Discovery › compare
Compare ADI vs NXPI vs MCHP
A portfolio-specific passage was removed from the public build.
0. Framing — this is not three of the same thing
All three are "broad-line analog/MCU cyclicals emerging from the 2024–25 inventory correction," and the market treats them as one trade. They are at radically different points on that curve, and that — not moat quality — is the dominant driver of relative value today.
| Prior peak revenue | TTM revenue | % of peak restored | Where in the cycle | |
|---|---|---|---|---|
| ADI | 12.31B (FY23) | 12.74B | 104% | Past the recovery. New highs, new growth vector. |
| NXPI | 13.28B (FY23) | 13.19B | 99% | Just closed the round trip. Fresh expansion. |
| MCHP | 8.44B (FY23) | 4.71B | 56% | Still deep in the hole. Recovering fast off a -57% trough. |
Read every valuation number below through that table. MCHP's cheap-looking forward multiple and ADI's expensive-looking trailing multiple are both artifacts of cycle position.
1. Side-by-Side Metrics
Market & pricing
| Metric | ADI | NXPI | MCHP |
|---|---|---|---|
| Price | $353.37 | $240.98 | $71.37 |
| Market cap | 172.1B | 60.8B | 38.8B |
| Enterprise value | 183.5B | 73.8B | 46.4B |
| 52wk range | 218.37 – 445.91 | 183.00 – 339.95 | 48.52 – 105.91 |
| % off 52wk high | -20.8% | -29.1% | -32.6% |
| P/E (ttm, GAAP) | 52.5 | 23.0 | 324.4 |
| P/E (fwd, non-GAAP consensus) | 23.5 | 13.3 | 17.0 |
| EV / TTM Revenue | 14.4x | 5.6x | 9.9x |
| EV / FCF (latest FY) | 42.9x | 32.4x | 53.3x |
| P/B | 5.10 | 5.57 | 6.01 |
| Beta | 1.19 | 1.80 | 1.73 |
| Analyst target (mean) | 457.73 (+30%) | 314.10 (+30%) | 113.08 (+58%) |
⚠️ Data flags. (1) All three forward P/Es are consensus non-GAAP figures — they add back large acquisition-intangible amortization ($188M/q at ADI, $108M/q at MCHP) and are not comparable to the GAAP trailing figures in the row above. (2) Yahoo's Jun-2026 quarter for NXPI reports diluted shares of 2.54B and EPS of $0.302 — both off by exactly 10x. Corrected to 254M shares / $3.03 EPS throughout this document. (3)
fin.pysnapshot yoy revenue growth for NXPI (12%) is stale; the actual Jun'26 vs Jun'25 figure is +19.5%.
Recent revenue trajectory (quarterly, $M)
| Quarter (approx. calendar) | ADI | NXPI | MCHP |
|---|---|---|---|
| Q1 (Mar/Apr '25) | 2,640 | — | 971 |
| Q2 (Jun/Jul '25) | 2,880 | 2,926 | 1,076 |
| Q3 (Sep/Oct '25) | 3,076 | 3,173 | 1,140 |
| Q4 (Dec/Jan '26) | 3,160 | 3,335 | 1,186 |
| Q5 (Mar/Apr '26) | 3,623 | 3,181 | 1,311 |
| Q6 (Jun '26) | — | 3,496 | — |
| Latest yoy | +37.2% | +19.5% | +35.1% |
| Sequential trend | accelerating | choppy but up | 4 straight up quarters |
ADI's most recent quarter is the standout: +14.7% sequential, which is not normal for a $3.6B/quarter analog business and points to genuine AI-datacenter/power content on top of the industrial restock. NXPI's Mar'26 dip (-4.6% q/q) is the auto-inventory wobble; it resumed +9.9% in Jun'26.
Margin structure
| Metric (latest quarter) | ADI | NXPI | MCHP |
|---|---|---|---|
| Gross margin | 67.3% | 57.3% | 61.0% |
| Gross margin, year ago | 61.0% | 53.4% | 51.6% |
| GM expansion yoy | +630bp | +390bp | +940bp |
| Operating margin (GAAP) | 38.1% | 30.6% | 17.1% |
| R&D as % revenue | 14.1% | 17.3% | 22.4% |
MCHP's gross-margin snap-back is the most violent (+940bp) because it is pure fab-utilization leverage off a starved base — that's mechanical, not structural, and it's the strongest part of the MCHP bull case. But its operating margin at 17.1% is still less than half ADI's, weighed down by $108M/quarter of amortization and an R&D line it cannot shrink on a revenue base 44% below peak.
