Munters Group

Phase A — Understand the business

Lens 1 · Company Overview

What it actually is. Munters is a 93-year-old Swedish industrial air-treatment company (founded 1955 by Carl Munters; roots to the 1930s) that engineers climate control — humidity, temperature and air quality — for environments where getting the air wrong is expensive or ruins the product. It is not a pure "data-center company." It is a desiccant-dehumidification and evaporative/precision-cooling specialist that has ridden three demand waves in sequence — food production, lithium-battery dry rooms, and now AI data-center heat rejection — using largely the same core competence: moving heat and moisture out of tightly-controlled spaces energy-efficiently.

Three business areas (as of FY2025):

• Data Center Technologies (DCT) — energy-efficient cooling for data centers: evaporative/adiabatic air cooling, chillers (via the Geoclima acquisition), coolant distribution units (CDUs), computer-room air handlers (CRAHs), and increasingly direct-to-chip liquid cooling (LCX system, 500 kW–1.5 MW; SyCool heat rejection; ZutaCore two-phase partnership). ~SEK 6bn revenue in 2025, ~1,000 employees. This is the growth engine.
• AirTech — the historical core: energy-efficient air treatment for industrial (battery gigafactory dry rooms, pharma, food, electronics), commercial (supermarkets, public infrastructure), and Clean Technologies (VOC abatement / gas purification via the Airprotech acquisition), plus a Service and Components business. Global leader in desiccant rotor dehumidification (patented Green PowerPurge).
• FoodTech — control systems, sensors and software for agriculture/livestock (after divesting the FoodTech Equipment line). Munters announced in 2026 it is exploring a full divestment of FoodTech to concentrate the group on DCT + AirTech.

How it makes money. Project-based capital equipment (design → manufacture → install → commission), increasingly on large multi-year orders booked into backlog and delivered over 1–3 years, plus a growing aftermarket Service/Components annuity. The DCT order book is now dominated by US hyperscaler and colocation "AI factory" deployments delivered 2026–2028.

Contract structure / payment terms. Large DCT orders are fixed-scope, milestone-delivered capital contracts (e.g. the Apr-2026 ~SEK 2.0bn / $183m modular-cooling order books in Q2 2026, deliveries early-2027 through Q1-2028). This is not recurring SaaS revenue and not classic take-or-pay — it is backlog that must be manufactured and installed, exposing Munters to input-cost, tariff and execution risk between order and delivery. Aftermarket Service is the recurring layer but is a minority of revenue.

Bottom line: a mission-critical niche industrial with a genuine 90-year heritage moat in dehumidification, now enjoying an AI-cooling demand shock that is filling the order book faster than it is filling the P&L.

Lens 2 · Supply Chain

Munters sits in the middle of the data-center thermal chain — it integrates components into cooling systems, it does not make chips or build the shell.

Upstream (inputs into Munters):

• Raw materials — steel, aluminium, copper (coils, heat exchangers), refrigerants/dielectric fluids, controls electronics, compressors, fans, pumps. Commodity-priced; exposed to US tariffs on imported finished chillers (the ~-4pp DCT margin drag in Q4-25/Q1-26).
• Component/technology partners — ZutaCore (two-phase waterless direct-to-chip; Munters SyCool provides heat rejection, ZutaCore Hypercool removes heat from GPUs); Geoclima (now owned — Italian air/water-cooled chillers, brought in-house Oct 2024); Airprotech (owned — VOC abatement). Vertical integration via M&A is a deliberate supply strategy.
• Manufacturing footprint — global plants; critically, Munters is standing up US chiller production (Virginia) to localize supply and dodge tariffs; this "dual-site" ramp is the source of the underutilization cost currently depressing margin. US chiller production "up and running during Q2 2026… fully replace imports by end of Q2, ramp in Q3".

Munters (the transform): designs and assembles cooling platforms — evaporative units, chillers, CDUs, CRAHs, liquid-cooling loops — tailored per data-center architecture (air / liquid / hybrid).

