United Microelectronics
Pure-play wafer foundry: manufactures chips to fabless/IDM customer designs on a per-wafer basis; revenue driven by wafer shipments x ASP x utilization, with mix shift toward higher-value specialty processes (22/28nm HV, RFSOI, BCD, embedded flash).
- 2026-08-04This market capitalisation previously read $62B (as of 2026-07-02). Restated to ~$47B on this refresh, roughly 24% lower.
- 2026-08-04This share price previously read $24.5. Restated to $18.69 on this refresh, roughly 24% lower.
Sources — 12 figures with citations
- Q2 2026 income statementfiled2026-06-30 (released 2026-07-29)Revenue NT$68,732,662k; gross profit NT$22,322,790k; operating income NT$14,949,990k; non-operating subtotal NT$30,236,131k; pre-tax income NT$45,186,121k; income tax expense NT$2,962,306k; net income attributable to parent NT$42,259,962ksec.gov — Consolidated Statements of Comprehensive Income, Q2 2026 consolidated financial statements filed as Exhibit 99.1 to the 6-K of 2026-07-29
- Q2 2026 headline metrics as announcedfiled2026-07-29Revenue NT$68.73B (US$2.18B); gross margin 32.5%; operating margin 21.8%; 22/28nm 37% of revenue; capacity utilization 85%; net income attributable NT$42.26B (US$1.34B); EPS NT$3.39; EPS per ADS US$0.537; 22nm 17.5% of sales; wafer shipments +10.6% QoQsec.gov — Exhibit 99.6 'UMC announced its operating results for the second quarter of 2026' in the 6-K filed 2026-07-29
- Q3 2026 guidance and FY2026 capexfiled2026-07-29Wafer shipments up high-single digit QoQ; ASP in USD remains firm; gross profit margin mid-30% range; capacity utilization 90%+; 2026 capex US$2.0 billion (raised from US$1.5B)sec.gov — 'Third Quarter 2026 Outlook & Guidance' in Exhibit 99.6
- Board-approved capital budget and capacity planfiled2026-07-29NT$148,680 million projected investment for capacity deployment; Singapore P4 cleanroom installation + tools for silicon photonics; construction of a new Tainan fab building shell to house future P7/P8, explicitly to 'minimize upfront depreciation'sec.gov — Exhibit 99.2 'The board meeting approved capital budget execution' in the 6-K filed 2026-07-29
- Balance sheet at 2026-06-30filed2026-06-30Cash & equivalents NT$124,706,465k; financial assets at amortized cost (current) NT$22,135,867k; short-term loans NT$2,806,983k; current portion of long-term liabilities NT$9,159,945k; bonds payable NT$34,075,713k; long-term loans NT$11,489,924k; total liabilities NT$222,048,583k; equity attributable to parent NT$443,920,249k; total assets NT$665,972,883k; net PP&E NT$259,212,411k; dividends payable NT$32,704,164ksec.gov — Consolidated Balance Sheets. Derived net cash: 124,706,465 + 22,135,867 - (2,806,983 + 9,159,945 + 34,075,713 + 11,489,924) = NT$89,309,767k. Excludes lease liabilities (NT$640,690k current + NT$5,276,626k non-current).
- H1 2026 cash flow and derived free cash flowderived2026-06-30Net cash provided by operating activities NT$55,678,697k (H1 2025: NT$45,923,736k); acquisition of PP&E NT$21,348,637k; acquisition of intangibles NT$1,339,043k; depreciation NT$30,813,547k; derived FCF NT$34,330,060k = 26.5% of H1 revenuesec.gov — OCF and capex lines are filed in the Consolidated Statements of Cash Flows. FCF derived: 55,678,697 - 21,348,637 = 34,330,060; margin = 34,330,060 / 129,770,564 = 26.46%. UMC does not report FCF. Capex intensity derived: 21,348,637 / 129,770,564 = 16.45%.