Health & balance sheet
| Metric | ADI | NXPI | MCHP |
|---|---|---|---|
| Total debt | 8.66B | 12.22B | 5.54B |
| Cash | 2.50B | 2.91B | 240M |
| Net debt | 6.16B | 9.31B | 5.30B |
| Debt / Assets | 18.1% | 46.0% | 38.5% |
| Debt / Equity | 25.8% | 104.0% | 87.7% |
| Current ratio | 1.75 | 2.24 | 2.09 |
| Interest expense (annual) | 318M | 466M | 215M |
| Interest coverage (TTM op inc) | 13.0x | 7.7x | 2.5x |
| Goodwill | 26.95B | 10.30B | 6.70B |
| Goodwill as % of equity | 80% | 102% | 104% |
| Tangible book value | ~$0 to slightly negative | negative | negative |
Every one of these is a serial acquirer, so book value is mostly purchase accounting. That single fact invalidates naive Graham/P-B comparisons across the trio (§4).
MCHP's 2.5x interest coverage and $240M cash balance are the hard constraint in this comparison. That is not a bankruptcy risk — it's a $38B company with revolver access — but it is a company with no slack, which is why the buyback is at zero and the dividend is being funded from the balance sheet.
Cash flow & capital allocation (latest fiscal year)
| Metric | ADI (FY Oct'25) | NXPI (FY Dec'25) | MCHP (FY Mar'26) |
|---|---|---|---|
| Operating cash flow | 4.81B | 2.82B | 962M |
| Capex | -534M | -537M | -91M |
| Free cash flow | 4.28B | 2.28B | 871M |
| FCF margin | 38.8% | 18.6% | 18.5% |
| Buybacks | -2.16B | -899M | 0 |
| Dividends (common) | -1.92B | -1.02B | -984M |
| Preferred dividends | — | — | -111M |
| Acquisitions | -46M | -1.18B | ongoing (Hailo) |
| Net debt change | +0.99B | +1.36B | -0.12B |
| Cash returned as % of FCF | 95% | 84% | 126% ⚠️ |
Growth trends (3yr CAGR, peak-to-trough distorted — read with §0)
| Metric | ADI | NXPI | MCHP |
|---|---|---|---|
| Revenue CAGR | -2.8% | -2.4% | -17.6% |
| Net income CAGR | -6.2% | -10.2% | -53.2% |
| FCF CAGR | +4.3% | -5.1% | -34.7% |
| OCF CAGR | +2.4% | -10.2% | -35.7% |
| Diluted share CAGR | -1.7% | -1.2% | -0.7% |
⚠️ Framework note (§1 of
analysis_notes.mdcalls for 5–8yr FCF CAGR).fin.pyonly returns ~4 annual periods, so these are 3yr CAGRs measured from the FY2022/23 cycle peak to the FY2025/26 trough-to-recovery — structurally negative for all three and not a durability read. ADI holding FCF CAGR positive through that window is the meaningful signal here; MCHP's -34.7% is the meaningful warning. A roic.ai pull would be needed for a true 8yr series.
Returns on capital
| Metric | ADI | NXPI | MCHP |
|---|---|---|---|
| ROE (reported) | 10% | 26% | 3% |
| ROA | 5% | 8% | 2% |
| ROIC (TTM NOPAT / invested capital) | 8.6% | 13.1% | 3.9% |
| ROIC ex-goodwill | 23.6% | 24.3% | 8.9% |
ADI and NXPI converge to ~24% on the operating business once you strip the goodwill from Linear/Maxim and Freescale respectively — both are genuinely excellent businesses that overpaid for scale. MCHP does not converge: 8.9% ex-goodwill is below its own cost of capital, and that is a trough-earnings artifact that only resolves if revenue keeps climbing.
