United Microelectronics Corp | World #3 pure-play foundry | 28nm + mature nodes
| | ADR (USD) | Underlying (TWD) | |--|--|--| | Price (2026-05-28) | $23.155 | ~TWD 143 | | 52w High | $23.69 | TWD 151 | | 52w Low | $6.58 (Aug 2025) | TWD 40.1 | | YTD 2026 | +195% | ~same | | 5y | +226% | ~same |
The +195% YTD is the critical context for this entire screen. UMC bottomed Aug 2025 at $6.58, was legitimately cheap at ~10x fwd P/E then. At $23.155 today it has already re-rated to 29x fwd — this is a post-recovery valuation screen, not a trough screen.
All financials reported in TWD; converted at TWD/USD 32.0. Each ADS = 5 ordinary TWD shares.
| Metric | Value | |--|--| | Market cap | ~$58B USD | | Trailing P/E | 36.8x | | Forward P/E | 28.9x | | P/B (ADR-adjusted) | 4.5x (distorted) / 1.74x TWD basis | | EV/EBITDA | ~16.6x (net-cash adjusted) | | P/S (ADR) | ~7.8x | | Dividend yield | 2.1% (last ADS div: $0.483) | | 5y avg yield | 5.83% | | Payout ratio | ~85% |
P/B via yfinance on the ADR inflates to 4.5x due to ADS conversion mechanics; the correct TWD-basis P/B (Finnhub) is 1.64–1.74x. Both are used in filter assessment below.
| Item | USD | |--|--| | Cash & equivalents | $3.46B | | Total debt | $1.87B | | Net cash | +$1.59B | | LT debt | $1.42B | | Current ratio | 2.72 | | Working capital | $3.59B | | Total equity | $11.44B | | Interest coverage | >20x (implied from EBIT/interest) |
Balance sheet is excellent — net cash, current ratio well above 2.0, LT debt far below working capital.
| Year | Revenue | EBITDA | Net Inc | FCF | Capex | |--|--|--|--|--|--| | 2022 | 8.71 | 4.86 | 2.80 | 1.97 | 2.59 | | 2023 | 6.95 | 3.45 | 1.87 | -0.25 | 2.94 | | 2024 | 7.26 | 3.34 | 1.52 | 0.08 | 2.85 | | 2025 | 7.42 | 3.40 | 1.26 | 1.53 | 1.58 |
Q1 2026 (quarterly): Revenue TWD 61.0B (~$1.91B), Net income TWD 16.2B (~$507M). Q1 2026 net income already equals 40% of full-year FY2025 — utilization recovery and margin expansion are real.
Capex halved from $2.85–2.94B (2023–24) to $1.58B in FY2025 — deliberate capex discipline, converting to FCF positive. Q2 2026 guided utilization low-80% range and ASP up low-single-digits (H2 price hike letters sent to customers).
| Year | USD/ADS | |--|--| | 2019 | $0.094 | | 2020 | $0.136 | | 2021 | $0.285 | | 2022 | $0.517 | | 2023 | $0.586 | | 2024 | $0.463 | | 2025 (paid Jun 2025) | $0.483 | | 2026 (ex-div Jun 24) | $0.483 announced |
Uninterrupted since 1994 (30+ years). Payout ratio ~85% is high relative to earnings, but FCF covers it in recovery years. The yield at $6.58 trough was ~7.3%; at $23.155 it's 2.1% — yield compression is the clearest sign of the re-rating.
No Form 4 insider transactions reported for the NYSE ADR (Finnhub returned empty). Consistent with a Taiwanese company where insiders transact on TWSE, not NYSE ADR. No positive insider buying signal available via this data source.
| | | |--|--| | Consensus (ADR, 5 analysts) | Sell | | Mean target (ADR) | $10.36 | | High target | $12.74 | | Low target | $7.40 | | Implied downside | -55% from current | | 90d revision trend | Improved vs Apr (6 buy + 18 hold + 3 sell + 4 strongBuy vs prior 3 buy + 17 hold + 10 sell) |
The May 2026 recommendation mix improved sharply after Q1 beat (buy+strongBuy went from 4 to 10, sell+strongSell dropped from 13 to 4 on 2303.TW). However the ADR-coverage price targets ($10–12) haven't updated to reflect the Q1 surge — a clear lag artifact. TW-listed analyst consensus would be more relevant; the ADR target is stale.
Q1 2026 (reported 2026-04-29): Revenue TWD 61.0B (+5.5% YoY). EPS TWD 1.29/share (+100% YoY). Utilization 79%. Gross margin 29.2%. Stock +6% on day. Key disclosures: (1) 22nm now 14% of revenue — record, 50+ tape-outs by year-end; (2) H2 price hike letters sent citing raw material, energy, logistics inflation + sustained demand; (3) Q2 guide: wafer shipments up high-single-digits, ASP up low-single-digits, gross margin ~30%, utilization low-80%.
Q4 2025 (reported 2026-01-28): EPS $0.129/ADS vs cons $0.131 (miss -1.4%). Utilization recovering from trough. FY2025 full year FCF $1.53B vs near-zero in 2024 — capex discipline working.
