
Vertiv Holdings (paired short overlay)
Designs, manufactures, installs and services critical digital-infrastructure hardware (thermal/liquid cooling, power, racks/integrated systems) sold to hyperscalers, colocation operators and enterprise data centers; ~one-third recurring services/aftermarket attached to a large installed base. Capital-equipment + services hybrid.
The thesis on this name
State of AI Compute
Cleanest public beneficiary of the thermal wall (co-equal bottleneck to packaging) and the electrical-gear+cooling toll — >$15B backlog (12-18mo coverage), ~$2.2B FCF, liquid-cooling premium capture — BUT the gear-toll margin-EXPANSION is a cycle-peak extrapolation, the VRT 'accumulate on <1.5x book-to-bill' entry is incoherent (entry trigger = de-rate trigger), and the 2-3yr moat half-life is shorter than the ~45-50x multiple implies; own on weakness, size for hyperscaler in-sourcing + microfluidic disintermediation.
State of Data-Center Power
The cleanest public pure-play on the thermal wall — the co-equal bottleneck to advanced packaging — and on the integrated electrical-gear + liquid-cooling rack toll. Backlog $15.0B (+109% YoY, fact), Q4-2025 book-to-bill ~2.9x, FY2026 guide raised to $13.5-14.0B rev and $6.30-6.40 adj EPS; adj operating margin expanded 430bp to 20.8% in Q1 2026 (fact). Captures a liquid-cooling premium as direct-to-chip displaces air. The catch is the price: >50x fwd P/E and ~$120-128B cap leave no margin of safety — this is the highest-quality but most-priced of the trio, so it is a compounder you scale on drawdowns, not a fat-pitch entry.
State of Data-Center Power
The liquid-cooling pure-play as racks go to 100kW+ — ~$15B backlog, ~2.9x book-to-bill, adj op margin 20.8% heading to 27% by 2030.
State of Data-Center Power
Tactical valuation hedge: a small short/put overlay on the most-expensive thermal name (P/E ~84) against the long electrical book, to cap AI-capex-unwind drawdown. Kept as the book's tail hedge: by the 5 Nov 2026 close Vertiv was $176.73, ~18x FY2027 consensus EPS and below every named power and cooling peer, so it is no longer the most expensive name; it stays until a broader instrument replaces it.
State of Data-Center Power
~$15B backlog, ~2.9x book-to-bill, margin 20.8%->27% path; sized below conviction purely on an 84x multiple — add on weakness.
State of Data-Center Power
Calibrated overlay on the highest-multiple name to offset ~1/3 of book beta in an AI-capex unwind; insurance, negative carry. Vertiv is no longer the highest-multiple name (~18x FY2027 EPS at the 5 Nov 2026 close); the overlay stays until a broader instrument replaces it.
Earnings, margins, COGS & capex
Vertiv is in a hyper-growth, margin-inflection phase driven by AI data-center buildout: Q3 FY26 sales +61% to $4.31B (+55% organic) with adjusted operating margin up 290bps to 25.2%, and management cited organic orders growth above 100% YoY; no dollar backlog or book-to-bill figure is given in the Q3 release or 10-Q (fact). FY26 guidance is $14.45-14.85B sales (37-39% organic) at 23.5-24.5% adjusted operating margin and $2.6-2.8B adjusted FCF (fact). The May 2026 investor day set a 2030 frame of 20-22% organic revenue CAGR and adjusted operating margin >27% (fact). The model is operating-leverage-led — incremental volume on a fixed manufacturing base plus higher-margin liquid-cooling mix drives the margin walk — and Vertiv is extending upstream: the UtilityInnovation Group acquisition (about $1.45B cash, closed late Oct 2026) adds microgrid controls and behind-the-meter power design for power-constrained data centers (fact).
Income statement — where each revenue dollar goes
% of revenueOf every $1 of revenue, ~62¢ is cost of goods and ~13¢ operating expense, leaving ~25¢ of operating profit (~18¢ net).