Downstream (who buys):

• DCT: US hyperscalers and colocation operators building AI factories — the buyer names are largely undisclosed in public filings (a concentration + opacity flag; see Lens 13). The Nov-2025 ~$215m order was from "a US hyperscaler"; the Apr-2026 ~$183m from "a US colocation customer".
• AirTech: lithium battery cell manufacturers / gigafactories (dry-room dehumidification — a $30m US battery order in 2025), pharma, food, electronics, supermarkets.
• FoodTech: poultry/livestock producers (controls).

Chokepoints / single-source dependencies: (1) the Virginia chiller ramp is a self-inflicted chokepoint — until it runs, tariffed imports compress margin; (2) customer concentration in US AI capex — a handful of large orders drive DCT; (3) skilled installation/commissioning capacity — scaling delivery of a SEK 15bn backlog is an execution bottleneck UBS explicitly flagged. Names present throughout — this lens is grounded, not generic.

Lens 3 · Competitive Advantages (moats)

Durable moats — real but narrow:

1. Desiccant-dehumidification heritage + IP (strong). Munters pioneered desiccant rotor dehumidification and holds patented tech (Green PowerPurge, LDP dehumidifiers) that gives up to 30% energy savings in battery dry rooms; it is the acknowledged market leader in gigafactory dehumidification. This is a genuine ~70-year process/know-how moat in a mission-critical, low-tolerance-for-failure niche.
2. Energy-efficiency positioning (moderate). In both AirTech and DCT the pitch is kWh saved — evaporative/adiabatic cooling and free-cooling reduce data-center PUE and OPEX. As power becomes the binding constraint on data centers, "cool the same load with less power/water" is a structurally advantaged message.
3. Application engineering + service installed base (moderate). Decades of climate know-how across industries; a Service/Components aftermarket that creates switching cost once installed.
4. Vertical integration via M&A (emerging). Buying Geoclima (chillers) and Airprotech (VOC) brought critical hardware in-house, widening the DCT product envelope to air + liquid + hybrid.

Where the moat is thin:

• DCT is the least moaty of the three. Data-center cooling is a crowded, well-capitalized field — Vertiv (~23% precision-cooling share), Schneider Electric, STULZ, Johnson Controls, Rittal, Daikin, plus liquid-cooling specialists (Boyd, Asetek, CoolIT). Munters is a share-taker, not the incumbent, competing partly on evaporative/energy angle and custom AI-factory integration. Vertiv and Schneider have larger global service networks and deeper hyperscaler relationships.
• Bargaining power is customer-favorable in DCT. Hyperscalers/colos are large, sophisticated, concentrated buyers; a supplier scaling into their backlog has limited pricing power — visible in the fact that record orders came with worse margins.

Net: the moat is genuine in dehumidification/AirTech (heritage, IP, leadership) and thinner in DCT (the growth story) where Munters is a fast-following challenger to entrenched giants.

Lens 4 · Segments

segments.csv is an empty stub — all figures , not . Munters reports by business area; exact per-area splits for FY2025 were not cleanly extractable from the image-PDF, so the following blends disclosed figures with estimates (labeled).

By business area (FY2025):

Trend and cause. The mix is shifting decisively toward DCT: DCT order intake grew +416% YoY (+453% organic) in one quarter (SEK 9.2bn of orders), while AirTech was the FY2025 drag (industrial/battery softness, underutilization) and FoodTech is being carved out. Geographically, growth is overwhelmingly the Americas (US AI data-center capex); precise Americas/EMEA/APAC splits were not cleanly sourced — n/a at the exact-percentage level, but management repeatedly attributes DCT strength to US demand.

---

Phase B — Measure performance

Lens 5 · Earnings Result (latest print — Q1 2026, reported ~Apr 2026)

Headline: exceptional demand, weakening earnings — the defining tension of the whole thesis.

