Skyworks Solutions

Phase A — Understand the business

Company Overview

Skyworks designs analog and mixed-signal semiconductors — its core franchise is RF front-end (RFFE) content for smartphones: power amplifiers, filters (SAW, TC-SAW, BAW), switches, tuners, and the integrated front-end modules that sit between a phone's transceiver and its antenna. It sells into smartphones, tablets, wearables, plus a "Broad Markets" portfolio spanning Wi-Fi/connectivity, automotive (power isolation for OBCs/BMS, telematics, infotainment), data center optical/timing, infrastructure, aerospace/defense, industrial and medical.

The business model in plain terms: Skyworks wins multi-year "design sockets" inside a customer's flagship product, ships highly-customized modules at high volume for that product's life, then re-competes at the next generation. Revenue is not recurring or take-or-pay — it is socket-by-socket, and a lost socket disappears fast. Roughly 6,900 customers and 4,900 products, but that breadth is a mirage at the top: Apple was 67% of net revenue in FY2025 (69% FY2024, 66% FY2023). The top three accounts receivable balances = 82% of gross AR. This is, functionally, an Apple RF supplier with a diversification project attached.

Stated competitors: Analog Devices, Broadcom, Cirrus Logic, Murata, NXP, Qorvo, Qualcomm, Texas Instruments. Of these, Qorvo and Broadcom are the ones that actually take Skyworks' sockets.

Supply Chain

Skyworks runs a hybrid IDM model — it owns front-end wafer fabs and assembly/test, unusual for a fabless-dominated industry, which is why it is so capex-light relative to revenue yet carries a real fixed-cost base.

Upstream → Skyworks → end customer, named:

• Owned fabs / process: GaAs and filter (SAW/TC-SAW/BAW) fabrication. PP&E by location: Japan $461.0M (acquired Panasonic filter operations — the BAW/TC-SAW base), Mexico $264.0M (Mexicali — assembly/test; 54% of the workforce is in Mexico ), United States $226.8M (Newbury Park, CA front-end fab + Irvine HQ), Singapore $215.0M (assembly/test/distribution). Total PP&E net $1,194.6M.
• Foundry / outsourced wafers: for digital and some specialty content Skyworks uses external foundries (industry: TSMC and others) — not named in the filing, label.
• Substrates / packaging materials, laminates, gold, GaAs/silicon-on-X wafers — commodity-ish inputs; the 10-KForm 10-KA company’s audited annual report to the US regulator. The most complete thing it publishes. flags raw-material and supplier availability as a risk.
• Channel: sells indirectly through distributors ($3,525.6M of FY2025 revenue) and direct to OEMs/CMs/ODMs ($561.3M). The Apple relationship runs through "multiple distributors, contract manufacturers, and direct sales."
• End customers (named): Apple (the whale), plus Amazon, Bose, Ciena, Cisco, Ericsson, Garmin, Google, Honeywell, Lenovo, LG, Microsoft, Nokia, Northrop Grumman, OPPO, Samsung, Sonos, Sony, Tesla, TP-Link, VIVO, Xiaomi.

Chokepoint: the BAW filter base in Japan + the Newbury Park front-end fab are the single-source process nodes; a disruption there hits the flagship-phone modules directly. The deeper chokepoint is commercial, not physical — one customer's socket decisions move the whole P&L.

Competitive Advantages (moats)

Real but narrow and eroding:

• Filter IP + process (genuine moat): high-performance BAW/TC-SAW filtering is hard; the analog/RF design + in-house filter fab is a years-deep capability. This is the durable piece and the strategic logic of buying Qorvo.
• Scale in RFFE modules: integration of PA + filter + switch + tuner into a single small module is a packaging/co-design advantage at flagship volumes.
• Switching costs (real within a generation, weak across one): once designed into a phone platform the part ships for that product's life. But the RFFE socket re-competes every generation, and FY2025 proved the switching cost does not survive a redesign — Apple moved content to a rival.
• Bargaining power — weak over its biggest customer, strong over the long tail. With Apple at 67%, Skyworks is the price-taker; Apple dual-sources deliberately and has been engineering Skyworks content out (Broadcom in). Over its 6,900 smaller customers Skyworks has normal supplier power.