Per-share metrics (owner's view)
| Metric | ADI | NXPI | MCHP |
|---|---|---|---|
| Revenue/share (latest FY) | 22.19 | 48.24 | 8.64 |
| FCF/share (latest FY) | 8.61 | 8.98 | 1.60 |
| Price / FCF-per-share | 41.0x | 26.8x | 44.6x |
| Book value/share | 69.27 | 43.25 | 11.87 |
| Shares out, 3yr change | -3.9% | -2.5% | -0.6% |
| Dilution overhang | none | none | mandatory convertible preferred ⚠️ |
MCHP's $111M/yr preferred dividend (from the ~$1.5B March-2025 mandatory convertible raise used to pay down debt) converts to common. Reported diluted EPS of $0.22 is already net of the preferred dividend, but the share-count dilution has not yet landed. Any per-share upside case for MCHP must be haircut for it.
Dividend profile
| Metric | ADI | NXPI | MCHP |
|---|---|---|---|
| Yield | 1.20% | 1.52% | 2.34% |
| 5yr avg yield | 1.65% | 1.69% | 2.00% |
| Yield vs own history | -27% (rich) | -10% (near fair) | +17% (cheap-looking) |
| Payout ratio (earnings) | 61% | 39% | 827% ⚠️ |
| Payout ratio (FCF) | 45% | 45% | 113% ⚠️ |
| Buyback support | yes, 2.16B | yes, 899M | suspended |
| Dividend safety | Strong | Strong | Strained |
2. Moat Comparison
ADI — the strongest moat, and the least contested
ADI sells the highest-precision analog and mixed-signal parts in the industry: converters, amplifiers, power management, RF. The moat is a stack of three reinforcing sources.
Intangibles / accumulated design IP. High-performance analog is the one corner of semiconductors where Moore's Law doesn't help you. Performance comes from decades of proprietary process recipes, layout craft, and hand-tuned circuit design that lives in the heads of senior engineers and can't be compiled from a spec. You can license an ARM core; you cannot license a 24-bit sigma-delta converter's noise floor.
Switching costs. ~75,000 SKUs across 100,000+ customers, most of them in industrial and automotive systems with 10–20 year production lives. Roughly half of ADI's revenue comes from products introduced more than a decade ago. A customer requalifying an analog front-end re-runs the whole system validation — the part might be $4, the switch costs six figures and nine months.
Efficient scale in fragmentation. The typical ADI part is a small dollar value in a low-volume socket. Nobody can profitably attack 75,000 individual sockets one at a time, which is exactly why the catalog itself is the defense.
Adversarial stress-test. A well-funded rival — say a Chinese national champion with state capital — can and does attack ADI's low end: SG Micro and 3Peak are winning commodity op-amp and power sockets in China on price. They cannot replicate the precision catalog on any timeline that matters, because the constraint is analog design talent, not capital. The realistic 10-year outcome is that ADI cedes the bottom of its range and holds the top at higher margins — which is what the 67.3% gross margin (up 630bp yoy, at a new record) is already telling us.
Where the moat is weak, and what management is doing. The moat was substantially bought: $27B of goodwill from Linear and Maxim is why reported ROIC is 8.6% against a 23.6% operating reality. Management has stopped acquiring ($46M last year) and is returning 95% of FCF instead — the correct response, and the reason ADI's per-share metrics are the best in the trio.
New growth vector. The +14.7% sequential quarter is not industrial restock alone. ADI's power-management and high-speed signal-chain content in AI datacenter racks and optical interconnect is a genuinely new demand pool for a company previously levered to factory automation and cars. This is the most important thing that has changed about ADI in five years, and it's the honest justification for a premium multiple.
Evergreen verdict: yes. Precision analog is as close to a forever business as semiconductors offers.
NXPI — best returns on capital, sharpest single risk
NXP is ~57% automotive: radar, in-vehicle networking, battery management, the S32 processor platform with its software stack. The rest is industrial/IoT, secure mobile (NFC and secure elements, where NXP has something close to a monopoly and a genuine regulatory/certification moat in eGov ID and payment credentials), and comms infrastructure.