Revenue trend
Margins
up (only ~30bps YoY in Q3; 37.9% for nine months vs 35.3%)
up (+290bps YoY; 22.6% in Q2, 20.8% in Q1)
up
up
mixed (28.3% in Q2; Q3 absorbed a $766.9M inventory build)
COGS structure
COGS is dominated by direct materials/components — copper, steel/aluminum, semiconductors and power electronics (IGBTs, capacitors), compressors, heat exchangers/CDUs, and purchased subassemblies — plus factory labor and freight. Liquid-cooling and integrated-rack mix carries richer margins than legacy air-cooling/UPS, so mix shift plus volume leverage on fixed plant is the main margin lever; tariffs and copper/steel inflation are the main COGS headwinds.
Capex
Capex is stepping up (~4.0% of FY26 sales per the company; 4.2% for nine months), funding capacity expansion for liquid cooling/CDUs (CoolChip/CoolPhase), integrated-rack/OneCore assembly, and regional manufacturing footprint (Americas-led). Growth is funded from operating cash flow ($2.66B in nine months), with ~$0.8B of net cash at Sep 30 2026 before the ~$1.45B UtilityInnovation Group payment; the heavier cash use is M&A (ThermoKey, Strategic Thermal Labs, Great Lakes/BiosCDU-type tuck-ins, UtilityInnovation Group) rather than greenfield capex.
Latest earnings
Beat against guidance: net sales $4,313M vs $3,650-3,850M guided (about 12% above the top of the range, computed); adjusted operating margin 25.2% vs 24.0-25.0% guided; adjusted diluted EPS $2.21 (+78% YoY) vs $1.77-1.83 guided; diluted EPS $1.95 (fact). Management raised full-year guidance across all key metrics. Consensus figures were not checked.
FY26: net sales $14.45-14.85B (+37-39% organic), adj op margin 23.5-24.5%, adj diluted EPS $7.10-7.20 (+69-71%), adj FCF $2.6-2.8B. Q4 FY26: net sales $4.31-4.51B (+49-51% organic), adj op margin 25.3-26.3%, adj EPS $2.22-2.28. 2030 frame: 20-22% organic CAGR, adj op margin >27% (May 2026 investor day).
- Backlog
- $15.0B at Dec 31 2025 (more than doubled YoY); no dollar backlog given in the Q3 FY26 release or 10-Q
- Orders
- Organic orders more than 100% above a year earlier in Q3 FY26 (company statement); ~2.9x book-to-bill and +252% organic orders in Q4 2025. Management said on the Q3 call that it does not disclose orders and does not plan to share orders data going forward (third-party transcript)
- Net cash / liquidity
- ~$0.8B net cash and $6.2B liquidity at Sep 30 2026 (computed / company release); roughly $0.7B net debt pro forma for the ~$1.45B UtilityInnovation Group payment in Oct 2026, before Q4 cash flow (computed)
- Adj op margin
- 25.2% Q3 FY26 → 23.5-24.5% FY26 guide (25.3-26.3% guided for Q4) → >27% by 2030
Growth drivers
- AI rack power density going to 100kW+ (GB200/GB300/Vera Rubin) making direct-to-chip liquid cooling mandatory — Vertiv's highest-margin, fastest-growing line
- Record >$15B backlog (more than doubled YoY) and ~2.9x book-to-bill at Dec 31 2025 (no dollar backlog or book-to-bill figure disclosed since; organic orders grew more than 100% YoY in Q3 FY26, company statement) giving revenue visibility into 2027+
- OneCore integrated power+thermal+IT platform converting point-product sales into larger, stickier full-data-center deals (Nvidia co-development)
- Hyperscaler + neocloud capex super-cycle and a structurally undersupplied grid-power/cooling bottleneck
- Operating leverage + mix shift driving the walk from 25.2% (Q3 FY26) to >27% adjusted operating margin by 2030
- Aftermarket/services attach on a growing installed base (recurring revenue)
Reported financials — SEC EDGAR
Audited GAAP figures pulled from SEC filings · latest filing 2026-02-13. The audited primary-source spine — not financial advice.
