Rockwell Automation

Phase A — Understand the business

Company Overview

Rockwell Automation is the largest pure-play industrial-automation company in the world — and the operative word is pure-play. Where Siemens, ABB, Schneider, Honeywell and Emerson bolt automation onto sprawling conglomerates (grid, power, building tech, aerospace), Rockwell sells one thing: the hardware, software and services that make a factory run. It traces an unbroken line to the Allen-Bradley Company, founded 1903, and was spun out of Rockwell International in 1996 when the aerospace/defense business went to Boeing. That lineage matters for the thesis: Rockwell is structurally barred from defense (it divested it) — which suits a no-defense coverage mandate exactly.

The business is plumbing for the physical economy. Its installed base of programmable logic controllers (PLCs) — the Logix control architecture with Allen-Bradley hardware — sits inside tens of thousands of plants, and the entire value proposition is that once a line is engineered around ControlLogix/CompactLogix and FactoryTalk software, ripping it out is a multi-year, multi-million-dollar risk no plant manager volunteers for.

Three segments:

• Intelligent Devices (FY25 $3,756M, 45% of sales) — drives, motion, safety, sensing, industrial components, and the OTTO/Clearpath autonomous mobile robots (the literal "robotics" of the coverage bucket). The hardware spine.
• Software & Control (FY25 $2,383M, 29%) — the Logix control platform, visualization, FactoryTalk, digital-twin/simulation, Plex (cloud MES) and Fiix (maintenance), network & cyber-security (Verve). The highest-margin, fastest-growing, most strategically important segment.
• Lifecycle Services (FY25 $2,203M, 26%) — engineered-to-order solutions, digital consulting, recurring cyber/safety/asset-management services, and (until April 2026) the Sensia oil-&-gas JV with SLB.

Go-to-market is distributor-led and concentrated. ~65% of global sales run through independent distributors, and the two largest distributors are ~20% of total sales in each of FY23/24/25. That channel is the single most important structural fact about this company — it is both the moat (entrenched local relationships, OEM/machine-builder reach) and the principal source of the cyclical whip (the 2024 destocking, see Lens 8). End markets span discrete (auto/EV/battery, semiconductor, e-commerce/warehouse), hybrid (food & beverage, life sciences, tire) and process (energy, mining, chemicals). Addressable market self-estimated at ~$120B.

Plain-terms verdict: a high-quality industrial razor with a software upgrade in progress — recurring-ish (installed base, ARR) but fundamentally a capex-cycle business whose orders rise and fall with US manufacturing PMI (sub-50 for all of Q4 FY25 ).

Supply Chain

Upstream → Rockwell → end customer, named at every node:

• Inputs (upstream): semiconductors and electronic components (microcontrollers, power semis, passives), copper and steel (drives, enclosures, motors), and — flagged explicitly in the 10-KForm 10-KA company’s audited annual report to the US regulator. The most complete thing it publishes. risk factors — rare-earth minerals subject to "embargoes, sanctions, and other trade restrictions". Rockwell discloses it "maintains several single-source supplier relationships" where alternatives don't exist or the incumbent is advantaged on quality/price; it does not name them. Chokepoint: single-source components for "higher volume and more profitable products" — the 10-KForm 10-KA company’s audited annual report to the US regulator. The most complete thing it publishes. states unavailability there "would be more severe".
• Manufacturing (Rockwell): ~10 principal plants — most significant in the U.S., Mexico, Poland, India, Canada, Singapore — plus two principal distribution hubs (one U.S., one Netherlands). Resiliency strategy = redundant manufacturing so high-value product lines can be built in more than one geography (the explicit tariff-mitigation lever).
• Channel (the critical node): ~65% of sales via independent distributors; the two largest ≈ 20% of revenue. In North America the distributor share is even higher. This is a deliberate asset-light reach strategy but it means Rockwell sells to the channel, and channel inventory swings drive reported revenue independent of true end-demand.
• Customers (downstream): OEMs/machine builders, system integrators, and end users — concentrated in the U.S. (>50% of sales ). New demand pull from data-center operators (cooling, power, build-out automation), warehouse/e-commerce automation, semiconductor fabs, and EV/battery plants.

Single-source + rare-earth + China exposure (China is a top-5 destination country) = the supply-chain tail risk. Tariffs are managed via pricing + alternative sourcing + redundant plants, and management guides tariff-neutral to EPS in FY26 — a claim worth tracking, not trusting blindly.