Moat sources. Automotive qualification (AEC-Q100, functional safety, 7–10 year design-win cycles with a Tier-1 or OEM) is a real barrier — being designed into a vehicle platform means revenue visibility through the model's entire production run. Secure identification is stronger still: certification regimes and government relationships that took twenty years to build. NXP earns the best ROIC of the three (13.1% reported, 24.3% ex-goodwill) which is the quantitative confirmation.
Adversarial stress-test — and this is the one that actually lands. The threat to NXP is not a Western rival; it's Chinese automotive localization. Chinese OEMs are now the largest auto market in the world, they are under explicit state direction to design out Western semiconductors, and unlike precision analog, an automotive MCU or a CAN transceiver is replicable by a competent domestic vendor with a decade of runway. Auto qualification is a barrier measured in years, not in impossibility. NXP's 57% auto concentration means this is not a diversifiable risk for them — it is the thesis.
Capital allocation flag. NXP spent $1.18B on acquisitions last year (Kinara for edge AI, TTTech Auto, Aviva Links) while adding $1.36B of net debt, on a balance sheet already at 46% debt/assets with negative tangible book. This is empire-building at the exact moment the cycle turned in their favor, and it's the reason FCF CAGR is negative while ADI's is positive. Watch it.
Evergreen verdict: qualified yes. The secure-ID and NFC franchises are evergreen. The automotive franchise is durable in the West and structurally contested in China.
MCHP — the moat that the correction partly disproved
Microchip's historical pitch was excellent: the broadest MCU + analog + FPGA + connectivity catalog, ~120,000 customers, a "total system solution" cross-sell that raised revenue per design win, extreme stickiness in 8- and 16-bit MCUs with 20-year product lives, and best-in-class 68% gross margins with monstrous FCF conversion. On paper it looked like ADI's moat at a lower price point.
What the 2024–25 correction revealed. MCHP's revenue fell 57% peak-to-trough — roughly triple the decline at ADI or NXPI in the same downturn, facing the same end markets. That differential is the single most important moat datapoint in this comparison. A moat built on switching costs should produce shallower cyclical troughs than peers, not deeper ones. The most plausible explanation is that a meaningful slice of peak-cycle revenue was distributor channel inventory rather than end demand, and that MCHP's direct-ship visibility into what customers were actually consuming was worse than management believed. The moat wasn't fake, but it was flattered.
Adversarial stress-test — the most attackable of the three. The mid-range MCU business sits exactly where the barriers are lowest. ARM Cortex-M cores are licensable off the shelf; RISC-V is free. Chinese vendors (GigaDevice, Sinowealth) already ship adequate parts with adequate toolchains at aggressive prices, and the legacy 8-bit franchise — MCHP's stickiest revenue — is the natural displacement target as 32-bit part costs fall below it. MCHP's real defense is its development-tool ecosystem and the inertia of 120,000 engineering teams, which is genuine but is a decaying asset rather than a compounding one.
Moat defense, and credit where due. Steve Sanghi's return as founder-CEO is the most credible thing about this story — he built the original operating model and he moved fast: Fab 2 closure, inventory purged from 251 days toward normal, opex cut, buyback suspended to protect the balance sheet. Four consecutive sequential revenue increases and +940bp of gross margin expansion say the operational fix is working. The Hailo acquisition (announced 2026-07-28) is a sensible strategic add in edge AI. But it is also cash going out the door at a company paying a dividend it doesn't cover, and that ordering of priorities is the tell.
Evergreen verdict: weakest of the three. The business survives; the question is at what margin and what share.
Moat ranking
ADI > NXPI > MCHP — and the gaps are not small. ADI's advantage is the least contestable and is currently widening (record gross margin, new AI demand pool). NXPI has the best returns on capital but carries the single largest structural risk (China auto). MCHP's moat is real but was overestimated by the market, by management, and probably by the moat itself.