- Composition of the non-operating gainfiled2026-06-30Share of profit of associates and joint ventures NT$23,605,970k in Q2 2026 (vs NT$446,242k in Q2 2025) and NT$26,421,106k in H1; other gains and losses NT$6,054,359k in Q2. Equity-method carrying values: HSUN CHIEH INVESTMENT NT$12,792,773k -> NT$34,229,937k, YANN YUAN INVESTMENT NT$13,722,026k -> NT$30,303,137k, SIS NT$3,562,947k -> NT$9,169,636k, UNIMICRON NT$14,428,352k -> NT$17,813,795k (2025-12-31 -> 2026-06-30); total equity-method investments NT$48,642,917k -> NT$96,474,468ksec.gov — Note 6(7) Investments Accounted for Under the Equity Method. Derived: non-operating share of pre-tax income = 30,236,131 / 45,186,121 = 66.9%; associates alone = 23,605,970 / 45,186,121 = 52.2%.
- Effective tax ratederived2026-06-306.56% Q2 2026 vs 12.86% Q2 2025sec.gov — Derived from filed lines: 2,962,306 / 45,186,121 = 6.56%; prior year 1,305,996 / 10,153,978 = 12.86%
- H1 2026 board-reported cumulative resultsfiled2026-06-30H1 revenue NT$129,770,564k; gross profit NT$40,141,280k (30.93% derived); operating income NT$26,226,401k (20.21% derived); pre-tax NT$61,829,951k; profit attributable to owners NT$58,431,436k; basic EPS NT$4.68sec.gov — Exhibit 99.1 'Announcement of board meeting approved the consolidated financial statements for the second quarter of 2026', period 2026/01/01~2026/06/30. Margins derived from the filed absolute figures.
- ADR price and market capitalisationmarket2026-08-03 (close)$18.69 per ADS close on 2026-08-03; provider market cap $45.84B; 12.58B ordinary shares outstanding; trailing P/E 17.47; 52-week range $6.56-$28.96stockanalysis.com — Closing price, not intraday. Derived cross-check: filed common stock NT$125,769,849k at NT$10 par = 12.577B ordinary shares; /5 per ADS = 2.515B ADS-equiv x $18.69 = ~$47.0B, ~2.5% above the provider's $45.84B (ADR-vs-TWSE price and FX timing).
- ADR price path across the reporting windowmarket2026-08-03 (close)$25.83 (2026-07-06) -> $21.19 (2026-07-21) -> $19.08 (2026-07-27) -> $17.11 (2026-07-29, print day) -> $18.94 (2026-07-30) -> $19.03 (2026-07-31) -> $18.69 (2026-08-03). -27.6% from 2026-07-06stockanalysis.com — All closes. Supports the 'market did not pay for the beat' point in the bear case.
- USD/NTD conversion basis used for derived USD figuresderived2026-06-30~31.53 NT$ per US$, implied by the company's own Q2 translation of NT$68.73B to US$2.18Bsec.gov — 68,732,662 / 2,180,000 = 31.53. Applied to H1 revenue (NT$129.77B -> ~US$4.12B), H1 capex (NT$21.35B -> ~US$677M) and net cash (NT$89.3B -> ~US$2.83B).
Earnings, margins, COGS & capex
UMC earns per-wafer fees across a diversified mature/specialty node base; profitability is a function of utilization and mix. FY2025 was roughly flat-to-modestly-up in NT$ (+2.3%, NT$232.3B to NT$237.6B) but +5.3% in USD with wafer shipments +12.3% YoY, as a sharp 22nm ramp (+93% YoY to a record) offset ASP/FX headwinds; full-year gross profit still fell 8.9% and operating income 14.8% YoY as pricing and depreciation bit. Margins held near 29% through the mature-node downcycle; Q1 2026 showed a demand recovery (utilization 79%) and a large net-income jump (+108% YoY, aided by non-operating/equity-method items).
Income statement — where each revenue dollar goes
% of revenueOf every $1 of revenue, ~68¢ is cost of goods and ~11¢ operating expense, leaving ~22¢ of operating profit (~18¢ net).
Revenue trend
Margins
up from 26.7% a year ago; guided ~30% Q2 2026
held near 29% through the downcycle; gross profit -8.9% YoY
operating income -14.8% YoY FY2025; recovering with utilization
down vs ~20% FY2024 on lower NT$ EPS (NT$3.34 vs NT$3.80)
COGS structure
Dominated by fixed fab costs — depreciation of fab equipment, cleanroom overhead, direct labor, plus variable materials (silicon wafers, photomasks, process chemicals/gases) and electricity. High operating leverage: gross margin swings sharply with utilization (79% in Q1 2026 vs 90%+ at peak). Depreciation is set to rise as new Singapore (Fab 12i) and Japan capacity and the Intel-Arizona 12nm program come online, pressuring margins near-term.