Versus rivals: Broadcom (taking Apple RF content) and Qualcomm are larger, more diversified, and can bundle RFFE with baseband/connectivity — a structural disadvantage for a pure-play. Qorvo is the symmetric pure-play rival, which is exactly why combining removes the most dangerous competitor and adds Qorvo's ultra-high-band PA / envelope-tracking strengths to Skyworks' low-band/diversity-module strengths. Ground note: the hardware commercial layer (kb/hardware/wiki/positioning.md) is AI-compute-centric (NVIDIA/AMD/Broadcom ASICs) and carries no RFFE matrix — moat read here is filing + web.

Segments (end-market, single reportable segment)

Skyworks adopted ASU 2023-07 in Q4 FY2025 and confirms one reportable operating segment (CODM = CEO, manages at consolidated level). The meaningful cut is the management end-market split disclosed on calls:

By end market (Q2 FY2026, quarter ended ~Apr 2026):

• Mobile ≈ 58% of revenue — "higher than expectations, driven by healthy sell-through at the top customer." This is the Apple-dominated, structurally-pressured book.
• Broad Markets ≈ 42% of revenue, +10% YoY — Wi-Fi (Wi-Fi 7), data center, automotive. This is the growth and the diversification thesis, and it is working at the margin.

By geography (FY2025, by OEM HQ): United States $3,157.1M (77%), Taiwan $259.1M, China $254.2M, South Korea $190.1M, EMEA $185.8M, Other APAC $40.6M. The US weighting is an artifact of Apple HQ, not US end-demand — the phones ship/assemble in Asia. EMEA jumped (≈$114.5M → $185.8M) on infrastructure/auto.

By channel (FY2025): Distributors $3,525.6M, Direct $561.3M.

Trend & cause: total revenue is decelerating then declining — FY2023 $4,772.4M → FY2024 $4,178.0M (−12.5%) → FY2025 $4,086.9M (−2.2%), with the company itself attributing the drop to "a decrease in market share at a significant customer, partially offset by … mobile and Wi-Fi". Mobile is shrinking; Broad Markets is the only thing growing.

---

Phase B — Measure performance

Earnings Result

The multi-year P&L (FY ends late Sept/early Oct):

GAAP EPS roughly halved in two years. The compression is double-barrelled: gross margin fell 300bps (lower-margin mix / under-absorption as Apple volume left) and R&D rose to 19.2% of sales ($606.8M → $785.5M) — Skyworks is spending into the downturn to rebuild content (iPhone 18, Android, Broad Markets), which depresses near-term margin by design.

Latest print — Q2 FY2026 (quarter ended ~2026-04-03):

• Revenue $943.7M, −1.0% YoY, but above the high end of guidance. H1 FY2026 revenue $1,979.1M (−2.1%).
• GAAP operating income $42.1M (4.5% margin) vs $97.3M (10.2%) a year ago; net income $35.6M, GAAP diluted EPS $0.24 (vs $0.43). Non-GAAP diluted EPS $1.15 — the $0.91 GAAP-to-non-GAAP gap is intangible amortization, SBC, and ~$19M of Qorvo transaction costs "].
• The margin optics are distorted by merger costs: GAAP R&D 22.5% and SG&A 12.7% of revenue in the quarter — SG&A is elevated by deal expense.
• Guidance — Q3 FY2026: revenue $900–950M (mid $925M), non-GAAP gross margin ~44.5–45.5%; mobile down slightly, broad markets up modestly.

Balance-sheet flags — none material; the balance sheet is a fortress. Cash + marketable securities $1,374M at FYE25 (cash $1,161.3M + securities $212.9M), rising to $1,413.3M cash by Q2 FY26. Total debt $995.8M — $500M 1.80% notes due 2026 (now classified current) + $500M 3.00% notes due 2031; net cash positive ≈ +$380M; $750M revolver undrawn; leverage covenant 3.0x with ample headroom. Operating cash flow $1,300.8M FY2025 (down from $1,824.7M / $1,856.4M) on just $195.0M capex → FCF ≈ $1,106M (~27% FCF margin). Even at trough earnings this is a cash machine.