Competitive Advantages (moats)

Rockwell's moat is real, narrow, and North-American.

1. Switching costs (the core moat). The Logix control architecture is the only platform "that can support many production disciplines — discrete, process, batch, safety, security, motion, robotics, and power control — in a single hardware and software environment". Once a plant standardizes on Allen-Bradley/Logix, the engineering, training, spare-parts and validation lock-in is enormous. This is why gross margins on Intelligent Devices and Software & Control sit at 18–35% segment-operating even through a down-cycle.
2. Installed base + distributor network (network effect). ~65%-distributor reach and a deep partner ecosystem mean Rockwell is the default in U.S. discrete automation. New machine builders design to Allen-Bradley because their North-American customers demand it — a self-reinforcing standard.
3. Brand / trust in critical infrastructure. Allen-Bradley, ControlLogix, PowerFlex, FactoryTalk are category-defining marks. In safety-critical and cyber-critical environments, "nobody got fired for buying Rockwell" is the U.S. analogue to IBM.

Bargaining power: Strong over customers in its installed base (switching cost), moderate over the channel (the channel is concentrated — losing a top-2 distributor would hurt), weak over single-source suppliers and rare-earth (admitted in risk factors).

The moat's boundary — be honest: it is geographic and discrete-centric. Outside North America, Siemens (TIA Portal / S7) is the entrenched standard and Rockwell is the challenger; in process industries, ABB, Emerson (DeltaV), Schneider/AVEVA, Honeywell are stronger. Rockwell is a fortress in U.S. discrete automation and a tenant elsewhere. Grounded in the 10-K's own competitor list: Siemens, ABB, Schneider, Emerson, Mitsubishi, Honeywell, Dassault.

Segments

All numbers (FY) and (fiscal Q2 FY26):

Revenue by segment ($M):

Segment operating margin:

Geography (FY25, by destination): North America $5,270M (+4%), EMEA $1,488M (−1%), Asia-Pacific $1,024M (−5%), Latin America $560M (−12%). By fiscal Q2 FY26 every region had flipped positive: NA +10%, EMEA +20%, APAC +13%, LatAm +9% reported.

The story the segments tell: FY24→FY25 was a trough — total revenue fell from $9.06B (FY23) to $8.26B (FY24) and only crept +1% to $8.34B in FY25, with volume −2pts offset by pricing +3pts. The recovery is led by Software & Control (+17% organic, 34.9% margin in Q2-FY26) — the secular software/ARR story finally showing in the P&L — while Lifecycle Services is the persistent laggard (organic still negative). Intelligent Devices is the cyclical swing factor turning up. Mix is improving toward the highest-margin segment, which is the entire margin-expansion bull case.

---

Phase B — Measure performance

Earnings Result (latest print: fiscal Q2 FY2026, period ended 2026-03-31)

The most recent printed quarter is the inflection, and it was a clean beat-and-raise.

• Sales $2,239M, +12% reported / +9% organic (currency +3pts, pricing +3pts). Broad-based: all four regions grew organically (NA +9%, EMEA +9%, APAC +10%; LatAm −1% the only soft spot).
• Diluted EPS $3.10 vs $2.22 (+40%); Adjusted EPS $3.30 vs $2.50 (+32%).
• Enterprise operating margin 22.5% vs 19.0% (+350bps) — driven by "price realization exceeding input costs, including productivity, higher sales volume, and favorable mix".
• Six-month FY26: sales $4,344M (+12%), Adjusted EPS $6.05 (+39%), enterprise margin 20.9%.
• vs consensus / guidance: the print "smashed expectations" and management raised FY26 guidance — see Lens 11.
• Cash: 6-mo operating cash flow $554M, free cash flow $445M (vs $464M PY — flat, as the cash recovery lags the P&L).
• Balance-sheet flags (FY25 year-end): Cash $468M; Receivables $1,931M (up from $1,802M — rising slightly faster than the +1% revenue, worth watching but modest); Inventories $1,247M (down — healthy post-destock); Goodwill $3,839M; total debt ≈ $3,224M (ST $608M + current LTD $2M + LT $2,614M); net debt ≈ $2.76B; commercial paper $522M @ 4.24%.
• Market reaction: the stock "boosted" on the beat and has compounded a ~58% rally off its 52-week low ($305) to ~$482 — i.e., a large amount of the recovery is already in the tape.

