United Microelectronics (UMC) Q4 2025 earnings review
Steady Execution in a Slow Lane
UMC delivered a solid Q4 beat with revenue hitting TWD 61.8B (+4.5% QoQ) and Gross Margins recovering to 30.7%, driven by a record surge in 22nm specialty products. However, the investment case is hamstrung by its market exposure. While the broader semi industry rides the AI wave (projected mid-teens growth), UMC's addressable mature market is forecasted to grow only in the 'low single digits' for 2026. Management is executing well—utilization is high (78%) and pricing is firm—but they are effectively the best house in a slow-moving neighborhood. Q1 guidance suggests a seasonal step back, with margins compressing again due to depreciation headwinds.
🐂 Bull Case
The strategy to pivot away from commodity nodes is working. 22nm revenue surged 31% sequentially in Q4, reaching >13% of total sales. This node commands better pricing and is sticky with customers (Wi-Fi, DTV), insulating UMC from the fiercest price wars.
Despite Chinese competition, ASPs remained firm in Q4 and are guided to remain firm in 26Q1. Management noted 'disciplined pricing' and mix optimization, dispelling fears of a race to the bottom in legacy nodes.
🐻 Bear Case
UMC's addressable market (TAM) is forecasted to grow only 'low single digits' in 2026, compared to 'mid-teens' for the wider industry and 'low 20s' for foundries generally. They are structurally excluded from the high-growth logic/memory boom until 2027+.
North American revenue contribution dropped to 21% in Q4 (down from 25% in FY24). While partly due to diversification, losing share in the highest-value region raises concerns about geopolitical decoupling or client losses.
⚖️ Verdict: ⚪
Neutral. UMC is a well-run company maximizing a low-growth hand. While the yield play and 22nm execution are excellent, the lack of immediate exposure to high-growth AI markets caps the upside until the Intel/12nm collaboration bears fruit in 2027.
Key Themes
22nm Node is the Engine
Accelerating. 22nm/28nm combined revenue reached 36% of sales, but 22nm specifically is the breakout star, growing 31% QoQ to a record high. This is not just volume; it's a mix shift that supports margins. Management expects double-digit growth for this platform in 2026, driven by display drivers and Wi-Fi logic.
Structural TAM Lag
Management explicitly quantified their growth ceiling: while the foundry market grows ~20% in 2026, UMC's specific addressable market will only grow 'low single digits.' This gap confirms that UMC is currently operating in the stagnant portion of the semiconductor cycle (consumer, commoditized legacy), waiting for 12nm and silicon photonics to expand their TAM in 2027.
Depreciation vs. Margins
Depreciation costs are rising ('low teens' increase in 2026) as new capacity comes online. This creates a headwind that offsets operational efficiency. Q1 gross margin is guided down to the 'high-20s' largely due to these higher costs, despite stable pricing. Peak depreciation is expected in 2026/2027.
CapEx Discipline
UMC is tightening its belt. 2026 CapEx is guided to $1.5 billion, down from $1.6 billion in 2025 (and $3.0B in 2024). With cash exceeding TWD 110 billion ($3.4B USD), this lower capital intensity sets the stage for potentially higher dividends or buybacks, although none were explicitly announced.
Geographic Mix Shift
North America's revenue share fell to 21% in Q4. Management framed this as a shift toward Asian consumer demand (Wi-Fi, DTV), but a structural decline in North American exposure could indicate 'friend-shoring' headwinds where US clients prefer non-Taiwanese/Chinese sourcing, or simply stronger recovery in Asian end-markets.
Other KPIs
Stable. The balance sheet remains a fortress. Cash exceeds TWD 110B, providing ample buffer for the $1.5B CapEx plan and dividend payouts. Net cash position is robust.
Decelerating. Down from NT$3.80 in FY24. While the second half of 2025 showed recovery, the slow start to the year dragged down full-year profitability.
Accelerating. Up from 29.8% in Q3 and 28.7% in Q2. This was the high water mark for the year, driven by the 78% utilization and 22nm mix, though guidance suggests this trend reverses in Q1.
Guidance
Stable. Following a 4.5% revenue rise in Q4, flat shipments indicate a pause in the recovery momentum, attributed to normal seasonality.
Decelerating. A step down from the 30.7% achieved in Q4. Management attributes this to higher depreciation costs rather than pricing pressure, as ASPs are guided 'firm'.
Decelerating. Down from 78% in Q4. This implies a slight pullback in production volume or capacity additions outpacing demand temporarily.
Stable/Positive. Management expects a better pricing environment in 2026 than 2025, driven by mix optimization. This contradicts fears of legacy node commoditization.
Key Questions
North America Decline
North American revenue contribution dropped to 21% in Q4. Is this purely a function of Asian consumer electronics recovery, or are you seeing structural share loss at key US customers due to geopolitical sourcing mandates?
Closing the TAM Gap
You forecast your addressable market growing only 'low single digits' in 2026 while the foundry market grows 20%. Beyond the 2027 12nm ramp, what specific bridge exists in 2026 to help you outperform this stagnant TAM?
Dividend Sustainability
With CapEx falling to $1.5B and a cash pile of TWD 110B, but earnings per share down YoY, how should investors think about the payout ratio for the upcoming dividend?