Market reaction (the tape's verdict): the stock has de-rated hard — see Lens 8. The Q2 FY26 beat-and-raise-ish print still saw shares dip on weak GAAP margin optics, then the complex rallied in May 2026.

Earnings Calls (sentiment trend)

No transcripts on the shelf; this is -sourced from quarterly releases/call coverage.

• Tone arc: Q1 FY2025 (Feb 2025) was the capitulation call — management confirmed Apple content loss and the stock fell ~25%. Since the Brace CEO transition (Feb 2025), the message has shifted from defending Apple to "diversify and execute": Broad Markets growth, new product launches, and disciplined opex.
• Recurring phrases now: "broad markets," "Wi-Fi 7," "data center," "content diversification," "multi-generational design win," "operational execution," and — post-Oct 2025 — "the combination with Qorvo" / "scale" / "synergies."
• What they stopped saying: the bullish "content gains at our largest customer" framing that dominated 2021–2023. The narrative has explicitly pivoted away from Apple content growth toward replacement demand (Android) and Broad Markets.
• The headline they want owned: Q2 FY26 release led with "Secured Multi-Generational Android OEM Design Win with Expected $1 Billion+ Revenue Through 2030" — management is selling a concrete, quantified offset to the Apple erosion. Sentiment: guardedly improving, execution-credible, but still in proving-it territory.

Comps

Peer pull from _index.json (hardware) + the obvious RF/analog set.

Read: SWKS at ~15x forward non-GAAP EPS sits at a deep discount to the broad-analog comps TXN/ADI (~35x) and roughly in line with the other Apple-/handset-exposed, lower-growth names (QRVO ~23x, CRUS ~16x). The market is pricing Skyworks as a structurally-challenged, single-customer RF supplier, not as a diversified analog franchise. The valuation gap to TXN/ADI is the embedded bear case in the multiple — and the re-rate optionality if Broad Markets + Android + Qorvo turn the growth line.

Stock-Price Catalysts (>5% moves, last ~5y)

Mostly:

• Feb 5–6, 2025 — the defining −25% crash. Q1 FY2025 print disclosed the company would lose RF content at Apple to a competitor (Broadcom) starting with iPhone 17 — content share −20% to −25%; B. Riley pegged ~$525M annual revenue at risk. Stock to ~$65. Morningstar cut fair value $95 → $70 the same week. This single event re-rated the equity.
• CEO change (Feb 5, 2025): Liam Griffin out, Philip Brace in — announced into the same window; read as the board acknowledging the strategic problem.
• 2025 grind to a 5-year low even despite a Q1 FY2026 (Dec-quarter) revenue beat ($1.035B, Wi-Fi 7) — sentiment stayed sour because Apple was still shrinking.
• Oct 27, 2025 — Qorvo merger announced. Cash-and-stock, ~$22B combined EVEnterprise valueWhat it would cost to buy the whole company: its market value plus its debt, minus the cash you would get with it. Often very different from market cap.; the strategic pivot from "lost Apple share" to "build the US RF scale leader."
• Feb 5, 2026 — FTC Second Request (extends antitrust review) — overhang event.
• May 5/2026 — Q2 FY26 beat + the $1B Android design-win headline; then ~May 22, 2026 the mobile-chip complex rallied (QCOM +12%, SWKS +9%, QRVO +7%). Apr 22, 2026 Barclays upgraded SWKS + QRVO to Overweight, PT $100.

What the tape tells you the market reacts to: (1) Apple content / socket news — the dominant driver, by a mile; (2) the Qorvo deal's regulatory progress; (3) Broad-Markets/Android traction as the offset. Macro and sector beta matter less than the idiosyncratic Apple variable.