Revenue — annual (GAAP)
Margins & balance sheet — FY’25
Bull & bear
Vertiv still owns the AI data center's two hardest physical bottlenecks — power and heat — and Q3 FY26 answered the execution question raised by Q2's organic shortfall: net sales of $4,313M (+55% organic) landed 12% above the top of the guided $3,650-3,850M range, adjusted operating margin reached 25.2% (+290bps), organic orders more than doubled (company statement), every key line of FY26 guidance was raised again, and the company moved upstream to the grid interconnect with the UtilityInnovation Group acquisition.
- Guidance raised again with the Q3 print (fact, 28 Oct 2026): FY26 sales to $14.45-14.85B from $13.8-14.2B, organic to +37-39% from +30-32%, adjusted operating margin to 23.5-24.5%, adjusted EPS to $7.10-7.20 from $6.65-6.75 and adjusted FCF to $2.6-2.8B from $2.4-2.6B — after Q3 net sales of $4,313M came in 12% above the top of the guided range
- The margin walk keeps stepping up: 20.8% (Q1) to 22.6% (Q2) to 25.2% (Q3, above the 24.0-25.0% guided), with Q4 guided to 25.3-26.3% and FY26 framed at a 24.0% midpoint, ~360bps above FY2025's 20.4%. The Q3 gain was operating leverage rather than gross margin: gross margin computes to 38.1%, only ~30bps above Q3 FY25, while SG&A fell to 12.9% of sales from 15.5%
- Cash generation stayed strong: adjusted FCF $652M in Q3 (+50%) and $2,230M for nine months (computed), the FY26 guide was raised to $2.6-2.8B, and Vertiv ended Q3 in a net cash position with $6.2B of liquidity — self-funding the ~4%-of-sales capacity build and the ~$1.45B UtilityInnovation Group purchase, which management said it would pay from existing resources
- EMEA, the regional crack flagged in June, has closed: organic growth went from -29% in Q1 FY26 to -2.4% in Q2 to +19.3% in Q3 (+22.8% reported), and growth is broad-based — Americas +66.2% and APAC +48.3% organic in Q3
- Deferred revenue has risen to $4,120.5M from $1,814.7M at year-end 2025 and inventories to $3,288.4M from $1,456.5M — consistent with staging for guided Q4 net sales of $4,313-4,513M rather than absorbing a demand pause, though inventories alone absorbed $766.9M of Q3 cash
- The valuation objection is materially weaker than when we wrote it: ~33.0x the FY26 guide midpoint and ~24.2x FY27 consensus EPS (computed; consensus is an estimate, aggregator), about the median of the named power and cooling peers (23.8x), against a company guiding +37-39% organic growth for FY26 and a 2030 frame of 20-22% organic CAGR to ~$26B at ~27%+ adjusted operating margin. The average analyst target of $339.07 (estimate, aggregator) sits ~43.7% above the Nov 3 close
The Q2 execution crack closed in Q3 (+55% organic against a +34-36% guide), so the case against Vertiv now rests on what the strong print costs and what it does not show. Inventories have more than doubled ($3.29B, +126%) and receivables are up 41% ($4.38B) since year-end 2025; the ~$1.45B UtilityInnovation Group purchase turns a net cash balance sheet into roughly $0.7B of net debt (computed, before Q4 cash flow) with up to $1.15B more payable on earn-out targets; backlog and book-to-bill are still not given in dollars; and the Q4 guide implies organic growth easing to +49-51% with sales only ~2% above Q3.