3. Cycle-Position Normalization
Because trailing multiples are meaningless at these cycle positions, here is what each business earns at a defensible mid-cycle, and what you pay for it.
| ADI | NXPI | MCHP | |
|---|---|---|---|
| TTM revenue | 12.74B | 13.19B | 4.71B |
| TTM GAAP net income | 3.31B | 2.98B (2.48B ex one-off gain) | 230M |
| TTM GAAP EPS | $6.73 | $11.74 ($9.78 normalized) | $0.22 |
| Latest-quarter run-rate EPS (annualized) | $9.60 | $12.12 | ~$0.84 |
| Mid-cycle normalized EPS estimate (GAAP) | $10–11 | $12–13 | $3.30–3.60 at full return to prior peak revenue |
| P/E on normalized GAAP EPS | 33–35x | 19–20x | 20–22x — and only if it fully recovers |
| Consensus non-GAAP fwd EPS (implied) | ~$15.06 | ~$18.19 | ~$4.20 |
| Distance from current run-rate to consensus | +57% | +50% | +400% ⚠️ |
The MCHP insight this exposes. At $71.37, MCHP trades at ~21x the EPS it would earn if revenue returned all the way to the FY2023 peak of $8.44B at healthy 65%/33% margins — a journey requiring revenue to grow another 79% from here, and then haircut again for preferred conversion. You are not buying a cheap recovery. You are paying a normal multiple for a recovery that has not happened, and financing an uncovered dividend while you wait. The consensus $4.20 forward number requires roughly quintupling the current quarterly non-GAAP run-rate; treat it as an analyst placeholder, not a forecast.
The NXPI insight. NXPI is the only one of the three where normalized earnings power is already substantially in the reported numbers and the multiple is below 20x. It has closed the round trip to prior peak revenue with higher operating margin than at that peak (30.6% vs 28.5%), meaning this is real operating improvement, not just volume.
4. Valuation Comparison
Applying analysis_notes.md §3 conditionally, with weights stated.
Graham's Intrinsic Value — √(22.5 × EPS × BVPS)
| EPS used | BVPS | Graham IV | Price | Price / Graham | |
|---|---|---|---|---|---|
| ADI | 6.73 (ttm) | 69.27 | $102.42 | 353.37 | 3.45x |
| NXPI | 11.74 (ttm, adj.) | 39.79 | $102.53 | 240.98 | 2.35x |
| MCHP | 0.22 (ttm) | 11.87 | $7.66 | 71.37 | 9.32x |
Weight: LOW (10%). Graham fails all three on the same grounds — book value is 80–104% goodwill from serial M&A, so BVPS overstates real asset backing; and trailing EPS is cycle-distorted (catastrophically so for MCHP). The ranking still carries information: NXPI cheapest, MCHP an extreme outlier at 9.3x. Do not read the absolute levels as fair value.
Bogle Expected Return — dividend yield + earnings growth ± multiple change
| Div yield | Mid-cycle earnings growth | Multiple change (5yr) | Expected annual return | |
|---|---|---|---|---|
| ADI | 1.2% | +10 to 12% | 35x → 28x ≈ -4.4%/yr | ~7 to 9% |
| NXPI | 1.5% | +8 to 10% | 19x → 19x ≈ 0% | ~10 to 12% |
| MCHP | 2.3% | +15 to 20% off trough | 21x → 18x ≈ -3%/yr | ~-5% to +14% (wide) |
Weight: HIGH (40%). This is the model that discriminates properly here. ADI's problem is not the business, it's that a good business at 33–35x normalized earnings has its return eaten by multiple compression. NXPI's advantage is that it needs no multiple expansion to deliver a double-digit return. MCHP's range is so wide it isn't a forecast, it's an admission of variance — and the distribution is skewed by an uncovered dividend and a 2.5x interest coverage ratio, which means the left tail is fatter than the arithmetic suggests.
Dividend Yield Theory
| Current yield | 5yr avg yield | Signal | Valid? | |
|---|---|---|---|---|
| ADI | 1.20% | 1.65% | 27% below avg → overvalued | ✅ valid |
| NXPI | 1.52% | 1.69% | 10% below avg → mildly rich | ✅ valid |
| MCHP | 2.34% | 2.00% | 17% above avg → looks cheap | ❌ trap |
Weight: MODERATE for ADI/NXPI (25%), ZERO for MCHP. DYT requires a dividend you can trust — that's the entire premise. MCHP's yield is elevated at 113% of FCF payout with a suspended buyback and $240M of cash. That is the textbook case DYT was never designed for: the yield is high because the risk is high, not because the price is wrong. ADI's -27% reading is the cleanest valuation signal in this entire report: by its own dividend history, ADI is expensive.