Unusual vs ROK's own history: the divergence between GAAP EPS $7.67 and Adjusted EPS $10.53 in FY25 ($2.86 gap) is the flag. FY25 GAAP was depressed by a $224M non-cash Sensia JV impairment, a $136M pre-tax legacy-asbestos accounting-method change ($0.91/sh), and restructuring — most of which is genuinely non-recurring, but the asbestos accounting change (from expense-as-incurred to full-future-accrual) is a one-time pull-forward that flatters future GAAP comparisons.

Earnings Calls (sentiment trend)

No transcripts on the shelf; reconstructed from filings + web. The arc over the last ~6 quarters is a clean capitulation → stabilization → confidence progression:

• FY24 (destock): repeated guidance cuts; Q2 FY24 EPS guide slashed from $11.24–12.74 to $8.80–9.80; CFO retirement announced. Tone: defensive, "managing through the channel correction."
• FY25 (trough): language pivots to cost productivity and margin expansion — the recurring phrase. The 10-K leans hard on "productivity," "price realization," and "margin expansion initiatives". Restructuring largely complete (FY24 charge $97M; FY25 reversed $5M).
• FY26 (recovery): tone is confident — "solid momentum… led by improving demand in warehouse automation, data center, semiconductor, and energy"; data-center revenue "doubling YoY," Logix ">20% growth globally". Phrases added: data center, AI, mega-projects. Phrase dropped: destocking / channel correction.

Sentiment read: management has earned some credibility back by delivering the margin recovery it promised at the trough — but the same team owns the FY24 air-pocket, so treat the new optimism as cyclically-aided, not vindication of a step-change.

Comps

Peer set = the 10-K's named competitors with public multiples.

Read: Rockwell is the most expensive name in the group on every comparable metric — ~35x forward earnings and ~24x NTM EV/EBITDA vs a peer-group EV/EBITDA average cited around 14x and the closest pure-ish peers (Emerson ~15.5x, Schneider ~16.9x) materially cheaper. ABB is the only peer in the same EV/EBITDA zip code (~22.6x). The premium is defensible on quality (pure-play, highest software content, best U.S. discrete moat) but it is a full premium with little margin for error — the market is paying up for the recovery it can already see.

Stock-Price Catalysts (>5% moves, last ~5 years)

The pattern is unusually legible: this stock trades on the order/destocking cycle and on guidance revisions, far more than on any single product. [all web unless noted]

• 2021–22: rode the post-COVID capexCapital expenditureMoney spent on long-lived things — buildings, machines, servers — rather than on running costs. + reshoring boom; total-return index $100→$135 (FY21) then back to ~$101 (FY22) as rates/multiples compressed.
• 2024 — the defining negative catalyst: distributor destocking drove FY24 revenue −8.8% to ~$7.5B and forced repeated guidance cuts; the Q2 FY24 EPS guide was slashed ~25% (to $8.80–9.80) and the CFO retired — multiple sharp down-moves. Q4 FY24 showed a 21% sales drop / 32% EPS decline and shares plunged.
• 2024–25 bottom: stock troughed near $305 (52-wk low).
• 2026 — the positive catalyst: the Q2-FY26 beat-and-raise plus the data-center/AI narrative drove a ~58% rally to ~$482, near the 52-wk high $486.

What the market actually reacts to: (1) order rates & channel inventory (the leading indicator), (2) guidance revisions (the trigger), (3) end-market narrative shifts (data center is the current one). It is not a single-customer story and not primarily a margin-surprise story — it is a cyclical-revisions stock. The practical implication for entry: the best risk/reward historically comes at the destock trough (FY24/early-FY25), not after the re-rate.

