The foundry industry is growing mid-teens, but their addressable market is growing low-single digits. That is a confession, not guidance.
Thesis: UMC is effectively demoting itself from the AI structural growth story. By explicitly stating their addressable market growth (low single digits) lags the industry (mid-high teens) by over 1,000 basis points, they are admitting they don't have the kit for the actual party. The 22nm story is a fine defensive moat, but with depreciation ramping 20%+ through 2027 and ASPs bleeding, margins are mathematically compressed. This is a yield play, not a growth play.
Verdict: HOLD — Conviction: MEDIUM
Catalyst: H2 2025 22nm ramp. Management pinned the recovery on this specifically filling the growth pipeline.
Key Risk: Depreciation growing at 'high 20%' in 2025 while revenue grows low single digits implies significant gross margin compression unless utilization surprises to the upside.
The Tell: Jason Wang admitting 'no plan to expand' silicon interposer capacity immediately after a scripted monologue about 'broadening' advanced packaging offerings. It reveals their 'AI strategy' is niche hunting, not volume participation.
Friction Level: MODERATE_FRICTION — Bulls see 22nm/packaging as a bridge to growth; Bears see a structural demotion where UMC serves only the commoditized overflow TSMC doesn't want.
Report not found
The report data is no longer available. Please return to the archive.