- The Q2 miss against management's own +20-24% organic guide (+17.8%, blamed on 'temporary supply chain congestion and multi-phased project execution') was recovered in Q3 at +55.0% organic against a +34-36% guide. That sets a high bar: Q4 is guided to net sales of $4,313-4,513M and +49-51% organic growth, about 2% above Q3 at the midpoint, so any further project-phasing slip would show immediately
- Backlog, book-to-bill and orders are still not given in dollars, and management said on the Q3 call that it does not disclose orders and does not plan to share orders data going forward. The $15.0B / ~2.9x / +252% datapoints are 31 Dec 2025 vintage; the Q3 release and 10-Q add only a growth rate (organic orders more than 100% above a year earlier, company statement), which is a direction, not a level. That leaves the strongest bull point resting on one percentage and qualitative pipeline commentary
- No premium to peers remains, but the stock still prices a hardware manufacturer on peak-cycle growth: ~33.0x the FY26 guide midpoint and ~24.2x FY27 consensus EPS (computed), against ~18.9x for Schneider Electric (estimates, aggregator). The de-rate mechanism fired precisely on cue — a ~3% revenue shortfall versus consensus, paired with a raise, still cost -17.3% in a single session and -25.8% from its June level; at the Nov 3 close the stock is still -27.5% from the Jun 25 mark despite the Q3 beat (computed)
- Capex intensity roughly doubled: ~4.0% of FY26 sales (company statement 28 Oct 2026), versus 2.15% actual in FY2025, with 9M FY26 at 4.2% ($425.0M). The asset-light framing does not hold, and inventories of $3.29B (+126% since year-end) plus receivables of $4.38B (+41%) are working capital at risk if the Q4 ramp slips
- Growth quality is diluting at the margin: ~6% of the 61.2% reported Q3 growth came from acquisitions, contingent-consideration charges ran $34.1M in Q3 and $96.1M for nine months, and GAAP operating margin of 22.8% sits 2.4 points below adjusted 25.2%. The UtilityInnovation Group deal adds up to $1.15B of further EBITDA-linked consideration to that line
- Concentration and lumpiness are unchanged and structurally unfixable — order flow rides a handful of hyperscalers and neoclouds, and Q2 showed that project phasing alone, with no cancellation at all, is enough to break a quarter
What it is worth
Forward P/E on FY2027 consensus adjusted EPS ($9.74; estimate, aggregator, Nov 2 2026), cross-checked vs power/electrical peers and the FY26 adj EPS guide of $7.10-7.20 and the 2030 frame (>27% adj op margin, 20-22% organic CAGR). Each case is that EPS times a named multiple: bull is the multiple the average analyst target implies, base is today's multiple (up to the named-peer median), bear is the cheapest named power or cooling peer. Peer multiples are Nov 3 2026 closes over FY2027 consensus EPS (Johnson Controls: fiscal year to Sep 2027), estimates from an aggregator.
~$184 (-22.0% from the close)
Schneider Electric's 18.9x, the cheapest named power and cooling peer, on FY27 consensus EPS of $9.74 (estimates, aggregator); consensus is held and the multiple carries the downturn — the AI-capex digestion / order-deceleration scare that is the paired-short payoff. The June bear of ~$180-220 assumed a compression toward a power-peer ~28-30x; the named peers trade at ~19-24x FY27 consensus EPS at the Nov 3 close, so the same price range needs a lower multiple on a higher EPS base. The $184 is 24% above the 52-week low of $147.82 (Dec 17 2025).
~$236 (0.0% from the close)
today's multiple, 24.2x FY27 consensus EPS of $9.74 (estimate, aggregator), with the FY26 guide delivered; the 23.8x median of the named peers is below it, so there is no step up. The June base of ~$330-360 ('multiple holds ~mid-40s on FY27 EPS') did not hold: that multiple no longer exists, and $330-360 is 34-37x today's FY27 estimate, above every named peer (the highest, Ecolab, is at ~32.4x).