---

Phase C — Judge people & books

Management

Blake Moret — Chairman, President & CEO:

1. Track record / tenure: CEO since July 2016 (~10 years), Chairman as well — a long-tenured, lifer-style operator (Rockwell career man). Total comp ~$15.4M, ~92% variable (stock/options/bonus) — reasonable equity orientation.
2. Capital-allocation history — mixed: Post-tax-reform he repatriated ~$2.3B (2018), took a $1B equity stake in PTC (a software bet that has had value swings — a PTC investment gain flattered FY23, its absence pressured FY24 ), raised the dividend (now $5.24/sh FY25, 15 consecutive-ish years of increases ), and ran steady buybacks ($419M FY25 @ $279 avg, $594M FY24; $927M remaining authorization). The strategy has now shifted to favor capex + buybacks over M&A ($2B+ five-year internal investment plan). The Sensia JV (with SLB) was a value-destroyer — dissolved April 2026 after a $224M FY25 impairment; honest cleanup, but it was their deal.
3. ROE/ROIC on his watch: ROE ~24% in a trough year (FY25) — structurally high because the model is asset-light and buyback-shrunk (treasury stock $3.5B, shares ~112M). But the FY24 destock shows the earnings are cyclical, so the through-cycle return is lower than the spot figure implies.
4. Red flags: the 2024 guidance-cut sequence is the real scar — repeated misses + CFO departure inside one fiscal year is an execution/forecasting black mark. Independent analysts frame the open question precisely: can Moret restore Intelligent Devices margin to 22%+ and avoid another order "air-pocket". Routine insider selling (Moret sold ~$8M of stock ) — normal for comp of this structure, not a standout flag.
5. Archetype: professional manager / steward, not founder-operator. For a mature compounder that's appropriate; it implies steady execution and shareholder returns, not bold reinvention.

Net: competent, aligned, shareholder-friendly steward who delivered the promised margin recovery — but carries a genuine forecasting-credibility ding from 2024 and one clearly-failed JV.

Forensic Red Flags

Acting as a forensic analyst across the three statements:

• Non-GAAP vs GAAP gap: the $2.86 FY25 wedge between Adjusted EPS $10.53 and GAAP $7.67 is the headline item. Drivers are mostly legitimate non-cash/one-time: $224M Sensia impairment, $136M asbestos accounting-method change, restructuring, intangible amortization. But note the asbestos change is a one-time accounting election (expense-as-incurred → accrue-all-future) that pulls a charge forward and makes future GAAP look better by construction — track that the Adjusted bridge stays clean as the cycle normalizes.
• Cash vs earnings: FY25 FCF $1,358M vs net income $869M — FCF exceeds GAAP earnings (coverage >1.5x), the opposite of a red flag; it reflects the non-cash impairments and working-capital release post-destock. Healthy. (Caveat: FY24 FCF was only $639M, so the series is volatile.)
• Receivables/inventory vs revenue: Receivables +7.2% on +1% revenue (FY25) — a mild watch-item, plausibly timing/mix. Inventories down — good. Nothing screaming channel-stuffing, though the distributor model inherently obscures true sell-through (sell-in ≠ sell-out is the structural caveat for this whole name).
• SBC / segment defns: Adjusted EPS excludes acquisition-intangible amortization and (new in FY26) Sensia dissolution costs/tax — the FY26 definition change is disclosed and immaterial-to-prior-years, but definition creep in the adjusted bridge is the thing to monitor.
• Presentation change: FY25 reclassified to break out Engineering & development from Cost of sales (~$162M/qtr) — explicitly "no impact on net income, EPS, cash flows, or segment earnings"; aligns with peers. Benign, improves comparability.
• Tax: effective rate rising (FY24 13.8% → FY25 18.3%), and BEPS Pillar Two (Singapore) adds ~+3pts from FY26 — a real, structural EPS headwind already in the adjusted tax rate (Q2-FY26 adjusted rate 20.6%).
• Off-balance-sheet / contingencies: legacy asbestos (net liability ~$151M, scheduled payouts to "thereafter") and environmental remediation as a potentially-responsible party — long-tail but quantified and modest vs a $54B market capMarket capitalisationThe share price multiplied by the number of shares. What the market says the equity is worth..

Regulatory findings (required sub-section):

• SEC Litigation Releases: None. "No LR found for this company in the search period [2021-06-29 → 2026-06-29]".
• SEC AAERs: None. "No AAER found".
• Non-SEC (FTC/DOJ/FDA/etc.): web search surfaced no material enforcement actions, consent decrees, fines, or penalties against Rockwell Automation in the period. (The company sells to regulated end-markets but is not itself a notable enforcement target.)
• 10-K Item 3 (Legal Proceedings): Rockwell points Item 3 to Note 17 — ordinary-course litigation plus the legacy asbestos (divested-business product-liability) and environmental matters; management states none are expected to be material to financial condition.
• Conclusion: No material regulatory or legal findings — verified via SEC EDGAR EFTS (LR, AAER), web search, and 10-K Item 3 / Note 17, as of 2026-06-29. Clean book.

