ETN › analyze
ETN · Analyze from before
This updates: Output/Stocks/Infrastructure/ETN/analyze-2026-03-21.md (2026-03-21, price ~$357,
conviction 7.5, target entry $300-320, no explicit trim). No health/moat/value/sentiment/
earnings file was written on ETN between the baseline and this pass, so the March file is the
sole prior record. 159 days elapsed.
What happened since the baseline
- Two large acquisitions closed in Q1 2026, funded almost entirely by new debt. Ultra PCS Limited ($1.53B, Jan 23) and Boyd Thermal ($9.55B, March 12) — combined ~$11B, versus a company that generated $3.55B of FCF in all of FY2025. Financed by $8.5B of new senior notes (six tranches, 3.85%-5.45%, 2028-2056 maturities) plus up to $2.5B of commercial paper.
- Boyd Thermal is a liquid-cooling business for AI data centers — Eaton explicitly frames it as completing an "industry-leading grid-to-chip solution," extending the moat from power distribution into thermal management, the layer hyperscalers cannot outsource around.
- The Mobility separation changed mechanism. The March baseline described a standalone spinoff. On 2026-06-11 Eaton instead announced a Reverse Morris Trust combination with Dana Incorporated — Eaton shareholders get ≥50.1% of the combined entity plus a ~$1.1B cash distribution; Dana holders get ~49.9%. Values Mobility at $5.1B (5.9x synergized 2026 pro forma EBITDA). Tax-free, still targeted to close Q1 2027 — the timeline held, the structure didn't.
- Two consecutive guidance raises. Q1 2026: organic growth guide raised to 10% midpoint. Q2 2026: full-year organic growth raised again to 11-13% (from the original 7-9%), adjusted EPS guide raised to $13.40-$13.60 (from $13.00-$13.50). Q2 adjusted EPS $3.15 was a Q2 record, above guidance; segment margin 23.1%, above the high end of guidance.
- US data center backlog is now quantified: 307 GW, ~15 years of 2025 build rate. Total electrical backlog +43% YoY; Americas book-to-bill 1.3; total company book-to-bill 1.2.
- GAAP diluted EPS fell YoY — TTM diluted EPS $9.81 vs FY2025's $10.45 — driven by the new interest expense (Q1 2026 interest expense $106M vs Q1 2025's $33M), acquisition-related purchase-accounting charges, and deal costs. This is a real divergence between GAAP and adjusted earnings, not a data error (see Phase 2 below).
- Buybacks paused in Q1 2026 ($0 repurchased) to preserve cash/debt capacity for the deals. Share count is flat, not falling (387.9M → 388.3M).
- Dividend raised 6% to $1.10/quarter ($4.40/yr) in Feb 2026 — the 18th consecutive annual increase — but the growth rate decelerated from the ~7-8% CAGR the baseline cited.
- Price ran from ~$357 to $416.04 (+16.5%), market cap $138.6B → $161.55B. Analyst mean target moved $396→$477, median $414→$487.
Data trap hit this pass
pitfall-vendor-forward-eps-is-the-wrong-fiscal-year fired on ETN, live, in this session's own
default tool. .mcp/fin.py ETN --news prints PE(fwd) 25.95 from Yahoo's forwardPE/forwardEps
($16.03) — that is the FY2027 consensus, not FY2026. The correct current-year field,
epsCurrentYear ($13.515) / priceEpsCurrentYear (30.78x), matches the company's own FY2026
adjusted-EPS guide midpoint ($13.50) almost exactly. True FY2026 forward P/E is ~30.8x, not
25.95x — a full multiple-turn-plus understatement, and it is the number this report's valuation
uses. Left uncorrected, it would have made ETN look ~16% cheaper than it is and manufactured a
false margin of safety, exactly the failure mode the pitfall describes.
Also applied, cleanly this pass: TTM reconstructed as the sum of four quarters, not read off
one column (pitfall-single-quarter-fcf-read-as-ttm). Q3'25 $1,173M + Q4'25 $1,573M + Q1'26 $314M
+ Q2'26 $874M = $3.93B TTM FCF. This matters here specifically because Q1 2026's $314M alone
would look like a collapse if mistaken for a full-year figure — it is one quarter of a business
that just spent $11B on acquisitions, not a trend.
