Energy
A boring enclosures-and-connectors company that bought and sold its way into the AI data-center build — now growing 50%+ with infrastructure orders up 100% organically; the re-rating to ~28x is largely earned, but the multiple now prices in the build continuing, and the legacy two-thirds of the business is still cyclical late-cycle industrials.
Research
The verdict
A boring enclosures-and-connectors company that bought and sold its way into the AI data-center build — now growing 50%+ with infrastructure orders up 100% organically; the re-rating to ~28x is largely earned, but the multiple now prices in the build continuing, and the legacy two-thirds of the business is still cyclical late-cycle industrials.
nVent is a "electrical connection and protection" company — it makes the physical hardware that connects and protects electrical systems: enclosures, cooling (liquid and air), bus systems, switchgear, cable management, electrical connections, fastening, and power management. It is an Ireland-incorporated plc, principally officed in London with US management in Minneapolis, listed on the NYSE as NVT. ~12,000 employees, ~48% in the US.
It was spun out of Pentair in April 2018 as the standalone electrical business — that lineage matters: this is a ~100-year-old collection of industrial brands (CADDY, ERICO, HOFFMAN, ILSCO, SCHROFF, TRACHTE) packaged into a focused mid-cap, not a startup.
Two reportable segments (renamed in 2025):
Business model. Engineer-to-spec and catalog hardware sold mostly through the channel — electrical distributors, retail, contractors, OEMs, panel builders — plus an increasing slug of direct, long-cycle "major projects" (data centers, power utilities) booked into backlog. Most catalog product ships within 90 days; infrastructure project work can take >1 year. Backlog was $2.3B at 2025-12-31 (up sharply on the EPG acquisition + data-center growth) and $2.6B by Q1-2026.
Customers / concentration. End markets: infrastructure, industrial, commercial & residential, energy. The largest customer was ~11% of consolidated net sales in 2025 — a real concentration, almost certainly a hyperscaler or a data-center integrator, given Systems Protection's mix. No other customer >10%. ~24% of net sales come from products sourced via non-US suppliers.
Upstream inputs. Principal raw materials are mild steel, stainless steel, plastic resins, copper, and aluminum — classic metal-bashing inputs. This makes gross margin directly exposed to commodity prices and, in 2025-26, to tariffs: management explicitly cites tariff-driven raw-material and labor inflation as a gross-margin headwind, and quantifies a ~$80M tariff impact embedded in 2026 guidance. ~24% of net sales rely on non-US suppliers — a second-order tariff/FX exposure.
Midstream (nVent). "Sell globally but serve locally" — regional manufacturing and regional supply chains. This is a deliberate moat against both freight cost and tariff exposure: localized production lets them serve data-center builds in-region.
Downstream / end customers (named buyer classes): hyperscalers, utilities, OEMs, panel builders, contractors, electrical distributors, retail. The single most important downstream node is the AI data-center buildout — the white-space (inside-the-rack: liquid cooling, PDUs, cable management) and gray-space (engineered buildings, enclosures, power connections) of hyperscale and colo facilities.
Chokepoints / single-source. The filing does not name single-source supplier dependencies, but two structural chokepoints exist: (1) copper and steel pricing — a margin chokepoint, not a supply one; (2) the largest-customer 11% concentration — a demand chokepoint. The supply chain is otherwise fragmented and regionally redundant by design. ``
The moat is mid-strength and mostly built on spec-in + brand + channel, not technology.
Where the moat is weak. Management itself calls the markets "highly competitive… large and well-established national and global companies, as well as regional and local companies and lower-cost manufacturers," with smaller players "compet[ing] primarily on price". Enclosures and liquid cooling have no IP moat strong enough to stop a Vertiv, Schneider, or Rittal from competing head-on for the same data-center dollar (Lens 7/13). nVent's bargaining power over hyperscaler customers is limited — when ~11% of revenue is one buyer in the middle of a capex super-cycle, the buyer has options.