• Order intake: +32% reported / +49% organic (currency −~17pp). DCT order intake +138% organic (currency −31pp).
• Net sales: MSEK 3,580, −4% reported (+9% organic, −12% currency — a strong SEK crushed reported figures).
• Adjusted EBITA: MSEK 390, margin 10.9% — down from 13.5% a year earlier. Record orders, falling margin..
• Net income: MSEK 124 (down from 198). EPS SEK 0.68 (vs 1.05).
• Order backlog: +88% YoY, mostly DCT for 2026–2028 delivery; DCT backlog ~SEK 15bn, DCT book-to-bill 1.6 (group ~1.3).

Why margin fell (management's own bridge):

• DCT tariff headwind ~−4pp (imported finished chillers into the US) — same drag as Q4-25; direct finished-goods tariff "gone by this quarter" as US production replaces imports.
• Underutilized factories + dual-site costs during the Virginia ramp.
• Currency — the SEK strengthened sharply, turning +9% organic sales into −4% reported.
• AirTech improving (cost savings, price increases, no more dual-site cost) — the one bright margin spot.

Guidance / tone. Reaffirmed DCT +30–40% full-year sales growth, H2-weighted; DCT margin expected to improve in H2-2026 as tariffs ease and Virginia ramps; capex + tax rate broadly in line with 2025. Tone: confident on demand, explicitly cautious on near-term margin.

Balance-sheet flags. Leverage ~2.9x net debt/adj EBITDA (elevated, M&A-driven); OWC/net sales ~7.3% (Q4-25) improving from working-capital release. Big AI orders require working-capital funding ahead of delivery — a cash-flow watch item as the backlog converts.

Market reaction. Shares fell on the print — record orders were already priced; the miss was margin. This tells you what the tape cares about now: not order intake (saturated as a positive) but incremental margin and delivery execution.

Full-year 2025 context (the year just closed):

• Net sales +8% reported (~+15% ccy-adj), ~SEK 15bn.
• Order intake +85% (record); backlog +53%; book-to-bill 1.6x.
• Adjusted EBITA margin ~12.7% (down YoY).
• EPS SEK 3.01 (vs 4.96 prior) — earnings fell despite record demand.
• Q4-2025 was ugly: net sales −8%, EPS −0.06 (vs 0.85) — tariff + ramp + FX all hit at once.
• Dividend proposed SEK 1.60/share (MSEK 292, 53% of continuing-ops net income).

Lens 6 · Earnings Calls (sentiment trend)

transcripts/ is empty — sentiment reconstructed from `` call coverage across FY2025→Q1-2026.

Trajectory of tone:

• Q4-2024 (early 2025): first crack — "overall record performance… while investing for the future," but AirTech margin miss (9.4%) and a −10% stock reaction. Management framing: growth strong, profitability disappointing, investment phase.
• Q2-2025: "growth amid challenges" — demand ramping, margin/execution caveats.
• Q3-2025: most bullish — "order intake surges 57%… data center business thrives," stock surged.
• Q4-2025 / FY (early 2026): the pivot to "exceptional demand, while earnings weakened" — explicit acknowledgment that tariffs, underutilization, dual-site and FX are eating the margin even as orders explode.
• Q1-2026: "good momentum across all business areas" — reaffirmed DCT growth, promised H2 margin recovery, stressed Virginia ramp on track.

Recurring phrases (what they keep saying): "record order intake," "data center backlog," "book-to-bill," "energy-efficient cooling," "H2 margin improvement," "US chiller production ramp," "tariff headwind easing."

What they've started saying: "AI factory," "liquid cooling / direct-to-chip," "CDUs," "modular cooling platform," "exploring divestment of FoodTech" (portfolio focus).

What the shift tells you: the story has moved from demand-doubt (2024: can AirTech grow?) to execution-doubt (2026: can DCT be delivered profitably?). Management is credible on demand, on-the-hook on margin — the H2-2026 margin-recovery promise is the single most-watched commitment.