Capex
US$1.5B budgeted for 2026 (vs US$1.6B in 2025) — disciplined relative to leading-edge peers. Directed at 22/28nm specialty capacity expansion (Singapore Fab 12i, Japan USJC) and the co-developed 12nm FinFET node at Intel's Ocotillo (Arizona) site.
Latest earnings
Beat, but the beat's composition matters more than its size. Revenue NT$68.73B (US$2.18B) was +17.0% YoY / +12.6% QoQ, gross margin expanded 330bps sequentially to 32.5%, operating margin to 21.8%, and net income attributable to parent hit NT$42.26B (US$1.34B), EPS NT$3.39 / US$0.537 per ADS — roughly 4.7x the year-ago NT$8.90B. But NT$30.24B of the NT$45.19B in pre-tax income (66.9%) came from NON-operating items, chiefly NT$23.61B of equity-method share of associates' profit, and the effective tax rate fell to 6.6% from 12.9%. Core operating income was NT$14.95B. The ADR nonetheless fell from $19.08 (2026-07-27) to $17.11 on the 2026-07-29 print day and closed at $18.69 on 2026-08-03.
Q3 2026 (given 2026-07-29): wafer shipments to increase by high-single digits QoQ; ASP in USD to remain firm; gross profit margin in the mid-30% range; capacity utilization 90%+. FY2026 capex raised to US$2.0 billion.
- Capacity utilization
- 85% in Q2 2026, up from 79% in Q1 2026; guided 90%+ for Q3 2026
- 22/28nm revenue share
- 37% of Q2 revenue (record high); 22nm alone 17.5% of sales
- Net income attributable to parent
- NT$42,259,962k (US$1.34B), vs NT$8,902,530k in Q2 2025
- Share of profit of associates (non-operating)
- NT$23,605,970k in Q2 2026 vs NT$446,242k in Q2 2025 — 52% of pre-tax income
- Effective tax rate
- 6.6% (NT$2,962,306k on NT$45,186,121k pre-tax) vs 12.9% in Q2 2025 (derived)
- H1 2026 operating cash flow
- NT$55,678,697k, up from NT$45,923,736k in H1 2025
- Silicon photonics
- First mass-production delivery of 12-inch photonic ICs to a customer (July 2026); platform open for general customer use in 2027
- Capacity plan
- Singapore P4 cleanroom + tools for silicon photonics; new Tainan fab building shell for future P7/P8 — phased, shell-first to defer depreciation
Growth drivers
- 22nm ramp — +93% YoY in 2025 to a record, UMC's highest-value node; grows within the 34% 22/28nm revenue envelope as 22nm displaces 28nm
- Specialty processes — OLED/display driver ICs (DDIC), power management (PMIC/BCD) for AI servers, RFSOI, embedded flash MCUs
- 12nm FinFET via the Intel Arizona partnership — mass production targeted 2027, opening a node beyond UMC's historical 22nm floor
- Geographic diversification (Singapore, Japan, US) as fabless customers de-risk supply away from Taiwan/China concentration
- Cyclical recovery in consumer/smartphone demand and utilization normalization into H2 2026, with price hikes flagged on some processes
Bull & bear
The mature-node cycle has visibly turned and UMC's operating leverage is doing exactly what it does on the way up: utilization 79% -> 85% with Q3 guided above 90%, gross margin 29.2% -> 32.5% with mid-30s guided, operating income +38% YoY. Record 22/28nm mix, a first mass-production silicon-photonics delivery, and a shell-first capacity plan that adds optionality without front-loading depreciation — all on a fortress balance sheet with ~NT$89B net cash.
