Impersonal research, traced to filings. An evidence-first read of whether ON Semiconductor (onsemi)'s move up the value chain is real, well-stewarded, and still underappreciated — the same for every reader, from SEC EDGAR 10-K + 8-K filings. This is not personalized advice or a recommendation to any specific person. As of 24 Jul 2026. Disclosures.
How it scores
Seven dimensions, 0–5 — the baseline everyone has (28%) plus the revealed-behaviour Analysis Bridge (72%). The method →
Baseline
Analysis Bridge
The evidence, dimension by dimension
Execution (do-side)
3.5/5Operationally onsemi is executing the hard parts: it delivered the restructuring (2,400 heads, $496M impairments, low-margin line exits), pulled AMG margin up in a down year, funded up-chain M&A internally, generated $1.76B operating cash flow and cut capex sharply — real do-side proof. But the top line has fallen two years running and FY2025 operating income collapsed to $84M (1.4% OPM); the ISG segment took a $230M inventory writeoff. Execution on the transition is credible; execution on financial results is cycle-suppressed. A solid, not stellar, execution grade.
Say/do credibility
3/5Say/do is mixed and must be weighted by track record. DELIVERED: capital allocation (buybacks, capex discipline), and the do-side restructuring actions promised in Q1-2025 were carried out. IN PROGRESS: the SiC / intelligent-power mix-shift is backed by funded M&A, not just slideware. MISSED/UNPROVEN: the headline 'meaningful gross margin expansion' promise runs directly counter to a 1,230 bps margin decline, and the 'secular over cyclical' framing is contradicted by two years of double-digit revenue declines. Management under El-Khoury/Trent (ex-Cypress team with a documented Cypress value-creation record) has a credible strategic voice, but investors were guided toward margin expansion and got a trough. Net: honest disclosure, real do-side actions, but forward promises not yet reconciled to delivered outcomes — mid-band credibility.
Strategy — IP & partnerships
4/5Strategy is genuinely up the value chain: from commodity discretes toward vertically integrated SiC (boule-to-module, the structural differentiator), the Treo scalable analog platform, intelligent sensing (CMOS/SPAD/SWIR for ADAS, robotics), and a fast-growing AI-data-center power franchise built via targeted IP acquisitions (Qorvo SiC JFET, Aura Vcore). Owns SiC crystal growth and wafering in-house — a hard-to-replicate moat vs peers who buy substrates. Only ding: the filing candidly notes it 'rationalized' its patent portfolio and is 'not substantially dependent on any single patent,' so IP protection is portfolio/trade-secret + vertical integration rather than a single blocking patent wall.
Governance & stewardship
3.5/5Experienced, aligned leadership (CEO El-Khoury and CFO Trent, the ex-Cypress team that 5x'd Cypress EV). Clean covenant compliance, conservative balance sheet, transparent restructuring disclosure, ~12-year average employee tenure. Disciplined capital return with a defined 3-year buyback window. No governance red flags surfaced in the 10-K; 8-K stream shows routine board/officer (5.02) and shareholder-vote (5.07) items and a Q4-2025 8-K Item 2.06 (material impairment) that was properly disclosed. Standard-good, not exceptional (single distributor at ~11% of revenue and China/Leshan JV concentration are watch-items, not governance failures).
Margins over the cycle
Operating margin and return on capital across the last decade of SEC filings. A structural climb up the value chain holds its gains through the down-cycle; a commodity name gives them back. Structural vs cyclical →
From SEC EDGAR XBRL.
The say/do credibility ledger
What management promised, versus what the filings show they delivered — the single largest weight in the read.
| Promise | When | Status | What the filings show |
|---|---|---|---|
| Deliver 'meaningful gross margin expansion' via footprint optimization and exiting low-margin product lines. | 10-K FY2025 (2026-02-09), Item 1 Business Strategy Developments | Missed | Consolidated gross margin FELL ~1,230 bps to 33.1% in FY2025 from 45.4% in FY2024 (47.1% in FY2023). Driven by volume decline, fab underutilization and $268M excess/obsolete inventory charges from the realignment. Margin-expansion goal is a forward promise the near-term numbers contradict; too early to call the structural target delivered, but the recent print is a clear miss vs the peak. 10-K FY2025, Item 7 MD&A Executive Overview + Gross Profit and Gross Margin |
| Exit non-core / commodity product lines that do not enhance gross margin ('right-size' the manufacturing footprint). | Q1-2025 Manufacturing Realignment announcement; reiterated 10-K FY2025 Item 1 & Item 7 | In progress | Do-side action delivered: terminated ~2,400 employees ($67.1M severance), $496.0M non-cash equipment impairments under held-for-sale accounting, $103.9M other exit costs, and inventory write-offs — i.e. real capacity reduction, not just words. AMG gross margin actually rose to 51.1% from 50.1% partly by shedding low-margin EFK manufacturing-services revenue, evidence the mix-up is executing at the segment level. Full margin payoff pending demand recovery. 10-K FY2025, Item 1 (2025 Manufacturing Realignment Program); Item 7 MD&A AMG gross margin discussion; 8-K 2025-02-25 Item 2.05 |
| Grow SiC and be a vertically integrated intelligent-power leader for auto EV + AI data center. | 10-K FY2025, Item 1 (PSG; 2025 acquisitions) | In progress | Executed bolt-ons that deepen the up-chain position: acquired Qorvo SiC JFET tech for $118.8M (Jan-2025) to address AI-data-center AC-DC power, and Aura Semiconductor Vcore power IP for up to $144M (Oct-2025). PSG other/AI-data-center revenue grew +$15.1M even as auto (-$438M) and industrial (-$120M) fell — the growth vector is real but small vs the cyclical drag. 10-K FY2025, Item 1 (Acquisitions); Item 7 MD&A Revenue from PSG |
| Return capital to shareholders / disciplined capital allocation. | Share Repurchase Program; Nov-2025 Board authorization | Delivered | Repurchased ~27.9M shares for ~$1,375M in FY2025, repaid $375M of revolver, and Board approved a fresh $6.0B buyback (2026-2028); already bought $175.6M under it through 2026-02-04. Capex cut hard (19%->6% of revenue) as the SiC build completes — capital discipline demonstrably delivered. 10-K FY2025, Item 7 MD&A Financing Activities / Key Financing and Capital Events |
| Business is driven more by content gains and secular drivers than by macro/industry cyclicality. | 10-K FY2025, Item 1 (Seasonality) | Too early | FY2025 revenue fell 15.3% and FY2024 fell 14.2% — two straight years of double-digit declines led by automotive and industrial volume, i.e. the cycle dominated the print. The secular-over-cyclical claim is not yet borne out in results; content/AI-datacenter gains exist but are too small to offset the auto/industrial downturn. Watch whether content gains re-accelerate revenue in the up-cycle. 10-K FY2025, Item 1 (Seasonality) vs Item 7 MD&A Revenue (FY2025 -15.3%, FY2024 -14.2%) |
Triangulation chain
A real migration lights up all five: stated → approved → funded → permitted → protected. An announced-only one stalls early.