---

Phase C — Judge people & books

Management

A near-complete C-suite refresh in 2025, all brought in to fix the Apple-dependence problem and execute the Qorvo deal:

• CEO — Philip Brace (since Feb 17, 2025). Semiconductor operator/turnaround pedigree: President & CEO of Sierra Wireless (2021–2023, sold to Semtech), executive chairman of Inseego (2024–), EVP at Veritas, President of Cloud Systems & Electronic Solutions at Seagate (2015–17); began at Intel and LSI. Currently on the boards of BlackBerry, Inseego, Lantronix. Granted a $30M performance award on appointment. Archetype: professional operator/value-rebuilder, not a founder — the right profile for a "diversify, cut what doesn't work, integrate an acquisition" mandate; the multi-board footprint is a mild attention-split flag.
• CFO — Philip Carter (since Sept 8, 2025) and SVP Sales & Marketing — Todd Lepinski (since June 2, 2025) — fresh finance and go-to-market leadership.
• Track record (the franchise, not the new team): under prior management Skyworks rode the 5G/Apple-content super-cycle to FY2023's $4.8B/$6.13 EPS, then let single-customer concentration become an existential risk. The new team inherits the problem.
• Capital allocation: historically shareholder-friendly to a fault — FY2025 returned $830.2M buybacks + $432.6M dividends = $1.26B against $1.30B operating cash; dividend $0.71/quarter (~$2.84/yr, 3.7% yield); $2B+ buyback authorized Feb 4, 2025. The 2021 Silicon Labs Infrastructure & Automotive acquisition ($2.75B, funded by the now-expired term loan) seeded Broad Markets — a reasonable diversification bet whose payoff is only now showing. Buybacks/dividends are now restricted by the Merger Agreement — capital return is paused/constrained while the deal is pending.
• Skin in the game: insider ownership is modest (professional-manager levels); our figures is not on the shelf — n/a.
• Red flags: none of the fraud/related-party variety. The honest flag is strategic: the prior team's failure to break Apple dependence; the new team is unproven here and is simultaneously running a turnaround and a $22B integration.

Forensic Red Flags

Acting as a forensic analyst on the FY2025 10-K + Q2 FY26 10-QForm 10-QThe quarterly version of the annual report. Lighter, and not audited. — the accounting is clean and conservative; the risks are economic, not forensic.

• Revenue recognition: point-in-time product sales through distributors/direct; standard, no long-term/percentage-of-completion games. Distributor channel (86% of revenue) carries the usual sell-in-vs-sell-through and price-protection/returns judgement — accrued customer liabilities $202.8M; not anomalous.
• Cash vs earnings — favorable. OCF $1,300.8M exceeds net income $477.1M; the gap is D&A + SBC, i.e. cash conversion is better than GAAP earnings, the opposite of a red flag. FCF ~$1.1B.
• Non-GAAP gap is large but legitimate — Q2 FY26 GAAP EPS $0.24 vs non-GAAP $1.15. Drivers: intangible amortization (acquired Silicon Labs/Panasonic assets), SBC, and ~$19M Qorvo transaction costs. Watch that SBC isn't flattering "non-GAAP" beyond peer norms, but the bridge items are real and disclosed.
• Goodwill/intangibles: Goodwill $2,176.7M (flat YoY — no impairment taken despite the revenue decline; reasonable given the cash flows) + intangibles net $809.0M, amortizing ~$184M/yr. A future Apple-driven revenue shock could pressure the impairment test — monitor.
• Receivables/inventory vs revenue: top-3 AR = 82% of gross AR — a concentration-of-credit-risk flag tied entirely to Apple, not a quality-of-earnings flag.
• Debt: plain-vanilla senior notes, in covenant compliance; the 2026 notes are reclassified current (refinancing/repay event) — routine.