All (FY) and (Q1). Continuing operations only (Thermal Management is discontinued — see Lens 5).
By segment — net sales ($M):
| Segment | 2023 | 2024 | 2025 | Q1-2025 | Q1-2026 |
|---|---|---|---|---|---|
| Systems Protection | 1,605.9 | 1,823.3 | 2,592.9 | 508.2 | 894.8 |
| Electrical Connections | 1,063.0 | 1,182.8 | 1,300.2 | 301.1 | 347.2 |
| Total | 2,668.9 | 3,006.1 | 3,893.1 | 809.3 | 1,242.0 |
By segment — reportable segment income ($M):
| Segment | 2023 | 2024 | 2025 |
|---|---|---|---|
| Systems Protection | 346.6 | 403.1 | 537.0 |
| Electrical Connections | 330.6 | 354.5 | 372.6 |
| Total | 677.2 | 757.6 | 909.6 |
Read. Systems Protection is the engine and is accelerating hard: +42% reported in FY2025 (vs +14% in FY2024), then +76% in Q1-2026 (organic + the Avail/EPG acquisition). Electrical Connections is the slower, steadier flywheel: +10% FY2025, +15% Q1-2026. The mix has tilted decisively toward Systems Protection — 67% of FY2025 sales and 72% of Q1-2026 sales — which is exactly the data-center-leveraged half. Cause: AI data-center demand + the EPG acquisition concentrated in Systems Protection.
By geography (FY2025, $M): Americas $3,158.3 (81%), EMEA $587.3 (15%), Asia-Pacific $147.5 (4%). This is a US-centric business — the AI data-center build is overwhelmingly a North American story for nVent, and Americas grew from $2,014.9M (2023) → $2,325.5M (2024) → $3,158.3M (2025), a +36% YoY 2025 jump.
Margin note. Both segments saw segment-income margin compress modestly in 2025 (inflation/tariffs on materials + labor, plus growth investment), partly offset by operating leverage and productivity. Segment income still grew 20% on 30% sales growth — operating leverage is real but margin is not expanding; it is being defended.
The number. Net sales $1,242.0M, +53.3% YoY (Q1-2025 $809.3M), of which ~34% organic. Reportable segment income $287.9M vs $189.3M.
EPS. GAAP diluted $0.87 vs $2.16 prior year — but the prior-year $2.16 included $1.64 of discontinued-operations gain (the Thermal Management sale closed Jan 2025); continuing-ops diluted was $0.86 vs ~$0.52. On the non-GAAP/adjusted basis management and the Street track, adjusted EPS $1.09, +62.7%, beating consensus by ~16%.
What drove it. Systems Protection +76%; infrastructure within Systems Protection grew >100% organically on AI data-center demand across gray space (engineered buildings, enclosures, power connections) and white space (liquid cooling, PDUs, cable management).
Orders / backlog. Organic orders +~40%, data-center-led (ex-data-center, mid-teens); backlog $2.6B, up low-double-digits sequentially, with order visibility into 2027. This is the most important single datapoint in the dossier: orders are outrunning revenue, so the growth is booked, not hoped.
Guidance — raised, twice. FY2026 reported sales growth lifted to 26-28% (from 15-18%), organic to 21-23% (from 10-13%); adjusted EPS to $4.45-$4.55 (from $4.00-$4.15) — with ~$80M of tariff drag still embedded. Management tone shifted clearly more confident QoQ.
Balance-sheet flags. FCF +21% YoY in Q1. No working-capital alarm in the filing; receivables/inventory are building with the backlog, which is normal for a project-heavy ramp but worth watching (Lens 10). Total debt $1,559.8M at 2025-12-31, down from $2,155.0M after the Thermal-sale deleveraging.
Market reaction. The tape has already paid for this: NVT is up from a 52-week low of $68.60 to $167.34 (2026-06-16), an all-time high $178.00 on 2026-06-02 — roughly a 2.4x in twelve months. The market is treating each beat-and-raise as confirmation of a structural AI-infrastructure re-rating, not a cyclical blip.