Dividend Discount Model
| Current div/share | Assumed g | r = 9% | DDM value | |
|---|---|---|---|---|
| ADI | ~$4.24 | 7% | ~$227 | |
| NXPI | ~$3.67 | 6% | ~$130 | |
| MCHP | ~$1.82 | 1% (token raises) | ~$23 |
Weight: LOW (5%). DDM structurally undervalues semis because dividends are a minority of FCF (45% at ADI and NXPI). Directionally consistent with everything else: all three trade well above dividend-only value, MCHP most extremely.
Cash-flow based cross-check
| EV / TTM FCF (est.) | FCF yield on EV | |
|---|---|---|
| ADI | ~34.6x | 2.9% |
| NXPI | ~28x | 3.6% |
| MCHP | 53.3x | 1.9% |
Synthesized fair-value ranges
| Fair value range | Current price | Assessment | |
|---|---|---|---|
| ADI | $255 – $310 | $353.37 | 14–28% overvalued. Great business, bad entry. |
| NXPI | $225 – $290 | $240.98 | Fair to modestly cheap. In the lower half of range. |
| MCHP | $45 – $65 | $71.37 | 10–37% overvalued on any normalized basis. |
Assumptions stated: ADI at 26–30x normalized GAAP EPS of $10–11, reflecting a real premium for the best moat plus the AI vector but not the current 35x. NXPI at 18–23x normalized $12.50, discounted for China-auto structural risk and the leveraged balance sheet. MCHP at 15–18x a probability-weighted recovery EPS of ~$3.00 — not the full-recovery $3.60 — with a further haircut for preferred conversion and thin interest coverage.
5. Winner by Dimension
| Dimension | Winner | Runner-up | Loser | Note |
|---|---|---|---|---|
| Revenue momentum | ADI (+37.2%, +14.7% q/q) | MCHP (+35.1%) | NXPI (+19.5%) | ADI's is off a higher base |
| Recovery upside remaining | MCHP (56% of peak) | NXPI | ADI | The only MCHP win |
| Gross margin | ADI (67.3%) | MCHP (61.0%) | NXPI (57.3%) | ADI at a record |
| Operating margin | ADI (38.1%) | NXPI (30.6%) | MCHP (17.1%) | |
| Balance sheet | ADI (18% D/A, 13x cover) | NXPI (46% D/A, 7.7x) | MCHP (2.5x cover, $240M cash) | Not close |
| ROIC | NXPI (13.1% / 24.3% ex-GW) | ADI (8.6% / 23.6%) | MCHP (3.9% / 8.9%) | Top two converge ex-goodwill |
| FCF margin & quality | ADI (38.8%) | NXPI (18.6%) | MCHP (18.5%) | |
| Capital allocation | ADI (95% returned, M&A stopped) | MCHP (deleveraging, forced) | NXPI ($1.18B M&A on levered B/S) | |
| Shareholder dilution | ADI (-3.9% 3yr) | NXPI (-2.5%) | MCHP (preferred overhang) | |
| Dividend safety | ADI (45% of FCF) | NXPI (45% of FCF) | MCHP (113% ⚠️) | ADI wins on coverage + growth |
| Moat durability | ADI | NXPI | MCHP | Gaps are wide |
| Concentration risk | ADI (industrial + auto + AI) | MCHP (broad MCU) | NXPI (57% auto) | |
| Absolute valuation | NXPI (5.6x EV/S, 19x norm.) | ADI | MCHP | |
| Risk-adjusted expected return | NXPI (~10–12%) | ADI (~7–9%) | MCHP (wide, left-skewed) | |
| Downside protection | ADI | NXPI | MCHP | |
| Upside optionality | MCHP | ADI (AI vector) | NXPI | If MCHP executes fully |
Tally: ADI 9 · NXPI 4 · MCHP 2
Note that ADI wins on almost every quality dimension and NXPI wins on the two that determine your return from today's price. That is the tension this comparison resolves.
6. Verdict
Overall recommendation: NXPI — best risk-adjusted entry. Conviction 7/10.
