Franklin 11.5% + ~4.7% voting proxy; Ashken 2.8%; Lillie 2.7%) retain ~17% combined — real skin in the game — but are systematically monetizing: ~$1.1B of sales over five years, including Franklin’s two discretionary 2026 blocks at $40.88/$42.08. The motivation read: builders who have mostly built, harvesting a 4x outcome — not promoters still buying their own story.

Valuation & Market Data

Is the stock an ADR, MLP, or K-1 issuer? No (FACT) — US domestic filer, NYSE-listed common, single class one-vote (plus the Series A preferred, 4.0M shares, one vote each, converting 12/31/26).

Dividend policy? (FACT.) None, ever: “We have historically not paid cash dividends and do not currently anticipate paying a cash dividend on our common stock” (FY25 10-K). Total-return vehicle via reinvestment and opportunistic buybacks.

How profitable is the business? (FACT.) FY25 adjusted EBITDA $1,041M (13.2% margin); TTM $1,122M (13.3%); FY26 guide $1,205–1,245M (~13.6–13.8% at midpoint). GAAP net income $302M FY25, $346M TTM. Adjusted net income $627M FY25, $696M TTM. At $41.54: 16.8x FY26E EV/adjusted EBITDA, ~23.9x FY26E consensus adjusted EPS, 4.8% adjusted FCF yield on market cap.

Is net income diverging from cash from operations? Divergence runs in the benign direction for EBITDA and the cautionary direction for per-share (INTERPRETATION on FACTS). GAAP OCF ($759M FY25) comfortably corroborates adjusted EBITDA ($1,041M) at 73% pre-capex conversion. But GAAP EPS is negative ((0.69) FY25) while adjusted EPS is $1.48, because the Series A appreciation dividend ($590M FY25) is charged below net income in the two-class calculation — a real, stock-settled transfer that adjusted EPS ignores until the 12/31/26 conversion. H1-26 FCF conversion (67.7%) running far below the FY guide (115%) is the other divergence to watch — seasonal, but the implied H2 swing is very large.

Risks & Downside

What factors would cause the stock to decline? (INTERPRETATION, ranked; full matrix in the Risk Analysis section.) (1) Specialty organic rolling over against the 2H-26 comps — the single highest-beta print risk; (2) margin expansion stalling below the ~75bps/yr glidepath as lower-margin project mix persists; (3) an FCF-conversion miss exposing lengthening project durations; (4) a goodwill/intangible impairment (Oct 1 annual test; GW+intangibles 153% of equity); (5) further founder distribution; (6) multiple compression from 16.8x toward the plain-industrial bracket if the recurring-growth narrative breaks. The Q1-26 precedent — a beat sold for −6% on mix — shows the expectation bar does the amplifying.

Risk of a catastrophic loss? Remote (INTERPRETATION). Demand is statutorily mandated; no customer >5% of revenue; capex 1.2% of revenue; leverage 2.2x with no maturity before 2029, first-lien covenant headroom wide (1.3x vs 3.75x springing), ~$1.1B of debt swapped to ~3.5% fixed; the model stayed cash-generative through FY20 (COVID) and FY22 (4.25x leverage into a rate shock). The honest disaster case is an impairment-plus-de-rating: a Specialty downturn stalling margins near 14%, forcing a goodwill re-test and a multiple migration toward 13x — worth roughly −23% (the Valuation section’s bear zone), not ruin.

Chance of a total loss? Negligible absent fraud or a multi-year depression in construction and code enforcement simultaneously. The five-year maximum drawdown is −49.6% (2022) — deep but a de-rating, not a solvency event; the company deleveraged through it.

Recent News & Events

Has the business environment changed recently? (FACT.) Yes, favorably at the top of the funnel: the AI/data-center/infrastructure capex wave is inflating project sizes (data-center fire projects from $7–8M toward ~$20M) and filling backlog to a record >$5B; Carrier’s 2024–25 exit from fire/security removed a scaled competitor; IIJA infrastructure funding runs through 2026 — and then expires, a dated tail-risk.

Significant acquisitions? (FACT, dated.) CertaSite $271M (2/2/26, Midwest inspection-first, ~$90M revenue); Onyx-Fire $522M (6/8/26, Canada, ~$190M revenue); WTech Fire Group ~$394M (7/1/26, Ireland-based, ~$175M revenue); seven immaterial deals $47M — >$1.2B deployed in H1-26, each described as accretive to the 10/16/60+ targets. Preceded by Elevated ($578M, 2024) and 14 bolt-ons for $233M (2025).