- Utilization improved to 85% from 79% and is guided 90%+ for Q3 2026 — the single most important variable for a high-fixed-cost foundry, and it is moving up two quarters in a row
- Gross margin 32.5% in Q2 (from 29.2% in Q1 and 28.7% in Q2 2025) with Q3 guided to the mid-30% range; operating margin 21.8%, operating income +38.2% YoY on +17.0% revenue
- Revenue NT$68.73B (US$2.18B), +12.6% QoQ on +10.6% wafer shipments with USD ASP guided to remain firm — volume growth is not being bought with price
- 22/28nm hit a record 37% of revenue with 22nm alone at 17.5% — the highest-value node in the portfolio keeps displacing 28nm
- First mass-production delivery of 12-inch photonic ICs (July 2026), with the silicon-photonics platform open for general customer use in 2027 — a genuine AI-interconnect attach point rather than a claim
- The capacity plan is capital-disciplined by construction: Singapore P4 cleanroom + tools now, Tainan building SHELL only for future P7/P8, explicitly to 'minimize upfront depreciation' while cutting future lead times
- Balance sheet: NT$124.7B cash + NT$22.1B amortized-cost securities against NT$57.5B of debt (~NT$89.3B / ~US$2.83B net cash), equity attributable NT$443.9B on NT$666.0B of assets
- H1 cash generation was strong even while investing: OCF NT$55.7B (+21% YoY), NT$34.3B of derived free cash flow after NT$21.3B of PP&E capex
- Equity-method holdings are now carried at NT$96.5B (from NT$48.6B at year-end 2025) — a real, marked-up stake in Taiwan's AI supply chain (Unimicron, SIS, Faraday) sitting inside a 17x-earnings foundry
Strip the portfolio marks and the quarter is good, not spectacular: two-thirds of pre-tax income was non-operating, NT$23.6B of it equity-method associate profit from investment-holding affiliates, and the tax rate halved. Meanwhile the capex budget was raised 33% to US$2.0B against an NT$148.7B multi-year plan — a depreciation wave aimed at a business whose margin is a utilization function, in a mature-node market where Chinese capacity is the structural ASP overhang. The ADR is down 27.6% in four weeks despite the beat.
- Earnings quality: NT$30,236,131k of the NT$45,186,121k pre-tax income (66.9%) was non-operating. Core operating income was NT$14,949,990k — so the 4.7x net-income growth headline is not a foundry result
- The single biggest driver, NT$23,605,970k of 'share of profit of associates', came from marked-up investment-holding affiliates: HSUN CHIEH INVESTMENT NT$12.79B -> NT$34.23B and YANN YUAN INVESTMENT NT$13.72B -> NT$30.30B carrying value in six months, plus SIS NT$3.56B -> NT$9.17B. These are portfolio gains, and they reverse
- Effective tax rate fell to 6.6% (NT$2,962,306k on NT$45,186,121k) from 12.9% a year earlier — a further non-operational flatter to EPS that will not persist
- Capex raised 33% to US$2.0B for 2026, on top of a board-approved NT$148.68B (~US$4.7B) capital-budget execution plan — the depreciation from Singapore P4 and the new Tainan fab lands into a cycle whose peak is unknown
- Depreciation already ran NT$30,813,547k in H1 2026 (up from NT$27,227,474k) while net PP&E FELL from NT$271.4B to NT$259.2B — the new build re-inflates a fixed-cost base that operating leverage cuts both ways on
- Equity-method investments are now NT$96.5B, 14.5% of total assets — an increasing share of book value and of reported earnings is a listed-equity portfolio, which is not what a foundry multiple is meant to capitalise
- The 12nm FinFET program with Intel in Arizona is still pre-revenue (mass production targeted 2027), so the node-extension thesis contributes nothing to current earnings while consuming R&D (NT$9,322,501k in H1) and capex
- The market is not paying for the print: ADR $25.83 (2026-07-06) -> $17.11 on the 2026-07-29 print day -> $18.69 (2026-08-03), a 27.6% drawdown across the reporting window
- Mature-node structural risk is unchanged: Chinese domestic capacity (SMIC, Hua Hong) plus specialty-node overcapacity keeps a lid on ASPs, and 'ASP remains firm' guidance is not ASP growth
- H1 inventory write-downs of NT$1,494M (against NT$324M of reversals) show the demand recovery is not uniform across the portfolio
What it is worth
Peer-relative multiple (vs GFS, TSM, VIS, TSEM) cross-checked with a reverse-DCF read on the current ADR price; UMC screens at PE ~40 (per aggregator data, 2026-07-02) and pays a ~1.7% dividend.
~30-50% downside
if the multiple normalizes toward its historical mid-teens on any combination of: renewed Chinese 28nm price cuts, a soft consumer cycle dropping utilization below ~75%, an Intel-Arizona 12nm slip, or NT$ strength — de-rating a cyclical mature-node foundry back to a cyclical mature-node multiple.