Unusual vs. its own history. A 53% revenue quarter and a doubling of infrastructure organic growth are completely outside nVent's historical ~5-10% organic band. This is a regime change, driven by one secular wave (AI capex) layered on top of M&A — both of which can reverse.
No transcripts are on disk (transcripts/ empty), so this is ``, anchored on the Q1-2026 call plus the Q4-2025 release.
What management is focused on (Q1-2026): the AI data-center buildout, explicitly framed as both white-space (liquid cooling — "AI data center liquid cooling orders surge") and gray-space (engineered buildings, enclosures), and the portfolio transformation paying off.
Tone shift over the last ~3 prints. Q3-2025 → Q4-2025 → Q1-2026 is a steepening curve: the recurring phrases moved from "data centers a growth vertical" to "record orders and backlog," "organic orders +40%," "visibility into 2027," and two guidance raises in two quarters. The thing they stopped hedging on is the durability of data-center demand — management now talks about it as a multi-year build, not a quarter.
What to listen for next. Any softening of the "ex-data-center mid-teens" line (the non-AI core), and any commentary on white-space content per rack as liquid cooling scales — that is the real bull lever.
Peers are the data-center electrical/power/cooling ecosystem: Vertiv (the pure-play AI-cooling/power name), Eaton (diversified electrical, big DC exposure), Hubbell (electrical products, DC + utility), plus private/foreign Schneider, ABB, Legrand, Rittal. Multiples ``, dated; where I could not source a clean figure I mark n/a.
| Company | Ticker | Mkt cap (USD) | Fwd P/E | EV/EBITDA | Div yield | 5-yr avg ROE |
|---|---|---|---|---|---|---|
| nVent | NVT | ~$27.1B | ~28-34x | ~24x | ~0.48% | ~13-16% |
| Vertiv | VRT | $112.9B | 44.4x | 49.1x | n/a | n/a |
| Eaton | ETN | $164.2B | 31.8x | 27.7x | n/a | n/a |
| Hubbell | HUBB | $26.0B | 24.9x | 19.6x | n/a | n/a |
Read. NVT trades below the AI-cooling darling (Vertiv, 44x/49x) and roughly in line with to slightly above Eaton (32x/28x) and Hubbell (25x/20x) on forward P/E, but its EV/EBITDA (~24x) sits above Hubbell and Eaton — the market is paying a premium-to-old-industrials, discount-to-Vertiv multiple. That is the central valuation fact: NVT is priced as a credible data-center-electrical play but not (yet) as a pure-play. If you believe the white-space/liquid-cooling content story, NVT looks like the cheaper way to own the same theme as Vertiv; if you think the legacy two-thirds drags, it looks fully priced versus Hubbell.
``. The 52-week range alone ($68.60 → $178.00) tells the story: this is a stock that re-rated on a theme, with the biggest moves clustered around (1) quarterly beat-and-raise prints and (2) portfolio/M&A events.
What the market actually reacts to: the data-center order book and guidance revisions, far more than the legacy industrial cycle. The pattern reveals NVT has become, in the market's mind, an AI-infrastructure capex derivative — which cuts both ways (Lens 13).
CEO: Beth Wozniak — Chair & CEO. Named CEO in 2017 ahead of the 2018 Pentair spin; 25 years at Honeywell (HVAC/critical technologies) before running Pentair's electrical business; added board Chair in 2023; also sits on Carrier and Parker Hannifin boards.
+ unless noted.
Regulatory findings (required).
"nVent" (FTC OR DOJ OR settlement OR fine OR penalty OR lawsuit) enforcement returned no material company-specific enforcement actions.Built bottom-up from management's own guidance + the order book. All outputs ``, inputs labeled.
FY2026 (base = management guide): Adjusted EPS $4.50 — the midpoint of the raised $4.45-$4.55 guide, on 26-28% reported / 21-23% organic sales growth, ~$80M tariff drag absorbed. This is unusually high-confidence for a base year because backlog ($2.6B) + order visibility into 2027 underpin it.