---

Phase D — Project & stress-test

Forward Projection (FY2026 → FY2028 Adjusted EPS)

Built bottom-up from the latest actuals + management's raised FY26 guidance. Base year anchored on guidance; out-years on revenue × margin × buyback.

Anchor (FY26 guidance, raised May 2026): sales ~$8.9B; Adjusted EPS $12.50–$13.10 (midpoint $12.80, ~+22% YoY); organic revenue 5–9% (midpoint 7%). Price realization ~250bps (100bps tariff pass-through + 150bps structural); S&C "several hundred bps" margin expansion.

Scenarios (FY27 Adj EPS):

• Bull ~$15.00: data-center scales from ~low-single-digit to ~5% of revenue, organic +10%, S&C margin >35%, full price-cost capture.
• Base ~$14.00: 7% organic, steady mix shift, Pillar-Two tax drag absorbed.
• Bear ~$11.50: a new destocking air-pocket (the FY24 playbook) — organic flat-to-down, ID margin stalls below 20%, operating de-leverage. This is the live tail, not a tail-of-the-tail — it happened two years ago.

Brier forecast (logged for calibration — base case committed):

• Question: ROK FY26 (ending 2026-09-30) non-GAAP Adjusted EPS ≥ $12.50 — p = 0.80 (already 6-mo Adj EPS $6.05 + raised guide floor $12.50; risk is only a sharp H2 reversal). Note: per --watchlist rules the our model create step is skipped in the unattended loop; recorded here for promotion if this goes to our position log.

Bull vs Bear

Bull case. Rockwell is the cleanest U.S.-listed pure-play on three durable secular tailwinds at once — reshoring (244k U.S. manufacturing jobs announced 2024, 88% in high-tech: EV/semi/electronics — Rockwell's exact end-markets ), data-center/AI build-out (data-center revenue doubling YoY, a market going from low-single-digit to ~5% of sales ), and the software/ARR margin mix-shift (Software & Control +17% organic at 34.9% margin ). The moat (Logix switching costs, 65% distributor reach) is intact; capital allocation is shareholder-friendly (dividend aristocrat-adjacent, steady buybacks); and the company just proved it can expand margin +350bps off the trough. FY24's destock cleared the channel, so the recovery has clean comps and operating leverage on the way up.

Bear case (2–3 permanent-impairment / de-rate risks).

1. It's a cyclical at a secular multiple. ~35x forward / ~24x EV/EBITDA prices ROK like a software compounder, but ~70% of revenue is still cyclical hardware/services tied to manufacturing PMI (sub-50 as recently as Q4 FY25). The single biggest risk is multiple compression — if the cycle wobbles, a 35x stock can lose a third of its value without any "impairment" at all (it did exactly this in 2024).
2. Channel fragility is structural, not solved. 65% distributor / top-2 = 20% means sell-in can decouple from sell-out at any time; the 2024 air-pocket was not a one-off bug, it's a feature of the model. Another destock resets the EPS path to the bear case (~$11.50).
3. The moat is geographically boxed. Siemens is winning the rest of the world and pushing software (TIA/Xcelerator); if Rockwell's premium depends on the data-center narrative and that decelerates, the "why pay 2x Emerson's EV/EBITDA" question gets ugly.

Pre-mortem (18 months out, thesis broke): US manufacturing PMI never sustained >50, the data-center automation TAM proved smaller/slower than the doubling-YoY headline implied (off a tiny base), distributors who over-ordered into the 2026 recovery began destocking again in late FY26, and a 35x multiple compressed to 25x on a single guide-down — a 30–35% drawdown with EPS only modestly lower. Plausibility: moderate-high — it is literally the 2024 movie replayed.

Are multiples too high? Yes, on an absolute and relative basis — ROK is the most expensive name in its peer group with the most cyclical revenue mix. The valuation is justifiable on quality but offers negative margin of safety at a 52-week high; consensus price targetConsensus price targetThe average of what published analysts think the share price should be. An opinion poll, not a forecast. ~$462 already sits below the ~$482 price.