Lens 7 · Comps

Read. At ~30x forward P/E Munters is cheaper than Vertiv (~44x) — the pure AI-cooling comp — but more expensive than Schneider (~25x) and well above the European building/industrial average (~20x). The valuation says the market is paying a growth premium for DCT but a discount to Vertiv — logically, because Munters is (a) smaller, (b) currently margin-impaired, and (c) a challenger not the incumbent. The bull case is that as DCT margin normalizes toward group levels, the earnings base re-rates the multiple down and the price up; the bear case is that ~30x already prices the fix. A hard EV/EBITDA on Munters could not be cleanly sourced — flagged rather than fabricated.

Lens 8 · Stock-Price Catalysts (moves >5%, ~last 2 yrs — ``)

The tradeable history shows the market reacts to margin and order-size, in that order of surprise:

• Q4-2024 result (early 2025): −10% — AirTech margin miss (9.4%), EBITA 6% below consensus. Margin misses get punished hard.
• Q3-2025: sharp rally — order intake +57%, "data center thrives".
• Large DCT order announcements (2025–26): positive spikes — the ~$215m US hyperscaler order (Nov-2025, "record") and ~$183m / SEK 2.0bn US colo order (Apr-2026) both moved the stock and drew UBS's bull PT (SEK 245).
• Q1-2026 result: shares fell — record orders, weak margin; the market had already priced the backlog.
• CEO succession (2026): announcement that Klas Forsström will step down at Q3-2026, succeeded by Stefan Aspman — a governance/continuity event to watch.
• FoodTech divestment exploration (2026): portfolio-reshaping catalyst.

Pattern: early in the AI-cooling story, orders moved the stock (upside surprise). The market has now saturated on orders as a positive and repriced to react to margin delivery — a classic maturation of a momentum name. 52-week range 109.8–212.4 SEK shows how violent the re-rating has been.

---

Phase C — Judge people & books

Lens 9 · Management

CEO — Klas Forsström (President & CEO since Aug 2019, ~6.8-yr tenure; stepping down at Q3-2026):

1. Track record: grew net sales from ~SEK 7bn to >SEK 15bn over his tenure — roughly doubled the company — building the DCT engine, an M&A cadence, and an "innovation engine." Prior: 20+ yrs senior roles at Sandvik (President of Sandvik Hard Materials, Sandvik Coromant). Uppsala MBA. A credible, growth-oriented industrial operator.
2. Skin in the game: owns ~0.11% (~SEK 37m) — meaningful in kronor, small in %; total comp ~SEK 26m/yr, 66% variable/bonus (aligned to performance, but a high variable share).
3. Capital allocation: acquisitive — Geoclima (chillers, DCT), Airprotech (VOC, AirTech), Hotraco + Automated Environments (FoodTech controls); ~12 acquisitions tracked; leverage rose to ~2.3–2.9x funding M&A. Now pruning (FoodTech Equipment divested; whole FoodTech under review) — a portfolio-focus discipline that reads well. Dividend policy 30–50% payout (SEK 1.60 for 2025). ROE/ROIC on his watch: n/a — not cleanly sourced (flag).
4. Red flags: high variable comp; acquisition-financed leverage into a demand peak; the near-term margin deterioration happened on his watch (though largely tariff/FX/ramp-driven). No related-party or promotional-behavior flags surfaced.
5. Archetype: professional manager (career industrial executive, not founder). Appropriate for a scaling, M&A-driven mid-cap — but the CEO transition at the exact moment the DCT margin thesis must be proven is a continuity risk.

Successor — Stefan Aspman (incoming CEO): internally appointed; track record at Munters scale not yet public-facing — a "prove-it" unknown.

Ownership backdrop — a quality tell. Largest shareholder FAM AB (~28%), the Wallenberg-linked industrial holding company. FAM board rep since 2017 (Kristian Sildeby). Institutional ownership ~80%. A patient, blue-chip Wallenberg anchor is a governance positive (long-term orientation, low raider risk) — the same stable-owner profile as other Investor/Wallenberg-sphere Swedish industrials.