onsemi frames itself as an 'intelligent power and intelligent sensing' company driving electrification/energy efficiency in automotive, industrial and AI data center; strategy is to 'increase profitable revenue through differentiated technologies,' 'right-size manufacturing footprint,' 'exit product lines that do not enhance gross margin or satisfy strategic objectives,' and pursue 'meaningful gross margin expansion.' Vertically integrated SiC (EliteSiC) is the flagship up-chain platform; Treo is the scalable analog/mixed-signal platform. (10-K FY2025, Item 1 Business ('Business Strategy Developments', PSG/AMG/ISG); Item 7 MD&A Executive Overview ('exiting product lines that do not enhance gross margin'))
Board-level actions executed: 2025 Manufacturing Realignment Program (announced Q1-2025) realigning internal fab capacity to long-term demand; Board approved a new $6.0B share-repurchase program in Nov 2025 (Jan-2026 through Dec-2028). Restructuring formalized via 8-K Item 2.05 (2025-02-25). (10-K FY2025, Item 1 ('2025 Manufacturing Realignment Program'); Item 7 MD&A Financing Activities (Nov-2025 $6.0B buyback); 8-K 2025-02-25 Item 2.05 (Costs Associated with Exit/Disposal Activities))
Self-funded: FY2025 operating cash flow $1,759.8M, cash + short-term investments ~$2,547.6M, $1.5B undrawn revolver, and in compliance with all covenants. Funded the up-chain M&A ($118.8M cash for Qorvo SiC JFET tech; up to $144M for Aura Vcore power IP) and $1,375M of buybacks while repaying $375M of revolver. Capex deliberately cut to ~6% of revenue (from 19% in 2023) as SiC capacity build completes; guides ~5% for 2026. (10-K FY2025, Item 7 MD&A Liquidity/Sources and Uses of Cash; Item 1 (Qorvo SiC JFET $118.8M; Aura Vcore up to $144M))
Owns/operates a global front-end fab and back-end network (US, Czech Republic, Japan, S. Korea, Malaysia; SiC boules in Hudson NH; SiC wafers in Roznov CZ and Bucheon KR). Material compliance with environmental, ITAR, FCPA, export-control regimes stated; no permitting gate flagged. Site/regulatory footprint already in place rather than pending — permitting is not the constraint for this transition (a mix-shift, not a greenfield build). (10-K FY2025, Item 1 (Manufacturing and Design Operations; Government Regulation))
Vertically integrated SiC (boule-to-module) is the moat: in-house SiC crystal growth (Hudson NH) plus SiC wafering supports 'supply assurance, cost competitiveness, and device performance.' Holds a substantial worldwide patent portfolio, 'strategically rationalized... to focus on higher-value inventions aligned to core technologies.' Long-term auto supply agreements, 45-120 day cancellation terms, and deep customer design-win relationships create switching costs. Caveat: filing states business is 'not substantially dependent on any single patent' and admits IP rationalization (fewer, higher-value patents). (10-K FY2025, Item 1 (PSG competitive strengths; Patents/Trademarks/IP; 'vertically integrated SiC manufacturing strategy'))
Catalysts
- AI data center power demand ramp; onsemi positions its intelligent-power portfolio for the 'complete power tree' as new AI processors/racks enter market (Qorvo SiC JFET + Aura Vcore acquisitions target this). · 2025 (10-K FY2025)
New, structurally growing revenue vector at higher value than commodity discretes; +$15.1M PSG other/AI-datacenter revenue in a down year is early proof.
- US and EU semiconductor manufacturing incentives (CHIPS-style funding) cited as available support for research/development/manufacturing. · 2025 (10-K FY2025 Risk Factors)
Potential subsidy for domestic SiC/fab investment; onsemi flags competitive risk if peers get more — a two-sided but net-supportive policy window.
- US-China tariff/export-control escalation; Leshan (China) 80%-owned assembly/test JV exposed; possible entity-list additions and export-license requirements for China customers. · 2025 (10-K FY2025 Risk Factors)
Geopolitical downside to a China-dependent back-end footprint and China end-demand; a real risk-window that could compress the transition's economics.
Sources
Bridged from SEC EDGAR 10-K + 8-K filings, latest 10-K filed 9 Feb 2026. Read the 10-K on SEC EDGAR → Provenance: agent-grounded-fulltext. Financials from SEC EDGAR XBRL; price from public market data.