The delta ledger
Lead rows are RETRACTED/SUPERSEDED — read those first. Type: Structural (business model) · Trend (multi-year series) · State (point-in-time) · Price (multiple/FV/zone) · Judgment (derived).
🔄 SUPERSEDED
| # | Claim (baseline) | Type | What replaced it |
|---|---|---|---|
| 6 | "Debt-to-assets stable ~25.5% — management is disciplined" | Trend | Debt/assets jumped to ~39.6% at 2026-03-31 (total debt $10.53B→$21.83B, total assets $41.25B→$55.09B) on the $11B debt-funded acquisitions. The discipline claim as stated no longer holds — management redirected leverage capacity toward growth M&A rather than holding the ratio flat. Not an error: a decision, made in Q1 2026, that the baseline could not have seen. |
| 7 | "Net debt/EBITDA 1.5x — very comfortable" | State | Naive TTM-standalone net debt/EBITDA is now ~3.1x ($20.56B net debt ÷ $6.64B trailing EBITDA, neither figure yet carrying a full year of the acquired businesses). Moody's assigned A3 to the new notes and models pro-forma leverage staying below 2.5x as acquired EBITDA annualizes — investment-grade, not distressed, but genuinely more levered than "very comfortable" describes today. |
| 11 | "Mobility separation: standalone spinoff by Q1 2027" | Structural | Restructured 2026-06-11 into a Reverse Morris Trust combination with Dana Incorporated — Eaton keeps ≥50.1% of the combined entity, receives a ~$1.1B cash distribution, and the deal is valued at 5.9x synergized EBITDA. Economically more favorable than a bare spinoff (Eaton gets paid to leave), but it adds Dana-shareholder-vote and regulatory-approval risk the baseline's simple spinoff didn't carry. Close timeline (Q1 2027) held. |
❌ RETRACTED
| # | Claim (baseline) | Type | Error and cause |
|---|---|---|---|
| 22 | "Target Entry $300-320 (confirmed from screen suggestion)" | Price | Stale, not wrong-when-written. The zone was correctly derived off the March fair-value case; the price has since run 31% past it (per the 2026-08-27 watchlist scan flag), and the fair value itself has moved on a real EPS-guide raise plus a real multiple re-rating. [[pitfall-stale-entry-zone-suppresses-a-name]] — re-derived below. |
📉 DRIFTED
| # | Claim (baseline) | Type | Level at which it breaks further |
|---|---|---|---|
| 5 | "Shares outstanding: net -2.5% over 3 years via buybacks, ~0.8%/yr" | Trend | Buybacks paused entirely in Q1 2026 ($0 repurchased); share count is flat (387.9M→388.3M), not shrinking. Not reversed — no net issuance — but the per-share tailwind stopped. Breaks further (becomes a real negative) if the pause extends past 2-3 more quarters while net debt/EBITDA sits above ~2.5x. |
| 21 | "Trading at the upper end of fair value, minimal margin of safety" (fwd P/E 26.9x at $357) | Price | True fwd P/E is now 30.8x at $416 — a genuine re-rating, not a data-quality artifact (the baseline's own 26.9x was already computed off the correct current-year field). The margin-of-safety gap the baseline flagged as thin has gotten thinner. Re-derived below; breaks further past ~34x (the trim line). |
⏳ UNTESTED this pass
| # | Claim (baseline) | Type | Why untested |
|---|---|---|---|
| 8 | "ROIC expanding 10.2%→14.6% (2022-2025)" | Trend | Post-acquisition ROIC needs a full quarter of Boyd Thermal/Ultra PCS invested-capital and NOPAT contribution to compute honestly; Q1-Q2 2026 carry partial-period purchase accounting that would distort a same-methodology recompute. Flagged for the next pass once FY2026 closes. |
| 17 | "Institutional ownership: net buying (1,016 added vs 938 decreased)" | State | Not re-pulled this pass — current level (88.1% institutional) is known, but the quarter-over-quarter flow direction was not independently re-verified. |
✅ CARRIED / 🔁 REFRESHED (compact)
| # | Claim | Type | Status | Note |
|---|---|---|---|---|
| 1 | Revenue growth ~9.8% 3yr CAGR | Trend | 🔁 REFRESHED | Accelerated — Q2'26 sales +21% YoY (14% organic + 7% acquired); FY26 organic guide raised twice to 11-13% |
| 2 | FCF 3yr CAGR 22.4%, "exceptional" | Trend | 🔁 REFRESHED | TTM FCF $3.93B vs FY25's $3.55B (+10.8%); FY26 guide reaffirmed $3.9-4.3B at Q2 |