Lens 10 · Forensic Red Flags

Accounting-risk map (web-only; no filings on disk — every figure /, and the absence of primary statements is itself a limitation flagged here):

1. Revenue recognition on long-dated project backlog (highest-attention item). A SEK ~15bn DCT backlog delivered 2026–2028 means percentage-of-completion / milestone revenue judgment matters enormously. Watch for: revenue recognized ahead of cash, margin-in-backlog assumptions, and any change in POC estimates. Not yet a finding — but the single account to scrutinize once IFRS statements are read. ``
2. Working capital / cash vs earnings. Big AI orders demand WC funding ahead of delivery; OWC/net sales ~7.3% (Q4-25) with management targeting 10–13% — a release flattered recent operating cash flow, which can reverse as the backlog builds inventory/receivables. Cash flow from investing −MSEK 2,031 in FY2025 (acquisition-heavy). Watch FCF divergence from adj EBITA.
3. Adjusted vs reported EBITA. Munters headlines "adjusted" EBITA — the gap to reported (restructuring, dual-site, integration, M&A costs) is where optics can hide. With dual-site/tariff "adjustments" currently large, the adjusted-to-reported bridge is a must-read. ``
4. Goodwill / intangibles from M&A. ~12 acquisitions (Geoclima, Airprotech, controls businesses) build goodwill; a demand-peak acquisition binge risks future impairment if AI-cooling growth normalizes. ``
5. FX translation noise. A strengthening SEK turned +9% organic into −4% reported (Q1-26) — not fraud, but it materially distorts headline trends; always read organic + constant-currency.

Regulatory findings:

• SEC/EDGAR: No CIK — Munters is not an SEC filer; no EDGAR enforcement search possible. total_sec_findings: 0 — because none is searchable, not because none exists.
• Non-SEC web search ("Munters Group" (FTC OR DOJ OR FDA OR CFPB OR consent decree OR settlement OR fine OR penalty) enforcement): no material enforcement actions, consent decrees, fines or penalties surfaced in web results as of 2026-07-06.
• Short-seller / market-integrity signal: Wellington Management International disclosed a public short of ~0.55% — the only public short position. Small, but a named institutional skeptic worth noting (thesis presumably: margin dilution / valuation, not fraud).
• Item 3 / Legal Proceedings equivalent: no 10-K on disk (foreign filer); Swedish annual-report legal-proceedings disclosure not cleanly extracted from the image-PDF — n/a from primary filing. No litigation surfaced via web.
• Conclusion: No material regulatory or legal findings surfaced — verified via SEC EDGAR EFTS (no CIK, N/A), non-SEC web search, and web coverage as of 2026-07-06. Note: the absence of accessible primary IFRS statements is a real limitation of this web-only pass; the backlog revenue-recognition and adjusted-EBITA bridge should be verified against the audited FY2025 report before conviction.

---

Phase D — Project & stress-test

Lens 11 · Forward Projection (EPS, next three fiscal years — ``, arithmetic shown)

Anchor (actuals/guidance, ``): FY2025 EPS SEK 3.01 (down from 4.96); group net sales ~SEK 15bn; adj EBITA margin ~12.7%; DCT ~SEK 6bn guided +30–40%; margin recovery guided H2-2026; ~178m shares.

Drivers: DCT the growth engine (+30–40% then decelerating); AirTech low-single-digit recovery; FoodTech divested (removes ~SEK 2bn revenue but is margin-neutral-to-accretive to focus); the swing factor is DCT incremental margin — does it climb from tariff-impaired ~single digits back toward the mid-teens group level as Virginia ramps and tariffs roll off.

Arithmetic (base FY2027): DCT ~SEK 6bn ×1.35 ×1.25 ≈ SEK 10.1bn; +AirTech ~SEK 6.5bn ×1.03² ≈ 6.9bn ≈ group ~SEK 17bn (FoodTech excluded); adj EBITA ~13.5% ≈ SEK 2.3bn; less D&A/interest/tax at ~historic ratios → net income ~SEK 0.8–0.85bn ÷ ~180m sh ≈ ~SEK 4.6 EPS. At ~181 SEK that is a forward ~40x FY26 / ~28x FY27 / ~23x FY28 P/E path on the base case.