~$339 (+43.7% from the $235.97 close): the multiple the average analyst target implies, 34.8x FY27 consensus EPS of $9.74 (target $339.07 across 30 analysts; estimates, aggregator) — sustained >27% margin walk + 20%+ growth holding a premium multiple. The June bull of ~$430-450 equals 44-46x today's FY27 estimate and sits above even the highest analyst target ($427, estimate, aggregator).
At $235.97 (fact, Nov 3 2026 close) the stock trades at ~33.0x the FY26 adj EPS guide midpoint of $7.15 and ~24.2x FY27 consensus EPS of $9.74 (computed; estimate, aggregator). Named peers on FY27 consensus EPS at the same close: Schneider Electric ~18.9x, Trane ~23.6x, Johnson Controls ~23.8x, Eaton ~24.0x, Ecolab ~32.4x; median ~23.8x (computed; estimates, aggregator). Super Micro trades at ~6.4x (fiscal year to Jun 2027) but assembles servers rather than power and cooling equipment, so it is not used. The ~46-48x forward multiple near a five-year high published in June no longer exists: the premium to peers is gone, and the June cases built on it (bull ~$430-450, base ~$330-360, bear ~$180-220 — about 44-46x, 34-37x and 18-23x today's FY27 estimate, computed) did not hold. Thesis is growth-durability; risk is a multiple de-rate on any AI-capex air-pocket, and the paired-short overlay no longer rests on a stretched multiple.
SWOT
Strengths
- Scaled, certified, full-stack power+thermal vendor — one of few that can bid hyperscaler liquid-cooling projects today; #1/leading liquid-cooling share
- Record >$15B backlog and ~2.9x book-to-bill at Dec 31 2025 (no dollar figure disclosed since; Q3 FY26 organic orders more than doubled, company statement) give strong revenue visibility
- Fast margin inflection — adj op margin 25.2% (Q3 FY26, +290bps YoY) with a credible >27% 2030 path on volume leverage + mix
- Strong balance sheet — ~$0.8B net cash and $6.2B liquidity at 30 Sep 2026 (before the ~$1.45B UtilityInnovation Group payment) — funds M&A and capacity from cash flow
- Deep Nvidia co-development (OneCore for GB200/GB300/Vera Rubin) embeds Vertiv in the reference architecture
Weaknesses
- Extreme customer concentration in a handful of hyperscalers/neoclouds — order timing is lumpy and AI-capex-dependent
- Valuation carries no discount to peers — ~24.2x FY27 consensus EPS against a 23.8x median for the named power and cooling peers (estimates, aggregator), and ~33x the FY26 guide midpoint — a multiple that still needs the Q4 ramp delivered
- EMEA organic swung from -29% in Q1 FY26 to +19% in Q3 FY26 — regional volatility beneath the Americas-led headline (Americas +66% organic in Q3)
- COGS exposed to copper/steel/semiconductor inflation and tariffs — gross margin not as transparently disclosed as adj op margin
- Backlog quality risk — large project orders can be pushed, re-scoped or cancelled if AI-capex sentiment turns
Opportunities
- 100kW+ → 250kW+ rack roadmap makes liquid cooling the default for any 2027+ hyperscale facility — TAM expansion
- Liquid-cooling market projected ~26% CAGR ($5.7B 2026 → ~$29B 2033) with Vertiv as a share leader
- Grid-power and on-site generation bottleneck pulls Vertiv up-stack into MV power, switchgear, and prefab power modules
- Services/aftermarket monetization of a rapidly growing installed base (recurring, high-margin)
- Bolt-on M&A in thermal IP (ThermoKey, Strategic Thermal Labs) to widen the cooling moat
- UtilityInnovation Group (closed Oct 2026) adds microgrid controls, microgrid switchgear and behind-the-meter power design, extending Vertiv to the grid interconnect for power-constrained sites
Threats
- AI-capex cyclicality/air-pocket — a hyperscaler digestion pause would hit orders and crush the multiple
- Schneider Electric and Eaton (post-Boyd Thermal) pushing full power+liquid-cooling stacks; Supermicro single-vendor rack-scale loops
- Hyperscalers in-sourcing/standardizing cooling (ODM/white-box CDUs) compressing price
- Commodity/tariff inflation and supply constraints on copper, power semis, and CDUs
- Multiple de-rate risk — at ~24.2x FY27 consensus EPS the stock sits at the named-peer median (23.8x), so a de-rate to the cheapest named power and cooling peer (Schneider Electric, ~18.9x) would take ~22% off the price even with strong execution (computed; estimates, aggregator)
- Integration and earn-out risk on UtilityInnovation Group — ~$1.45B paid in cash at about 13x expected 2027 EBITDA (company figure), with up to $1.15B more contingent on EBITDA targets
Moats, dependencies & bottlenecks
Moats
Scale + certifications to bid hyperscaler liquid-cooling at production volume One of a handful of vendors with the manufacturing footprint and qualifications to win large GB200/GB300 cooling projects today.