| 3 | FY26 FCF guidance $3.9-4.3B | State | ✅ CARRIED | Restated unchanged at Q2 2026 (OCF $5.0-5.4B, FCF $3.9-4.3B) despite the new interest burden — organic cash generation is absorbing it |
| 9 | Dividend: 17yr streak, ~7-8% CAGR, 46% FCF payout | Trend | 🔁 REFRESHED | Streak now 18yr (Feb 2026, +6%); growth rate decelerated to 6% from the 7-8% cited |
| 10 | R&D modest, "engineering-driven not R&D-intensive" | Structural | ✅ CARRIED | No contradicting evidence |
| 12 | NVIDIA partnership | Structural | ✅ CARRIED | No contradicting news; Boyd Thermal reinforces the same positioning |
| 13 | Siemens Energy JV | Structural | ✅ CARRIED | No contradicting news found |
| 14 | Macro tailwinds (DC power, grid modernization, IRA/IIJA, electrification) | Structural | 🔁 REFRESHED, strengthened | Now quantified: 307GW US DC backlog (~15yr of 2025 build rate), total backlog +43% YoY, book-to-bill 1.2-1.3 |
| 15 | Analyst consensus Buy, mean target $396 | Price | 🔁 REFRESHED | Mean $477 / median $487 (range $333-534), still consensus Buy |
| 16 | Insider activity "mixed, no concerning patterns" | State | ✅ CARRIED | One director (G. Johnson) made small, genuine open-market buys repeatedly through Aug 2026 ($68k-$390k each); officer sales remain routine RSU-driven diversification. No new red flag. |
| 18 | Moat sources: switching costs/spec-in, product breadth ("grid-to-chip"), scale, installed base, regulatory barriers | Structural | ✅ CARRIED, strengthened | Boyd Thermal explicitly extends grid-to-chip from power into liquid cooling; Ultra PCS deepens aerospace breadth |
| 19 | Adversarial stress test (Chinese competition low, software-layer capture manageable, tech disruption low, customer-concentration low-moderate) | Structural | ✅ CARRIED | No predicted risk materialized. The hyperscaler-vertical-integration fear the baseline flagged as its main watch item moved the other direction — the 307GW backlog says Eaton is being pulled deeper into the supply chain, not disintermediated |
| 20 | Evergreen rating 8/10, Moat grade A- | Judgment | ✅ CARRIED | See adjudication below |
How the close calls were decided
Is the leverage jump a moat/quality problem or a financing-structure problem? The Fundamentals read (debt/assets 25.5%→39.6%, naive net debt/EBITDA 1.5x→~3.1x) looks alarming in isolation. Set against it: (1) both deals are strategically coherent, not empire-building — Boyd Thermal fills the one gap (thermal) in an otherwise complete power-to-chip stack, and Ultra PCS deepens the higher-margin aerospace segment; (2) Moody's re-affirmed investment-grade A3 on the new notes and models leverage returning under 2.5x as acquired EBITDA annualizes — a rating agency with access to the same acquisition disclosures does not read this as a credit problem; (3) FCF guidance was reaffirmed, not cut, meaning management itself does not expect the interest burden to break cash generation. Verdict: real added risk, correctly downgrades the "very comfortable" claim to SUPERSEDED, but does not on its own downgrade the moat or the business-quality read — it is priced as a financing decision, weighed most heavily by Fundamentals and Valuation, not by Moat.
Is the GAAP EPS decline a red flag or a reporting artifact? Both, in different proportions.
Real: interest expense genuinely tripled quarter-over-quarter (Q1 2026 $106M vs Q1 2025 $33M) and
purchase-accounting amortization on $11B of newly acquired intangibles is a real, recurring GAAP
charge, not a one-time add-back to wave away. Artifact-adjacent: adjusted EPS — the metric
management guides to and the metric the Q2 beat-and-raise was measured against — grew and beat
guidance. Both numbers are load-bearing: the GAAP decline is the honest cost of the acquisitions
this year; the adjusted growth is the honest read of underlying operating momentum. Reporting the
GAAP figure alone (as the naive EPS(ttm) field does) without the reconciliation would have
understated the business; reporting only adjusted EPS would have hidden the debt cost. This report
carries both.