This is not a claim that NXPI is the best business. ADI is, clearly and by a wide margin. It's a claim about §0 of the framework: it's not just how good the company is — it's how well the market has priced it.
Why NXPI wins. It has genuinely completed its cycle recovery at a higher operating margin than its prior peak (30.6% vs 28.5%), earns the best returns on capital in the trio (24.3% ex-goodwill), and trades at 19–20x normalized GAAP earnings and 5.6x EV/revenue — roughly a 60% discount to ADI on sales and 40% on normalized earnings. Its dividend is covered at 45% of FCF. Bogle math delivers 10–12% annually without requiring any multiple expansion, which is the definition of a margin of safety in an expected-return framework. It sits 29% below its 52-week high with fundamentals inflecting up.
Named risks on NXPI (these are real, not boilerplate): 1. China automotive localization — the sharpest structural risk in this comparison. 57% auto concentration against an explicit state policy of designing out Western semis in the world's largest auto market. Auto qualification is a multi-year barrier, not a permanent one. This is the reason NXPI is cheap, and it may deserve to be. 2. Leverage — 46% debt/assets, negative tangible book, $12.2B gross debt. Fine in an upcycle, constraining in the next downturn. 3. M&A discipline — $1.18B spent while adding $1.36B net debt, right as the cycle turned. If this becomes a pattern, the ROIC advantage erodes and the thesis weakens. 4. Beta 1.80 — this will draw down hard in any semi selloff.
Suggested NXPI zones: starter position at current $241; add zone $200–225; trim above $320. Break trigger: two consecutive quarters of China auto revenue decline attributable to design-out rather than cycle, or a debt-funded acquisition above $2B.
ADI — the best business here. Wrong price. Watch, don't buy. Conviction 8/10 on business, 4/10 on entry.
ADI is the highest-quality company in this comparison on essentially every measure that matters for a long-term hold: widest and least contestable moat, record 67.3% gross margin, 38.8% FCF margin, 18% debt/assets, 13x interest coverage, 95% of FCF returned to shareholders, M&A discipline restored, and a genuinely new AI-datacenter demand vector that justifies rerating the growth outlook upward.
The problem is arithmetic. At 33–35x normalized GAAP earnings and 14.4x EV/sales, and with its dividend yield 27% below its own five-year average, you are paying full price for a company that is already past its cyclical recovery. Bogle gets you 7–9% annually and most of that is eaten by the multiple compression the model assumes. Buying the best business in a sector at the top of its own valuation range is the "Great + Expensive → Wait/watch" cell of the framework table, and it is exactly the discipline the notes call for: rather miss an opportunity than overpay.
Suggested ADI zones: watch below $300 (yield ~1.4%, approaching its historical average); accumulate $255–290; back up the truck below $240. The AI-content story means you may not get that entry — accept that. The 52-week low was $218.
MCHP — Pass. Not a value trap yet, but priced as if the turnaround already worked. Conviction 3/10.
The operational turnaround is real and deserves acknowledgment: four consecutive quarters of sequential revenue growth, +940bp of gross-margin expansion, inventory purged, Fab 2 consolidated, founder-CEO back and executing. If you believe MCHP returns to $8.4B revenue at 33% operating margins, there's roughly $3.30–3.60 of EPS in there.
Three things stop this from being a buy:
- You are already paying for the recovery. At $71.37, that's ~21x the EPS MCHP earns only after growing revenue another 79% and only at restored peak margins. The reward for being right is a normal multiple. The consensus $4.20 forward EPS requires roughly quintupling the current non-GAAP run rate — treat it as a placeholder.
- The dividend is not covered. $984M common + $111M preferred against $871M of FCF is a 126% payout, funded from a balance sheet down to $240M of cash with the buyback at zero and interest covered only 2.5x. Management is prioritizing an uncovered dividend and an acquisition (Hailo, 2026-07-28) over balance-sheet repair. That is a governance signal, not just a cash-flow one.
- The moat underperformed its own stress test. A -57% peak-to-trough revenue decline — roughly triple ADI's or NXPI's into the same end markets — is evidence that the switching-cost moat was partly channel inventory. And the mid-range MCU franchise is the most exposed position in this trio to licensable ARM cores, free RISC-V, and Chinese price competition.