Nvidia reference-architecture co-development (OneCore / CoolChip CDU) Designed-in for GB200/GB300/Vera Rubin; advantage real but Nvidia also partners with rivals, so not exclusive.
Large global installed base of power/thermal gear generates recurring service revenue and high switching costs.
Single-vendor power+cooling+rack lowers integration risk for buyers; Schneider/Eaton chase the same bundle.
Heat-exchanger/CDU know-how widens the cooling lead but is replicable via acquisition (cf. Eaton/Boyd).
Dependencies
Google, AWS, Meta, Oracle, CoreWeave et al.) Demand and order timing are dictated by a few buyers' AI buildout cadence; a digestion pause hits orders directly.
Liquid-cooling pull is keyed to Nvidia accelerator densities and volumes; a GPU air-pocket flows straight through.
COGS and lead times exposed to commodity inflation, tariffs, and power-electronics availability.
Whole thesis rides a single, potentially front-loaded capex super-cycle; reversal would compress growth and multiple together.
Power-interconnect and permitting bottlenecks can delay the projects that drive Vertiv orders (double-edged: also expands its power TAM).
Advantages
- First-mover scale and certifications in production liquid cooling for 100kW+ racks
- >$15B backlog and ~2.9x book-to-bill at Dec 31 2025 (no dollar figure disclosed since) — visibility few hardware peers can match
- Full power+thermal+rack platform (OneCore) wins larger, stickier, single-vendor deals
- Strong balance sheet (~$0.8B net cash and $6.2B liquidity at Sep 30 2026, before the ~$1.45B UtilityInnovation Group payment) self-funds growth and M&A
- Rapid margin expansion (25.2% adj op margin Q3 FY26 → >27% 2030 target) on operating leverage + mix
- Deep Nvidia co-development presence in the AI reference architecture
Weaknesses
- Customer concentration in a handful of AI-capex buyers — lumpy, sentiment-sensitive orders
- Valuation in line with named peers (~24.2x FY27 consensus EPS vs a 23.8x median; estimates, aggregator) leaves a smaller error budget than in June, when the forward multiple was ~46-48x
- Regional volatility — EMEA organic swung from -29% in Q1 FY26 to +19% in Q3 under an Americas-led headline (Americas +66% organic in Q3)
- COGS exposed to copper/steel/semiconductor inflation and tariffs; gross margin less transparently disclosed
- Backlog conversion/cancellation risk on large AI projects
- Intensifying competition from Schneider, Eaton (Boyd Thermal), Supermicro, and ODM in-sourcing
Bottlenecks
- Order timing controlled by a few hyperscalers — revenue is lumpy and visibility hinges on backlog conversion, not steady demand
- Manufacturing/assembly capacity for CDUs and integrated liquid-cooled racks must scale ahead of orders without over-building
- Copper, steel, and power-semiconductor supply + tariffs constrain COGS and lead times
- Grid-power interconnect and data-center permitting can delay the customer projects Vertiv is slotted into
- Skilled field-engineering/installation labor to deploy and service complex liquid-cooling at hyperscale pace
Top signals & trends
Top signals
The leading indicator. Q3 FY26 gave only a growth rate (organic orders more than 100% above a year earlier, company statement of 28 Oct 2026), and management said on the call that it does not disclose orders and does not plan to share orders data going forward, so book-to-bill cannot be observed from here and the signal is qualitative; before this, any deceleration below ~1.0x book-to-bill was the first crack in the thesis.