Does the mobility deal's new complexity (Dana shareholder vote, regulatory approval) undercut the "catalyst" framing? Partially — it is genuinely more conditional than a pure spinoff. But it also converts a value-neutral separation into a value-positive one (Eaton receives $1.1B cash and 50.1%+ of a larger combined entity rather than simply distributing shares). Net: SUPERSEDED (mechanism changed) rather than DRIFTED (thesis weakening) — the strategic logic (shed the lower-growth, lower-margin Mobility segment) is intact and the economics improved.
Thesis persistence and conviction delta
Thesis persistence (Structural + Trend claims, CARRIED or REFRESHED): 9 of 13 = 69%. High persistence — the moat widened rather than eroded, the demand backdrop strengthened and got more quantified (307GW backlog), and the dividend streak extended. The four non-survivors were the leverage/discipline claim (SUPERSEDED), the buyback-driven share count claim (DRIFTED), the mobility-separation mechanism (SUPERSEDED), and ROIC (UNTESTED) — none of them a moat or demand-thesis failure.
Conviction: 7.5 → 7.0. This is not the standard "business held, multiple ran, trim the conviction gap" case alone — three forces moved at once, and they partially offset:
- Business quality/demand: up. Backlog visibility (307GW, +43% YoY), two consecutive guidance raises, and a moat-widening acquisition (Boyd Thermal) are a stronger position than March.
- Valuation: more expensive, on the corrected basis. True fwd P/E 26.9x→30.8x. Not a data error either time — a genuine 16.5% price move against a smaller (~2%) EPS-guide raise.
- Balance sheet risk: new and real. Debt roughly doubled to fund two large, near-simultaneous acquisitions; buybacks paused; GAAP EPS fell YoY. Investment-grade and guided to delever, but a risk that did not exist in the March file.
Rows 6, 7, 21 (leverage, net debt/EBITDA, valuation) drove the downgrade; rows 1, 2, 14, 18, 19 (revenue/FCF trend, tailwinds, moat, stress-test) argued the other way and kept the downgrade to a half-point rather than a full step.
What is genuinely new (no baseline counterpart)
- Boyd Thermal / Ultra PCS acquisitions and their financing (Section "What happened," items 1-2).
- The Dana Incorporated RMT structure replacing the standalone Mobility spinoff.
- The 307GW US data center backlog figure — the baseline cited qualitative DC tailwinds without a quantified backlog number this large.
- The GAAP-vs-adjusted EPS divergence as a named, tracked issue.
Updated verdict
Valuation (re-derived from scratch, per convention)
| Metric | Baseline (Mar'26, $357) | Now ($416.04) |
|---|---|---|
| True current-year fwd P/E | 26.9x (FY26 adj EPS $13.25) | 30.8x (FY26 adj EPS guide midpoint $13.50) |
| Trailing P/E | 34.1x | 42.4x |
| EV/EBITDA | 23.9x | 27.65x |
| P/FCF (TTM) | 39.0x | 41.1x ($161.55B ÷ $3.93B) |
| Dividend yield | 1.23% | 1.05% |
| 5yr avg yield | ~1.1-1.4% | 1.55% |
DYT flag: current yield (1.05%) sits below its own 5yr average (1.55%) — the yield fell because price ran ahead of a decelerating dividend, not the reverse, so this is a straightforward "got expensive" signal, not the [[pitfall-dyt-inverts-when-price-caused-the-yield]] trap (that trap concerns a high yield manufactured by a falling price; here the yield is low, manufactured by a rising price — the ordinary direction). Reverting to the 5yr average yield implies ~$284 — consistent with Graham's $107 and DDM's $156-238 in direction (all low-weight for a goodwill-heavy, buyback-return compounder per the baseline's own reasoning, retained here), but all three now point further below spot than they did in March.