Regulatory findings:

• SEC Litigation Releases: none naming Skyworks since 2021-06-30 (SEC EDGAR EFTS, LR).
• SEC AAERs: none (SEC EDGAR EFTS, AAER).
• 10-K Item 3 (Legal Proceedings): the company discloses ordinary-course litigation; the material disclosed legal item is the Qorvo merger-related stockholder litigation — two lawsuits filed in NY state court in January 2026 plus stockholder demand letters, alleging disclosure deficiencies in the joint proxy/prospectus and seeking more disclosure / injunctive relief / damages. These are boilerplate merger-objection suits, routinely mooted by supplemental disclosures — low materiality.
• Non-SEC enforcement (web): no material FTC/DOJ/FDA fines or consent decrees found against Skyworks as an operating matter. The one live regulatory event is the FTC Second Request on the Qorvo merger (Feb 5, 2026) — an antitrust review, not an enforcement penalty (see Lens 12/13).
• Conclusion: No material accounting-fraud or enforcement findings — verified via SEC EDGAR EFTS (LR, AAER), web search, and 10-K Item 3 as of 2026-06-30. The only legal/regulatory matters are merger-related.

---

Phase D — Project & stress-test

Forward Projection (EPS, next 3 fiscal years — standalone, pre-merger)

Note: the Qorvo merger (expected close early CY2027) will reshape FY2027+ entirely; this projection is standalone Skyworks (the cleaner base to track), with the merger treated as a separate step-change in Lens 12. All output; inputs labeled.

Base anchors: FY2025 revenue $4,086.9M, non-GAAP EPS ~ $4.65; consensus FY26 ~$4.78, FY27 ~$5.06 (non-GAAP). Q3 FY26 guide mid $925M, GM ~45% non-GAAP.

Base call: standalone non-GAAP EPS recovers toward ~$5.00–5.20 by FY2028 — i.e. roughly back to the FY2024 level after two lost years, driven by Broad Markets + Android replacing lost Apple content rather than by Apple re-accelerating. At ~$72 that is ~14x FY28E — undemanding if the offset materializes. (Forecast.ts create skipped — --watchlist rule; this base is recorded here, not yet committed as a Brier line.)

Bull vs Bear

Bull case. Skyworks is a capex-light (~5% of sales) cash machine — ~$1.1B FCF even at trough — trading at ~15x forward EPS and ~3.7% yield with a net-cash balance sheet. The franchise has a genuine filter/RF process moat, and three things are turning at once: (1) Broad Markets +10% YoY (Wi-Fi 7, data center, automotive) is now 42% of revenue and growing — the diversification finally has traction; (2) the multi-generational Android OEM design win = $1B+ revenue through 2030 is a concrete, quantified plug for the Apple hole; (3) the Qorvo merger turns the most dangerous symmetric competitor into a partner, creating a $7.7B-revenue / $2.1B-adj-EBITDA US RF leader with >$500M synergy target, with a combined customer base that dilutes Apple concentration and adds Qorvo's UHB/envelope-tracking IP. Contrarian view: the market is pricing terminal decline into a company whose growth segment is accelerating and whose biggest competitive threat is being absorbed.

Bear case (permanent-impairment risks). (1) Apple (67%) is structurally engineering Skyworks out — Broadcom took iPhone 17 RF content, iPhone 18 could cut further, and Apple has every incentive to keep dual-sourcing and eventually in-house more RF; a franchise where the top customer is actively reducing you is the definition of a melting ice cube. (2) RFFE is a low-growth, deflationary, ~6%-share-each commodity-ish market dominated by larger, diversified Broadcom/Qualcomm who can bundle — pure-play RF has structurally lower terminal margins. (3) The Qorvo deal might not close — the FTC Second Request (Feb 2026) materially raises antitrust risk for combining the #4 and #5 RF players (overlap in PA/filters); a block leaves Skyworks standalone, shrinking, with deal costs sunk and a $100M reverse termination fee owed to Qorvo.

Pre-mortem (18 months out, thesis broke): the FTC sued to block the merger (or forced value-destroying divestitures); the deal collapsed; meanwhile Apple confirmed another content cut on iPhone 18; the Android $1B win ramped slower than promised; Broad Markets growth decelerated with the broader semi cycle; standalone EPS fell back toward $3.50 and the stock re-rated to ~10x = high-$30s/low-$40s.