Continued growth confirms multi-year visibility; flattening/decline would signal AI-capex cresting. No dollar backlog has been disclosed since Dec 31 2025, so this signal cannot be read directly; organic orders growth above 100% in Q3 FY26 (company statement) is the only current datapoint.
Vertiv's demand is downstream of these prints — watch for any digestion language.
Margin walk is the under-modeled value driver; a stall reframes the multiple.
Share retention as rivals add liquid cooling determines pricing power.
The original pillar of the paired-short overlay was a ~46-48x forward multiple near a five-year high. At $235.97 the stock trades at ~24.2x FY27 consensus EPS and ~33.0x the FY26 guide midpoint, about the 23.8x median of its named power and cooling peers (computed; estimates, aggregator), so that premium is gone. The overlay rests on AI-capex cyclicality and the working-capital and disclosure risks in the bear case rather than on a stretched multiple.
Trends
Converts liquid cooling from optional to mandatory — Vertiv's core, highest-margin growth line.
Primary demand engine; also the single biggest cyclical risk if it inflects.
Large, fast-expanding TAM where Vertiv is a share leader.
Expands Vertiv's power TAM — its UtilityInnovation Group acquisition (closed Oct 2026) targets this constraint with microgrid controls and behind-the-meter power design — but can delay customer projects that aren't yet powered.
Erodes the cooling-vendor scarcity premium over time; pressures price/share.
OneCore aligns with buyer desire to de-risk integration — larger, stickier deals.
Ecosystem & competitor graph
Suppliers feed the company; customers pull from it. Line thickness shows the strength of each tie (supply-chain dependency, customer earnings contribution). Hover to isolate a tie.
Copper / steel / aluminum commodity suppliers Core raw materials for heat exchangers, busbars, enclosures — primary COGS input and inflation exposure.
power modules) Infineon and peers supply power electronics for UPS/power distribution.
Power/analog semis used in power-conversion and control electronics.
Italian heat-exchanger specialist — vertically integrated cooling IP/supply.
Strategic Thermal Labs / Great Lakes-type tuck-ins (acquired) Liquid-cooling and rack/integration capability brought in-house via M&A.
Mechanical components for thermal/CDU systems.
Hyperscaler — among the largest AI data-center capex buyers.
Hyperscaler data-center power/cooling demand.
Hyperscaler AI buildout.
Large AI training cluster buildout — liquid-cooling demand.
Rapidly expanding AI cloud capacity.
CoreWeave / neoclouds + Digital Realty / Equinix colos Neoclouds and colocation operators (DLR, EQIX) are fast-growing direct buyers.
Closest full power+cooling rival; ~18.9x FY27 consensus EPS (Nov 3 2026 close; estimate, aggregator); expanding liquid cooling (Motivair). Direct hyperscaler competitor.
Full power+cooling; added liquid cooling via Boyd Thermal (Mar 2026); ~24.0x FY27 consensus EPS (Nov 3 2026 close; estimate, aggregator). Direct, well-capitalized threat.
Single-vendor rack-scale liquid-cooled servers with integrated cold plates/CDUs — bypasses standalone cooling vendors.
HVAC/thermal scale player pushing into data-center cooling; ~23.6x FY27 consensus EPS (Nov 3 2026 close; estimate, aggregator).
Building/data-center thermal and chillers; broad HVAC reach into data centers.
CoolIT (Ecolab-owned) is a leading direct-to-chip/CDU specialist supplying OEMs and hyperscalers.