Revisit triggers (both required): (a) dividend covered by FCF — needs roughly $1.5B+ annual OCF, i.e. quarterly revenue near $1.5B, versus $1.31B today; and (b) price below $55. Meeting only the second condition without the first is the value trap.
Tension surfaced honestly
The genuine disagreement in this analysis is ADI vs NXPI, and it is close. If you believe the AI-datacenter content vector is a durable multi-year secular addition to ADI's TAM — and the +14.7% sequential quarter is real evidence for that, not noise — then ADI's 33–35x is a growth multiple rather than a cyclical-peak multiple, and paying up for the best moat in the sector is defensible. The bear case on NXPI is symmetrically real: China auto localization could permanently impair 20–30% of its business, in which case its "cheapness" is accurate pricing of a structurally shrinking franchise and the 5.6x EV/sales is a value trap of a subtler kind than MCHP's.
The recommendation of NXPI over ADI rests on a valuation gap wide enough (roughly 40% on normalized earnings) to absorb being partly wrong about China. It does not rest on NXPI being the better company. It isn't.
The MCHP verdict is not close — all three analytical lenses (fundamentals, moat, valuation) agree, which is itself worth noting.
7. Watchlist Recommendation
NXPI stands out. Suggested Watchlist.md entry:
- Sleeve: ⚡ AI & Infrastructure Capex — or 🔧 Re-Rating Plays. Re-Rating is the better fit: the thesis is a cheap multiple on restored-and-improving earnings power, not AI capex exposure (NXP's AI content is thin).
- Conviction: [7]
- Entry zone: $200–225 · Currently: $240.98 (👀 On Deck, just above zone)
- Trim zone: >$320
- Thesis: Auto/industrial analog-MCU leader back to prior-peak revenue at higher margins (30.6% vs 28.5%), best ROIC in its peer group (24.3% ex-goodwill), at 19–20x normalized earnings and 5.6x EV/sales — a 40–60% discount to ADI on quality that is maybe 15% lower.
- Break trigger: Two consecutive quarters of China auto revenue decline attributable to design-out rather than cycle; or debt-funded M&A above $2B; or debt/assets above 52%.
- Concentration note: 57% automotive. Would correlate with any existing auto, industrial-cyclical, or China-exposure holdings.
ADI also merits a watchlist slot — as a 🏛 Evergreen Compounder with conviction [8] and an explicit out-of-zone flag. It's the best business in the sector and the entry is 14–28% too high; that's precisely what a watchlist is for.
MCHP: do not add. Place in 🪦 Graveyard as 🌱Revisit with the dual re-entry trigger above (FCF-covered dividend AND sub-$55).
8. Data Quality & Limitations
| Issue | Impact | Mitigation used |
|---|---|---|
fin.py returns only ~4 annual periods |
Framework §1 calls for 5–8yr FCF CAGR; 3yr CAGRs here span peak→trough and are structurally negative for all three | Flagged explicitly; supplemented with quarterly trajectory. A roic.ai pull would materially improve the durability read |
| Yahoo NXPI Jun'26 quarter: shares 2.54B, EPS $0.302 | Both off by exactly 10x | Corrected to 254M / $3.03 throughout |
| Yahoo NXPI snapshot yoy revenue 12% | Stale | Recomputed from quarterlies: +19.5% |
| Forward P/Es are non-GAAP consensus | Not comparable to GAAP trailing; MCHP's implies +400% from run-rate | Flagged; normalized §3 table built instead |
| NXPI Mar'26 includes +$621M other income | Inflates TTM net income by ~$500M after tax | Normalized TTM NI to $2.48B and EPS to $9.78 |
| ROIC computed by hand (NOPAT / invested capital) | May differ from roic.ai's methodology | Both reported and ex-goodwill figures shown |
| Mid-cycle EPS estimates are judgment calls | Drives the entire valuation section | Assumptions stated inline; ranges not point estimates |
| No earnings-call transcripts or insider-activity data pulled | Management credibility and insider signal assessed only from public news | Noted as a gap — a /analyze NXPI would close it |
Generated by Financebot · framework: analysis_notes.md §0–§5 · tools: .mcp/fin.py, Yahoo Finance MCP (quarterly statements)