Roughly range-bound / modest — earnings recover with the cycle and 22nm mix, but a PE ~40 already discounts much of it; total return is carried mainly by the ~1.7% dividend plus low-teens earnings growth, with multiple compression a headwind. Fair value near a high-teens-to-low-20s PE.
~30% upside
if the recovery sustains — GM holds >30% on 22nm mix, utilization climbs to mid-80s+, 12nm/Intel ramps on schedule in 2027, mature-node pricing firms into 2027, and consolidation optionality re-rates the group. Multiple sustained/expands on visible growth.
At ~US$62B market cap / PE ~40, UMC trades well above its historical mid-teens multiple and above mature-node peer GlobalFoundries — the price embeds a sustained mature-node upcycle, continued 22nm margin-mix gains, and successful 12nm/Intel execution in 2027. That is a demanding bar for a business that grew NT$ revenue +2.3% and saw EPS decline in FY2025 (NT$3.34 vs NT$3.80). The multiple, not the franchise, is the central valuation risk.
SWOT
Strengths
- Scale + 40+ year operating history as the world's #3-4 foundry and #2 Taiwan pure-play; deep, sticky fabless/IDM customer relationships
- Specialty-process leadership (22/28nm HV for OLED drivers, RFSOI, BCD, eNVM) that commands better margins than commodity mature-node capacity
- Strong balance sheet — net cash, low leverage (D/E ~0.19), consistent dividend (~1.7% yield on the ~NT$2.61/share declared for FY2025), self-funds capex
- Geographic diversification across Taiwan, Singapore, Japan and (via Intel) the US — a de-risking advantage as customers seek non-China/-Taiwan supply
Weaknesses
- No leading-edge (sub-12nm) capability — structurally excluded from the highest-growth AI-compute and advanced-mobile wafer demand that TSMC monopolizes
- High fixed-cost/operating-leverage model — margins and earnings swing hard with the mature-node cycle and utilization (FY2025 gross profit -8.9%, operating income -14.8% YoY)
- NT$ revenue only +2.3% in FY2025 and EPS down YoY (NT$3.34 vs NT$3.80) — growth is modest and FX-exposed
- Elevated valuation (PE ~40) leaves little room for a demand or pricing disappointment
Opportunities
- 12nm FinFET with Intel (production 2027) extends the addressable node range and adds a US-manufactured, geopolitically-favored offering
- AI-server buildout drives mature-node demand for power ICs, PMICs, and interface/DDIC chips — not just leading-edge
- Mature-node price recovery as utilization tightens across the industry into H2 2026/2027
- Reshoring / China+1 supply-chain diversification favors non-mainland specialty foundries; recurring mature-node M&A speculation could re-rate the group
Threats
- Aggressive mature-node capacity buildout and price cutting by mainland-China foundries (SMIC, Hua Hong), heavily state-subsidized — the primary structural overhang on 28nm pricing
- Mature-node overcapacity industry-wide keeping utilization below peak
- Geopolitical/tariff risk to Taiwan-centric manufacturing and cross-strait tension
- Cyclicality of consumer electronics/smartphone end-markets; FX (NT$/US$) volatility
Moats, dependencies & bottlenecks
Moats
Moderate-strong Qualification-locked customer designs (e.g. OLED display drivers, PMICs) create switching costs and better-than-commodity margins; harder for subsidized Chinese entrants to replicate quickly on the high-value nodes.
A new mature-node fab costs billions and years; UMC's installed base and 40-year process library are a real barrier — but one China is actively buying past with state capital.
Fabless customers (MediaTek, Realtek, Novatek) qualify designs to specific UMC processes; requalification is costly, but not insurmountable when Chinese foundries undercut on price.
A differentiator vs pure-Taiwan or pure-China peers as customers demand supply-chain resilience; the Intel-US footprint is unique among mid-tier foundries but still ramping.
Dependencies
Qualcomm, Realtek, Novatek, Texas Instruments) Customer / demand Revenue is downstream of consumer electronics, smartphone and increasingly AI-server IC demand; concentration in a handful of large fabless accounts.
Applied Materials, Lam Research, KLA, Tokyo Electron) Tool availability and cost gate fab expansion; less acute for mature nodes than for EUV-dependent leading edge.
SUMCO, GlobalWafers, Entegris) Wafer/chemical/gas supply and pricing feed directly into COGS.