Change in accounting policies? (FACT.) None adverse. Two presentation changes worth noting: the segment recast (HVAC Safety→Specialty, retroactive, 5/2/25) and the FY26 FCF-conversion denominator change (adjusted EBITDA → adjusted net income), which makes the 115% guide not directly comparable to the prior 80%-of-EBITDA framework (normalization: ~115% × adjusted NI ≈ ~80% × adjusted EBITDA at current leverage).

Recent changes — new markets, facilities, management? (FACT.) New geographies via M&A: first Canadian platform (Onyx) and first European platform bolt-on since Chubb (WTech). First elevator bolt-on of 2026 (the $1B+ elevator thesis in motion). Management: CFO transition completed — Kevin Krumm resigned 11/22/24; David Jackola interim then confirmed EVP & CFO 3/31/25; IR leadership transitioned Adam Fee → Adam Walters (2/25/26). No C-suite changes since. Financing: $500M 5.750% 2034 notes (May 2026), TLB extended to 2033, revolver upsized to $1.0B. The standing strategic event of the next five months is the Series A auto-conversion on 12/31/26, which ends the preferred appreciation-dividend mechanism and makes 2027 the first clean GAAP EPS year.

APPENDIX B — Source Appendix

Companion to the research memo (2026-08-04). All sources used in this report. Primary sources before secondary.

1. SEC Filings (CIK 0001796209; 60-month corpus: 405 filings indexed via EDGAR full-text/submissions API; 190 Form 4 XMLs parsed)

Periodic reports (load-bearing):

Earnings releases (8-K EX-99.1 exhibits):

Guidance / M&A 8-Ks:

Ownership / insider:

• Schedule 13D/A (Franklin / Mariposa entities): 2025-01-03, 2026-01-02, 2026-05-06 — Franklin 50.4M shares beneficial (11.5%), plus ~4.7% voting proxy.
• Form 3/4/5 corpus: 195 insider filings indexed; 190 Form 4 XMLs parsed and deduped on exact keys (485 unique in-window transactions).
• 424B2 prospectuses: 2021-09-16 (19.75M shares @ $20.25, Chubb funding); 2024-04-18 (11.0M shares @ $37.50).
• EDGAR companyfacts XBRL (CIK 1796209): five-year fundamentals spine, reconciled to 10-K values exactly.

2. Earnings-Call Transcripts

3. Company IR Materials

• APi Group 2025 Investor Day deck, 2025-05-21 (via the company IR site) — 10/16/60+ framework, ISM mix 40%→54%, service/project mixes by segment, $3–4 service-per-inspection-dollar, TAM estimates, Chubb value-capture/branch economics, 2028 segment targets.
• Long-term targets press release, 2025-05-21 — https://ir.apigroupcorp.com/News/press-releases/news-details/2025/APi-Group-Announces-New-Long-Term-Financial-Targets/default.aspx
• Chubb acquisition investor deck, 2021-07-27 — ~13.3x with-synergies multiple framing, 4.25x pro forma leverage, <3.0x deleveraging goal.
• IR press-release index (ir.apigroupcorp.com); SEC EX-99.1 exhibits used as the release text of record.

4. Third-Party Data Tools (labeled third-party; caveats in the Valuation section)

• AZI (azitrading.com) — daily OHLCV 2017-10-11 → 2026-08-04; valuation percentiles (pulled 2026-08-04): composite 73.1st, P/E 30.7th, P/B 94.1st, P/S 94.5th percentile. Caveat: CSV carries a 1.5 split factor dated 2025-07-01; adjusted series is internally consistent (returns unaffected) but noted against the real 3-for-2 split (Q2-2025).
• FactorsToday API (pulled 2026-08-03/04) — factor loadings (Base R²=0.427, All-Factors R²=0.498; Momentum L1-zeroed; Industrials +0.554, Infrastructure +0.264), annualized risk-adjusted track record, stock-specific volatility (21.7% annualized), related stocks (DOV/GTES/URI/CR/PH), and factor-return history.
• yfinance (pulled at the 2026-08-04 close) — comp multiples for EME/FIX/PWR/PRIM/URI/DOV/GTES/PH/JCI and APG stats; TTM-based with yfinance’s own EV conventions (flagged; reconciled to filings where load-bearing — APG evToEbitda 18.7x matches the pro forma TTM figure).
• MarketBeat (pages refreshed through 2026-08-04) — consensus snapshot: https://www.marketbeat.com/stocks/NYSE/APG/price-target/ (Moderate Buy, 9 analysts, mean PT $52.57, range $47–55); short interest: https://www.marketbeat.com/stocks/NYSE/APG/short-interest/ (9.30M shares, 2.64% float, 3.0 DTC, 2026-06-15 settlement). Q2 consensus EPS ~$0.41 via stockevents.app (secondary aggregator — flagged).