Tracked Brier forecast — NOT logged. Per --watchlist rules, the forecast.ts create step is skipped in the breadth loop; the base call (MTRS FY2026 EPS ≥ SEK 3.3 / DCT +30–40% sales growth achieved) is recorded here for /thesis to log if it promotes this to a call.

Lens 12 · Bull vs Bear

Bull case. Munters is a mission-critical picks-and-shovels play on the AI-power/cooling constraint with a heritage moat and a SEK ~15bn contracted DCT backlog stretching to 2028 — visibility most industrials would kill for. The margin drag is transitory and diagnosable: tariffs (rolling off as US chiller production replaces imports by Q3-2026), dual-site underutilization (ends as Virginia ramps), and FX (mechanical). As DCT incremental margin normalizes toward the group's mid-teens, earnings inflect sharply in H2-2026 and 2027, re-rating a stock that already trades at a discount to Vertiv (~30x vs ~44x). Optionality: liquid-cooling/direct-to-chip (LCX, SyCool, ZutaCore) rides the highest-growth thermal segment; a FoodTech divestment sharpens focus and could fund deleveraging or buybacks; a Wallenberg (FAM 28%) anchor underwrites patient capital. Contrarian bull: the market is treating a margin-timing problem as a quality problem.

Bear case (permanent-impairment risks).

1. DCT margin never reaches group level. If hyperscaler/colo buyers keep pricing power, Munters is a capacity-constrained challenger fulfilling low-margin backlog for entrenched giants — record revenue, mediocre returns. The "H2-2026 recovery" is a promise, not a print, and has already slipped once.
2. Order-book is a demand-peak artifact. AI data-center capex is famously lumpy and concentration-heavy; a pause, cancellation, or push-out in US AI buildout (or a single large customer) hollows the backlog. Backlog is orders, not cash — subject to renegotiation.
3. Execution/ramp failure. UBS flagged Virginia-site ramp hiccups and delivery-scaling as key risks; a mid-cap scaling a SEK 15bn backlog with a new CEO is an execution tightrope.

Pre-mortem (18 months out, thesis broke). It's early 2028. AI-cooling order intake decelerated in 2027 as hyperscalers digested capacity; the Virginia ramp ran late and over-cost; DCT margin plateaued in low double-digits because colos squeezed price; FX stayed a headwind; the new CEO's first year was consumed by the FoodTech carve-out. Consensus FY2027 EPS was cut ~20%, the ~30x multiple compressed to ~20x, and the stock round-tripped toward the low-100s (52-wk low was 109.8). Nobody was defrauded — the market simply paid a growth multiple for a margin-timing bet that didn't land on schedule.

Are multiples too high? ~30x forward is defensible only if the H2-2026→2027 margin inflection is real. On current (impaired) margins it is expensive. The valuation is a call on execution, not on demand.

Contrarian view (what the market refuses to see): the real durable business here may be AirTech dehumidification (heritage moat, gigafactory leadership, recovering margin) — the boring segment the AI narrative ignores — while DCT is the exciting-but-commoditizing part the multiple is built on. If DCT margin disappoints and AirTech quietly re-rates, the stock's composition of value is the opposite of how it's priced.