Key events:
China overcapacity: SMIC, Hua Hong, Nexchip, XMC have collectively added massive mature-node capacity. Chinese foundries expected to hold >25% of top-10 mature capacity by end 2025. This is structural, not cyclical. UMC's response is specialization: 22nm leadership, display drivers, specialty processes (BCD, HV, embedded flash) that Chinese foundries cannot easily replicate. The H2 ASP hike is bold in this context — it implies UMC's differentiated capacity is sold out while commodity 28nm+ faces Chinese pricing pressure.
Cycle position: Recovery phase, not trough. Utilization 79%→80%+ with ASP hikes confirms cycle inflection. The thesis hook was correct — but the market figured this out before this screen ran.
| Filter | Threshold | Result | Verdict | |--|--|--|--| | F1 Market cap | > $2B | ~$58B | PASS | | F2 Current ratio | >= 2.0 | 2.72 | PASS | | F2 LT debt < WC | | $1.42B < $3.59B | PASS | | F3 Positive EPS | All 5–10y | Positive since at least 2015 | PASS | | F4 Dividend uninterrupted | 5–20y | 30+ years (since 1994) | PASS | | F5 EPS growth >=33% | 10y normalized | ~100% est (7.17%/yr x 10y) | PASS | | F6 P/E <=15 (or <=20) | <=20 quality | Trailing 36.8x / Fwd 28.9x | FAIL | | F7 P/E x P/B <=22.5 | | 36.8 x 1.74 (TWD P/B) = 64x | FAIL |
Score: 5/7. Both failures are valuation failures caused by the +195% YTD re-rating — not structural/fundamental failures. Six months ago UMC likely passed all 7 at ~$8 price.
Nature of FAIL: Not value-trap (F3/F4/F5 all solid), not cyclical trap (shr-017 — earnings are recovering, not deteriorating). Pure re-rating: the thesis was correct, the entry window closed.
| Growth rate | V (trailing $0.63) | MoS | V (forward $0.80) | MoS | |--|--|--|--|--| | g = 0% | $5.36 | -77% | $6.80 | -71% | | g = 3% | $9.13 | -61% | $11.60 | -50% | | g = 5% | $11.65 | -50% | $14.80 | -36% | | g = 7.5% | $14.80 | -36% | $18.80 | -19% | | g = 10% | $17.96 | -22% | $22.80 | -2% |
Break-even growth implied by current price:
For a mature-node foundry facing Chinese overcapacity at 28nm+, sustaining 10–14% EPS growth perpetually is a demanding assumption. UMC's 5y historical EPS CAGR is 7.17% and the current cycle has EPS declining 55% from 2022 peak. Even the optimistic bull case (fwd EPS g=10%, MoS = -2%) barely justifies current price.
| Category | Signal | Color | |--|--|--| | Price action | +195% YTD, at 52w HIGH | RED — no margin of safety | | Earnings trend | Q1 2026 strong beat, H2 ASP hike | GREEN | | Insider direction | No NYSE Form 4 data; TWSE data unavailable | NEUTRAL | | Analyst revisions | May improvement, but stale ADR targets ($10–12) vs $23 price | RED — mean target -55% | | Dividend status | Maintained, 30+ year streak | GREEN | | Balance sheet | Net cash $1.6B, CR 2.72, strong | GREEN | | Short interest | 1.15% of float — negligible | NEUTRAL | | Dilution | None detected | GREEN | | China risk | Structural overcapacity at 28nm+, not yet fully resolved | AMBER | | Geopolitical (Taiwan) | Persistent tail risk, not new | AMBER | | Valuation vs IV | Trading at or above all but the most optimistic Graham IV | RED |
Three REDs: price action at peak, analyst targets ~55% below current, and Graham IV requires perpetual 10%+ EPS growth to justify entry.
RED — DO NOT ENTER AT CURRENT PRICE.
The thesis was correct: UMC was a clean Graham profile foundry at $6–8 (Aug–Nov 2025). The opportunity has been captured by the market. At $23.155 the stock requires 10–14% perpetual EPS growth to justify Graham IV — an extreme ask for a mature-node foundry facing structural Chinese competition at the commodity 28nm layer. Five Graham filters pass (balance sheet, earnings continuity, dividend, size) but the two valuation filters (F6, F7) fail decisively, not due to cyclicality but due to a completed 3x re-rating. China overcapacity risk is structural and unresolved; UMC's 22nm specialization partially hedges it but does not eliminate it.
WATCHLIST trigger: Re-enter watchlist if price retraces to $10–12 (where analyst consensus now sits and Graham IV at 7.5% growth = $14.80 provides meaningful margin of safety). That implies a -50% drawdown from current — possible if the 22nm cycle disappoints or China ASP pressure accelerates. Alert: add pm watch add UMC below 12.00.
VERDICT LINE: UMC's Graham profile is structurally sound but the entire trough opportunity closed with the +195% YTD re-rating — the stock now requires perpetual 10%+ EPS growth to justify its $23 price, making it a WATCHLIST re-entry at $10–12 only.