Are multiples too high? No — at ~15x forward the multiple is already discounted; the risk is that earnings (the E), not the multiple, fall. The debate is about the denominator, not the ratio.

Devil's Advocate (short-seller)

Dismantling the bull case:

• The whole thing rests on one customer that is leaving. 67% Apple, and Apple is the one customer with the scale, the silicon team, and the motive to design Skyworks out. "Diversification" has been the story for a decade and the company is still 58% Mobile. Every "Android win" and "Broad Markets +10%" is fighting a 67% headwind.
• The merger is the bull's crutch — and it's the weakest link. You are underwriting an outcome controlled by the FTC, which has already issued a Second Request on a combination of the #4 and #5 players in a concentrated market. If antitrust forces divestitures of the overlapping filter/PA assets, the synergy math ($500M) evaporates and you've bought a more-complex, more-indebted (Goldman $3.05B bridge) entity at a worse moment. The bull case quietly assumes the deal closes clean in early 2027 — a heroic assumption.
• The moat is weaker than bulls think. RFFE content gets designed out, not switched out — FY2025 is the proof. Filters are hard, but Broadcom and Murata also make them, and the OEMs hold the whip. ~6% global share is not a moat, it's a seat at a crowded table.
• Capital-allocation own-goal: the company spent ~$1.26B/yr on buybacks/dividends during the good years instead of diversifying away from Apple faster or building scale — and now buybacks are frozen by the merger agreement exactly when the stock is cheap.
• If growth disappoints 20–30%: standalone EPS to ~$3.50 and a 10–12x trough multiple → high-$30s. Downside is real and ~40–50% from here.
• Single scenario that permanently impairs: FTC blocks Qorvo and Apple confirms a further iPhone-18 content cut in the same window — Skyworks is then a sub-scale, shrinking pure-play with sunk deal costs and no strategic answer.

Management Questions (ordered by information value)

1. The FTC issued a Second Request in Feb 2026 — what specific product overlaps are they focused on, and are you prepared to divest filter/PA assets to clear it, or will you walk and pay the $100M fee?
2. Quantify Apple content for iPhone 18 vs iPhone 17 — is the content decline finished, stabilizing, or still falling, and what is your committed dollar content per device through the next two cycles?
3. The $1B+ Android design win through 2030 — name the revenue ramp by fiscal year, the gross margin vs corporate, and what protects it from the same socket-loss dynamic that hit Apple.
4. If the Qorvo deal closes, what is the realistic timeline and split of the >$500M synergy target between COGS and opex, and how much is revenue dis-synergy from overlapping customers?
5. What does the standalone Skyworks operating-margin recovery path look like to 2028 if the merger is blocked?
6. Broad Markets is +10% and 42% of revenue — what is the sustainable through-cycle growth rate, and which sub-segment (Wi-Fi, data center, auto) carries it?
7. R&D rose to ~19% of sales — at what revenue level does that normalize, and what is the target long-term opex model for the combined company?
8. With buybacks frozen by the merger agreement, what is the pro-forma capital-allocation framework (leverage target, dividend, buyback) the day after close, given the Goldman $3.05B bridge?
9. What is your honest assessment of customer concentration two years post-Qorvo — does Apple drop below 50% of the combined entity, and by when?
10. Where does in-house BAW/filter capacity (Japan, Newbury Park) sit on utilization, and is there stranded fixed cost as Apple volume leaves?
11. How do you defend RFFE pricing against Broadcom and Qualcomm's ability to bundle RF with baseband/connectivity?
12. What is the AI-smartphone content thesis worth in dollars-per-phone, and when does it actually inflect volumes/content?
13. Automotive and data-center content — what is the design-win pipeline value, and what is the lag to revenue?
14. What goodwill/intangible impairment sensitivity exists if mobile revenue falls another 15%?
15. What is the board's succession and incentive structure to ensure the new team is paid for diversification and integration, not for a near-term EPS rebound?

---