Lens 13 · Devil's Advocate (short-seller)

Dismantling the bull case:

• Where revenue is concentrated: DCT growth is US-centric AI capex funneled through a handful of large, undisclosed hyperscaler/colo orders. The public filings name buyers only as "a US hyperscaler" / "a US colocation customer" — opacity that hides concentration. If two or three end-customers drive most of the SEK 15bn backlog, a single deferral is a profit warning.
• Why the moat is weaker than bulls think: in DCT — the growth story and the multiple — Munters is a share-taker against Vertiv (~23% share, deeper hyperscaler ties, larger service network), Schneider, STULZ, JCI, Daikin, plus liquid-cooling specialists. Its differentiator (evaporative/energy efficiency, custom AI-factory integration) is replicable by better-capitalized incumbents. Pricing power is customer-favorable — proven by the fact that record orders came with lower margins.
• Most dangerous underestimated competitor: Vertiv — if it aggressively bundles liquid + air + service for hyperscalers, Munters' custom-integration edge narrows and it's left with the price-taker slices.
• Worst capital-allocation risk: acquisition-financed leverage (~2.9x) into a demand peak — Geoclima/Airprotech/controls goodwill is impairment-exposed if AI-cooling growth mean-reverts; buying at the top is the classic late-cycle industrial error.
• Assumptions that must hold for ~181 SEK: (1) DCT +30–40% and margin recovers to mid-teens by 2027; (2) no AI-capex air-pocket; (3) Virginia ramps on time/on cost; (4) CEO transition is seamless; (5) FX stops being a ~12pp drag. Five things must go right.
• If growth disappoints 20–30%: base FY2027 EPS ~SEK 4.6 → ~SEK 3.3–3.7; at a de-rated ~20x that's ~SEK 66–74 — roughly a 60% drawdown toward/below the 52-wk low. The downside is not academic; the stock has already ranged 110→212 in a year.
• Single scenario that permanently impairs: a structural, not cyclical, cap on DCT margins — if being a challenger in a giant-dominated field means Munters never earns group-level margins on data-center hardware, then the entire DCT re-rating thesis is void and the company reverts to a ~SEK 6–7bn AirTech industrial with a low-margin cooling appendage. Plausibility: moderate — this is the crux short.

Named skeptic already on the tape: Wellington (0.55% public short).

Lens 14 · Management Questions (ordered by information value)

1. DCT incremental margin: on the SEK ~15bn backlog, what is the gross/EBITA margin embedded in backlog vs the current group margin — and specifically what margin are the largest AI orders booked at?
2. Margin bridge to H2-2026: quantify the H2-2026 DCT margin recovery — how many points come from tariffs rolling off vs Virginia utilization vs pricing, and what is the exit-2026 DCT margin run-rate?
3. Customer concentration: what share of the DCT backlog is the top 1 / top 3 customers, and how much is hyperscaler vs colocation?
4. Cancellation/renegotiation terms: are the large AI orders firm/non-cancellable, and what are the penalty/repricing provisions if a customer defers?
5. Virginia ramp: exact timeline and capex/opex to full US chiller production — what's the cost of the ramp still to flow through, and the risk it slips again?
6. Through-cycle DCT margin target: where can DCT margins settle structurally vs Vertiv/Schneider — is group-level margin realistic, or is DCT a permanently lower-margin mix?
7. FoodTech divestment: timeline, expected proceeds/multiple, and use of proceeds — deleverage, buyback, or DCT/AirTech M&A?
8. CEO transition: what are Stefan Aspman's first-100-day priorities, and how is continuity of the DCT margin plan assured through the handover?
9. Working-capital funding: how much WC is required to convert the backlog, and what's the FCF profile through 2028?
10. Competitive positioning in liquid cooling: how defensible are LCX/SyCool/ZutaCore vs Vertiv and the direct-to-chip specialists — and what's the win-rate on contested hyperscaler bids?
11. AirTech recovery: is the Q1-2026 AirTech margin improvement durable, and what's the normalized AirTech through-cycle margin?
12. Leverage: comfort level on ~2.9x — any covenant proximity, and the deleveraging path?
13. Order-intake sustainability: how much of 2025-26 record intake is pull-forward vs a durable multi-year demand plateau?
14. Goodwill: any impairment-testing sensitivity on Geoclima/Airprotech goodwill if DCT growth normalizes?
15. Capital-allocation discipline: what IRR/return hurdle governs future M&A given you bought several assets near a demand peak?

---