FY2027 (base): Adjusted EPS ~$5.30. Visibility "into 2027" supports continued double-digit growth but the law of large numbers and tougher comps argue against repeating 2026.
FY2028 (base): Adjusted EPS ~$6.00. Bull ~$6.70 / Bear ~$5.00 — by 2028 the swing factor is whether AI-DC is still a growth vertical or has normalized into a base.
Path summary ($ adjusted EPS):
| Case | FY2026 | FY2027 | FY2028 |
|---|---|---|---|
| Bull | 4.65 | 5.75 | 6.70 |
| Base | 4.50 | 5.30 | 6.00 |
| Bear | 4.20 | 4.70 | 5.00 |
Valuation cross-check. At $167.34 and base FY2026 $4.50, NVT trades ~37x current-year and ~31.6x FY2027 / ~27.9x FY2028 base. That is a full multiple for a company two-thirds of whose revenue is still cyclical industrials — the price already capitalizes several years of the DC build going right. Note: forecast.ts create was intentionally skipped per the unattended --watchlist rules.
Bull case. nVent is the best risk-adjusted way to own the "picks-and-shovels of the AI data hall" below Vertiv's nosebleed multiple. It sells both gray-space (enclosures, engineered buildings, power connections — where it just bought EPG) and white-space (liquid cooling, PDUs, cable management), so as data halls densify and liquid cooling goes mainstream, its content per megawatt rises. Orders are +40% organic with visibility into 2027 — the growth is booked, not narrative. Management has proven it can allocate capital (sold Thermal at the top, bought growth, deleveraged, bought back stock). The legacy Electrical Connections core (+15%) gives a real, non-AI floor. At ~28-34x forward vs Vertiv's 44x, there is room to re-rate toward the pure-play if white-space mix keeps climbing. The contrarian-bull take the market under-appreciates: nVent is quietly more of a white-space liquid-cooling story each quarter, and is still mentally filed by many investors as a sleepy enclosures company — a perception lag that closes as the mix shifts.
Bear case (2-3 permanent-impairment risks).
Pre-mortem (18 months out, thesis broke). It's late 2027. Hyperscalers guided down 2027 capex in mid-2026; nVent's +40% organic orders went to flat by Q4-2026; the $2.6B backlog shipped but wasn't replaced; two guidance cuts followed; and the stock round-tripped from $178 toward $100 as the multiple compressed from ~30x to ~18x on a now-decelerating base. The tariff drag came in worse than $80M, squeezing the metal-input margin. Most likely actual cause: an AI-capex digestion pause colliding with a full valuation.
Are multiples too high? For the base path, ~31.6x FY2027 is rich but defensible if DC growth persists; the asymmetry is unattractive — limited re-rating headroom toward Vertiv, large de-rating downside toward Hubbell. The risk/reward at $167 is worse than it was at $90.
Dismantling the bull case.
A consolidating monopolist on a leveraged treadmill — RUN's GAAP "profit" is an HLBV mirage, but the OBBBA's asymmetric kill of 25D (not 48E) hands the TPO leader the residential market it can't yet profitably finance; the bet is whether ~$15B of non-recourse debt rolls before rates or a securitization-market hiccup forces a dilutive reset.
A real turnaround that has already been paid for — six straight quarters of margin repair and the return to positive operating cash flow are genuine, but at ~$60 the stock prices in a clean, AMPTC-independent recovery the filings explicitly say does not yet exist (ex-45X credits, SolarEdge is still gross-loss-making), so the asymmetry from here is poor.
A re-accelerating solar-EBOS niche-monopolist (FY26 guided +30%, record $758M backlog, IP win pending) trading at ~27x fwd P/E on a near-empty balance sheet ($1.9M cash, revolver 91% drawn) — the growth is real but the equity is priced for it and the funding runway is the actual risk, not the (now-settled) shrinkback overhang.