Contrarian view (what the market refuses to see): the bulls are extrapolating a doubling off a low-single-digit data-center base into a structural growth re-rate, while under-weighting that this is the same distributor-driven cyclical that cut guidance four times in 2024. The market is paying a software multiple for a business whose order book it cannot actually see in real time. The asymmetry is wrong here — the time to own ROK is the next destock scare, not the post-beat melt-up.

Devil's Advocate (short-seller)

Dismantling the bull case:

• Where revenue is concentrated / what breaks it: ~65% through independent distributors, top-2 = ~20%. A single large distributor consolidating, defecting, or over/under-ordering moves reported revenue independent of demand. The "data center doubling" is off a low-single-digit base — management's own framing is ~5% of revenue near-term — so it cannot offset a discrete-hardware downturn; it's a narrative, not yet a needle-mover.
• Why the moat is weaker than bulls think: outside North America Rockwell is the challenger to Siemens, and in process industries it trails Emerson/ABB/Honeywell. The Logix lock-in is real in U.S. discrete but does not travel — and software-led entrants plus open-automation initiatives chip at the edges.
• Most dangerous competitor bulls underestimate: Siemens — bigger R&D budget, global installed base, and an aggressive software/digital-twin push (Xcelerator) that directly attacks Rockwell's highest-margin S&C growth engine.
• Worst capital-allocation moves: the Sensia JV (impaired $224M, dissolved 2026) and the $1B PTC stake whose mark-to-market swings have whipsawed reported earnings (gain flattered FY23, absence pressured FY24). Management has used the balance sheet for bets that added volatility, not durable value.
• Assumptions that must hold for today's price (~$482, ~35x fwd): sustained ≥7% organic growth, S&C margin holding ~33–35%, no destocking relapse, data-center scaling on schedule, and the Pillar-Two tax drag fully absorbed without an EPS miss. That's a lot of "and."
• If growth disappoints 20–30%: FY27 EPS to ~$11.50 (bear) and a multiple de-rate to ~25x → ~$290, roughly −40% — and note that's near the actual 52-week low of $305, so it's not a hypothetical floor, it's a recent one.
• Single scenario that permanently impairs the business: hard to construct — the franchise is durable; the realistic damage is valuation, not solvency. The honest short thesis is "great company, wrong price, wrong moment in the cycle," not "broken business."

Management Questions (ordered by information value)

1. Your reported revenue is ~65% distributor-routed with the top two at ~20% — what real-time sell-through visibility do you now have that you lacked in 2024, and how confident are you that the FY26 recovery isn't channel re-stocking that reverses?
2. Data-center automation is "doubling" but ~low-single-digit % of sales — at what revenue share does it actually move enterprise growth, and what's the realistic FY28 number?
3. Intelligent Devices margin is back to ~21% in Q2 — what gets it sustainably to the 22%+ target, and is that volume-dependent or structural?
4. Software & Control is at 34.9% margin and +17% organic — how much is durable ARR vs. one-time license/project, and what is the actual ARR base and growth rate in dollars?
5. You've shifted capital allocation to capex + buyback over M&A ($2B+ plan) — what ROIC hurdle are you underwriting, and why is organic the better use than acquisitions now?
6. Tariff-neutral to EPS in FY26 — what specifically offsets the cost, how much is price vs. redundant-manufacturing, and what's the FY27 exposure?
7. Pillar-Two adds ~3pts to the tax rate from FY26 — is ~20% adjusted the new structural rate, and are there mitigations?
8. Siemens is pushing Xcelerator hard globally — where are you winning and losing share in software outside North America, specifically?
9. The Sensia dissolution closed April 2026 — what's the pro-forma revenue/margin impact, and what did you learn about JV structures for future partnerships?
10. Your single-source and rare-earth dependencies are flagged in risk factors — which product lines are most exposed, and what's the dual-sourcing roadmap?
11. ARR is guided to high-single-digit growth — what's the path to double-digit, and which offerings (Plex, Fiix, Verve, FactoryTalk) drive it?
12. How are you embedding AI into the Logix/FactoryTalk stack in a way customers will pay incrementally for, not just expect for free?
13. With the stock at ~35x forward and near all-time highs, how does that change your buyback pace vs. dividend vs. debt paydown?
14. What leading indicators are you watching that would tell you a demand air-pocket is forming before it hits orders?
15. Five years out, what share of revenue is recurring software/services, and what does that do to the through-cycle margin floor?

---