5. Prior Analysis of Peer Companies

Used for industry cross-read (fire-services structure, specialty-contracting cycle, value-chain margin hierarchy, peer multiples) — the author’s prior analysis, not external sources:

• Johnson Controls (prior analysis, 2026-06-13) — fire/security industry framing (“a scale game layered onto code-driven, recurring inspection revenue”), JCI’s own fire margin catch-up thesis.
• Comfort Systems (prior analysis, 2026-06-12) — US MEP market structure, craft-labor constraint, data-center cycle risk, ~50x trailing P/E bracket.
• EMCOR (prior analysis, 2026-06-21) — fire segment $1.6B, ~26% recurring mix, data-center RPO +60% to $4.46B, ~9% operating margin / ~31% ROIC upper-tier benchmark.
• Arcosa (prior analysis, 2026-07-31) — value-chain margin hierarchy (self-perform trades > asset-light design > lump-sum EPC); peer multiples (ACM 10.8x fwd P/E / 9.2x EV/EBITDA; EME ~20x EV/EBITDA; FIX ~50x trailing P/E).

6. News & Market Sources (URLs as captured)

• Carrier completes $3.1B Chubb sale (2022-01-03) — https://www.corporate.carrier.com/news/news-articles/202201_carrier-completes-31b-sale-chubb-fire-security-business-api-group-corporation.html
• StockStory Q2-2026 beat coverage (2026-07-30) — https://stockstory.org/us/stocks/nyse/apg/news/earnings/apis-nyseapg-q2-cy2026-beats-on-revenue-stock-soars
• MarketBeat Q2-2026 call highlights (2026-08-01) — https://www.marketbeat.com/instant-alerts/api-group-q2-earnings-call-highlights-2026-07-30/
• ENR Top 600 Specialty Contractors, October 2025 (2024 fire-protection revenue ranking: APi $3,509M #1, EMCOR $1,620M #2, J.F. Ahern $286M #3).
• Analyst-action timeline (JPM 2025-07-15; Zacks 2025-09-15; CJS 2025-12-11; RBC 2026-02-26; Truist 2026-05-01; Barclays 2026-05-07; Weiss 2026-06-11; UBS 2026-07-20; Citi and Baird 2026-07-31) — aggregated via MarketBeat/news sweep; individual notes not retrieved in full.
• GuruFocus Q2 page — GF Value “overvalued” flag noted in the bear framing; secondary source.

Data Quality & Reconciliation Notes

1. Filing-reconciled figures govern. Where aggregators conflict with filings (yfinance EV conventions, AZI percentiles, stockevents.app consensus), the 10-K/10-Q/8-K values were used; aggregator figures are labeled and flagged in the memo.
2. Net leverage definitions differ by source: 10-Q covenant first-lien 1.3x; company “net leverage” ~2.2x at Q2-26 (per management commentary); this report’s TTM arithmetic 2.4x ($2,672M / $1,122M). All three appear in the memo with their bases.
3. FY26 adjusted EPS is not company-guided (verified in the Q2-26 EX-99.1: guidance covers revenue, adjusted EBITDA, and FCF conversion only). All FY26/FY27 EPS figures are consensus (~$1.74 / ~$1.95–2.00) or derived (~$1.70) and labeled ASSUMPTION.
4. Split basis: 3-for-2 stock split effected Q2-2025; all historical share/per-share figures in this report are post-split unless explicitly labeled pre-split.
5. Form 4 corpus: duplicate-content files from two fetch passes verified identical by md5; transactions deduped on (owner, code, date, shares, price) before aggregation. Related-entity double-reporting (e.g., FD Juno parent/subsidiary) corrected where identified (Blackstone 8/14/23: $71.4M economic, not $142.7M).
6. Unresolved data gaps (carried in the Open Questions section): no verbatim Q2-26 transcript; CertaSite/Onyx/WTech multiples undisclosed; the unexplained 2026-06-10 −5.3% session and 2025-02-19 +7.8% session; final 2026 Series A dividend magnitude; Safety NA/International intra-year split.