Partner / roadmap UMC's node-ceiling breakthrough depends on Intel executing the Arizona ramp on schedule (certification targeted end-2026, production 2027); Intel's own turmoil is an execution risk.
Pricing is set at the margin by state-subsidized Chinese capacity; UMC is a price-taker on commodity 28nm.
Advantages
- Highest-value 22nm ramping to record levels — margin-accretive growth within the 34% 22/28nm mix
- Net-cash balance sheet self-funds capex and dividend through the cycle
- Multi-geography footprint (unique US presence via Intel) as customers de-risk
- Deep specialty-process library and sticky, qualification-locked customer base
Weaknesses
- No leading-edge roadmap — structurally excluded from the fastest-growing wafer demand
- Cyclical, high-fixed-cost earnings with modest secular growth (FY2025 operating income -14.8% YoY)
- Directly exposed to subsidized Chinese mature-node price competition
- Valuation (PE ~40) prices in an upcycle with little margin for error
Bottlenecks
- Leading-edge exclusion — no path below 12nm caps addressable market and growth ceiling
- Mature-node overcapacity / utilization — margins gated by industry-wide 28nm supply, largely out of UMC's control
- Depreciation ramp from Singapore/Japan/Arizona capacity pressures near-term margins before the revenue arrives
- FX translation (NT$ reporting) can mask or erode USD-basis growth
Top signals & trends
Top signals
22nm +93% in 2025 to a record; 22/28nm held ~34% of revenue as 22nm displaces 28nm — the clearest margin-mix positive to track.
Bullish if it climbs toward mid-80s+ · High operating leverage means utilization is the single biggest margin driver each quarter.
Bullish catalyst, execution-dependent · On-schedule per May 2026 updates; slippage or Intel disruption would remove the key growth optionality.
Bearish overhang · Subsidized price cuts are the main structural risk to UMC's commodity mature-node margins.
Management flagged process price increases and new orders on rising material costs — watch for realization in Q3/Q4 prints.
Trends
PMICs, BCD power ICs, and DDICs for AI infrastructure add mature-node demand UMC can serve.
Favors UMC's Singapore/Japan/US diversification — a structural tailwind for non-mainland specialty foundries.
SMIC/Hua Hong flooding 28nm caps industry pricing and threatens share over time.
Node migration in UMC's specialty sweet spot lifts value per wafer.
Positive (optionality) · Recurring reports of interest in combining mid-tier mature-node foundries; unconfirmed, treat as optionality not base case.
Ecosystem & competitor graph
Suppliers feed the company; customers pull from it. Line thickness shows the strength of each tie (supply-chain dependency, customer earnings contribution). Hover to isolate a tie.
Lithography systems (DUV for mature nodes).
Deposition/etch/CMP process equipment.
Etch and deposition tools.
Process control / inspection/metrology.
Coater/developer and etch equipment.
Silicon wafer substrate supplier.
Advanced materials, filtration, and process chemicals/gases.
Major fabless customer — smartphone/connectivity SoCs on mature/specialty nodes.
RF/PMIC and companion chips fabricated at specialty foundries.
Connectivity/networking/audio ICs — core mature-node fabless customer.
Display driver ICs — anchor demand for UMC's 22/28nm HV specialty.
Display drivers and timing controllers on specialty nodes.
IDM that also outsources selected analog/power capacity.
Automotive/industrial MCUs and analog — mature-node outsourcing.
The dominant foundry (~65%+ share) and the leading-edge monopolist; competes with UMC on mature/specialty nodes but plays in a different weight class overall.
The closest US-listed mature/specialty peer — automotive, RFSOI, FDX; recurring mature-node combination speculation. Direct comp on nodes and end-markets.
Taiwan mature-node specialist (PMIC, DDIC); partnering with NXP on a Singapore 12-inch fab — direct specialty-node competition. TSMC-affiliated.
Niche analog/specialty foundry (RF, power, sensors); smaller (~1% share) but overlaps in RFSOI/BCD specialty.
Taiwan mature-node/memory foundry; overlaps in DDIC and commodity mature logic.
China's largest foundry; state-subsidized mature-node capacity and aggressive 28nm pricing is UMC's primary structural competitive threat. Named for analysis only.
China specialty mature-node foundry (power, eNVM, RF); subsidized capacity adds to mature-node oversupply. Analysis context only.
Leading-edge + mature foundry; competes at the high end and selectively on specialty.