Fair value range:
| Scenario | Fair value | Basis |
|---|---|---|
| Bear | $325-350 | 24-26x FY26 adj EPS ($13.50) — macro slowdown, integration stumble, or leverage concern re-rates the multiple down |
| Base | $365-405 | 27-30x FY26 adj EPS — quality premium sustained, current guide holds |
| Bull (12-18mo) | $480-560 | 32-35x FY27E adj EPS (~$15-16, guide-consistent extrapolation, not vendor consensus) |
Bogle expected return: 1.05% yield + ~10-11% guided EPS growth − ~4.4%/yr if fwd P/E reverts 30.8x→~24x over 5yr ≈ ~7%/yr. Flat multiple: ~11.6%/yr. Reasonable, not exceptional, for the price paid — consistent with the baseline's own read at a lower price.
Note for the file, not yet a verdict change: the FY26 segment-margin guide (24.1-24.5%) is
already historically high for this business; per [[pattern-margin-expansion-is-a-finite-growth-lever]],
the 100-150bps/yr margin-expansion component of Eaton's EPS algorithm has a ceiling nobody has
named. Worth extrapolating explicitly at the next pass once a few more years of the post-Boyd
margin trajectory are visible.
Entry, trim, and triggers
- Fair value: $325-405 (central ~$365, base case)
- Entry: $325-350 — the bear-case band, ~15-20% below base-case central, replacing the stale $300-320 zone. At $335 midpoint: fwd P/E ~24.8x, yield ~1.31%.
- Strong-buy / back-up-the-truck: <$310 (near the 52-week low $311.92) — a level that would require a real macro or integration scare, not a routine pullback.
- Trim: 34x fwd (FY26 adj EPS basis, $13.50) → implies ~$459, ~10% above spot. Basis: fwd. Set at the top of the premium band a wide-moat, backlog-visible industrial compounder can sustain before the multiple itself becomes the primary risk — below where faster-growing AI-capex names in this file trade (ANET ~49x, AMD ~63x true fwd) but above ETN's own pre-AI-cycle historical range, reflecting the Boyd Thermal-widened moat.
Break triggers (would falsify this pass, not just move the price)
- Net debt/EBITDA fails to trend toward the <2.5x Moody's models by FY2026 year-end (integration cost overrun or EBITDA underperformance on the acquired businesses)
- Buybacks stay paused past Q1 2027 while leverage remains elevated
- Data center backlog growth decelerates or book-to-bill falls below 1.0
- GAAP-adjusted EPS gap widens further without a stated normalization path
- Dana RMT fails shareholder or regulatory approval (reverts Mobility to a plain spinoff — economically worse but not a moat break)
- Segment margin guide is cut (would falsify the "margin expansion still has room" assumption this pass is not yet challenging)
Upgrade conditions
- Price re-enters the $325-350 band with the thesis intact → ACCUMULATE
- Net debt/EBITDA visibly trending under 2.5x with two consecutive quarters of buyback resumption → removes the balance-sheet objection and would support conviction back toward 7.5-8.0 even above the entry band
What this pass did NOT test
- ROIC post-acquisition (row 8) — needs a full quarter of Boyd Thermal/Ultra PCS contribution cleanly separable from purchase-accounting noise. Explicitly UNTESTED, not assumed to still hold.
- Institutional ownership flow (row 17) — level confirmed (88.1%), quarter-over-quarter direction not re-pulled.
- GAAP FY2026 EPS guide — this report used the adjusted EPS guide ($13.40-$13.60) throughout, consistent with how the company and consensus frame it, but did not independently source an updated GAAP EPS guide figure to quantify the full GAAP-adjusted gap in dollar terms. Single-source risk: the GAAP decline is confirmed from Yahoo's quarterly statements (primary-adjacent, SEC-filed data) and is not itself in question, but the forward GAAP gap is not sized.
- Boyd Thermal / Ultra PCS standalone financials (margin, growth rate pre-acquisition) — not independently pulled; this report relies on Eaton's own strategic framing and the purchase price multiple (Boyd Thermal at $9.55B) rather than a bottom-up check of what was bought.
- Credit-rating detail beyond the headline A3 — did not pull the full Moody's/S&P rationale document, only the assigned rating and the outlook-relevant headline figures via search.
Sources: .mcp/fin.py ETN --news, yahoo-finance MCP (get_stock_info, quarterly income/cashflow/
balance sheet, get_holder_info insider_transactions), WebSearch (Eaton Q1/Q2 2026 earnings
releases and call transcripts, Dana/Eaton RMT announcement, Boyd Thermal/Ultra PCS acquisition
press releases, Moody's rating action), Knowledge/Playbook pitfalls cited inline.