Stock LLM Wiki

報告_Citi_聯電2303_20260910

更新 2026-09-12

PDF 原檔:報告_Citi_聯電2303_20260910_original.pdf

原始內容

citivelocity.com

10 Sep 2026 15:25:39 ET │ 12 pages

UMC (2303.TW)

Demand Inflecting Up

CITI'S TAKE

UMC attended our Global TMT Conference. The management sees a clear improvement in foundry demand versus the past two to three years, with utilization approaching ~90% and customer visibility extending through year-end, compared with only roughly two months previously. Customers are increasingly concerned about potential capacity tightness, partly because AI demand is absorbing capacity across the semiconductor ecosystem. UMC has implemented selective price increases in 2H26, mainly for incremental or rush orders, while the more important development is the ongoing 2027 contract renegotiation. Key catalysts ahead: 1. Better UTR and GM/ASP into next year; 2) Intel 12nm engagement; 3) silicon photonics expansion ; 4) advanced packaging.

Prepared for [讀者資訊已遮蔽]

Flash |

Buy Buy
Short-Term View: Upside, expires28-OCT-26 Short-Term View: Upside, expires28-OCT-26
Price (10Sep2613:30) NT$142.50
Target price NT$125.00
Expected share price return -12.3%
Expected dividend yield 2.5%
Expected total return -9.7%
MarketCap NT$1,791,238M
US$56,890M

Laura (Chia Yi) Chen AC

+886-2-8726-9090 laura.cy.chen@citi.com

See Appendix A-1 for Analyst Certification, Important Disclosures and Research Analyst Affiliations

Prepared for [讀者資訊已遮蔽]

  • 2027 Pricing Looks More Favorable. Management hopes to implement broader-based price increases across AI and non-AI products to offset higher wafer, energy and other input costs and support new capacity investments. Pricing should therefore turn more positive in 2027 if utilization remains high. Management noted that every 1% ASP increase could improve gross margin by roughly 0.5-0.6ppt. However, margins are unlikely to revisit the >45% COVID-era peak because depreciation is structurally higher following the current investment cycle.
  • Intel 12nm cooperation: Capital-Light Path, contribution starting in late 2027 . The Intel collaboration is strategically important because it gives UMC a capital-light path from its strong 22/28nm franchise into 12nm. Intel provides existing Arizona manufacturing equipment and capacity, while UMC contributes technology conversion, RD resources and its existing customer relationships. Initial customers are primarily Wi-Fi and communications-related, with the first meaningful revenue expected around late 2027; broader contribution should build gradually thereafter as customers validate the first wave. UMC expects to book the wafer revenue, with Intel manufacturing costs flowing through UMC's cost structure and economics broadly shared between the two parties. Expansion into smaller nodes or advanced packaging with Intel remains possible, but management stressed that 12nm must first demonstrate competitive yield, manufacturing performance and profitability.
  • Silicon Photonics: Singapore Fab Becomes a Major AI Growth Platform. The company is building dedicated 12-inch capacity for silicon photonics in Singapore, with the new facility expected to be ready in roughly 14 months. The platform can encompass PICs and related components such as TIA, with management describing demand as stronger than its available engineering resources. UMC also has exposure to TFLN through an affiliate focused initially on 6-inch technology, with potential migration to 8-inch manufacturing supported by UMC. Current economics are attractive because photonics wafers can command multiples of conventional CMOS wafer pricing. Longer term, UMC wants to combine silicon photonics, specialty materials and packaging capabilities into an integrated platform.
  • Advanced Packaging: Building an Open-Ecosystem Second Source. Advanced packaging represents UMC's third major structural growth driver. Rather than competing directly with TSMC's fully integrated leading-edge ecosystem, UMC aims to become a credible second-source, open-ecosystem provider. Its developing capabilities include interposers, wafer-to-wafer and hybrid bonding, DTC, and related packaging processes, with silicon photonics potentially integrated into the same broader solution. UMC is constructing two new advanced-packaging facilities in Taiwan. Construction is expected to take more than two years, while existing facilities can provide limited capacity during the interim.
  • Specialty Strategy Provides Defense Against China Competition. UMC continues to differentiate itself through specialty technologies rather than competing primarily on commodity mature-node capacity. Management estimates 50-60%+ of revenue now comes from specialty processes, which provides meaningful insulation from Chinese foundry pricing pressure. China's semiconductor market is also becoming increasingly segmented into 'China-for-China' and 'non-China-for-non-China' supply chains, limiting direct cross-border price competition. UMC's two China fabs remain highly utilized and represent roughly 10% of group capacity. Overall, UMC's strategy is shifting from being primarily a mature-node foundry toward a specialty and heterogeneous-integration platform while improving utilization and pricing provide a stronger cyclical backdrop for 2027.

Prepared for [讀者資訊已遮蔽]

Short-Term View on UMC (2303.TW)

Direction:

Upside

Duration:

Within 90 Days (expires 28 Oct 2026)

Date Added:

29 Jul 2026

Category:

Thematic driven

We see UMC's improving sales supported by higher UTR and pricing trend to support its share price performance in the coming months.

Prepared for [讀者資訊已遮蔽]

UMC

Valuation

Our target price for UMC of NT$125 is based on 15x 2027E EPS, which is higher than the midpoint of its 3-year forward PE. We believe our target multiple is justified by our belief that UMC should benefit from the robust demand in 22/28nm, TSMC's increasing order outflow opportunity, long-term opportunity in advanced packaging and photonics and that its niche technology offerings should help it outperform its trailing-edge foundry peers. Our target price equates to 2026E/27E PB of 3.5x/3.2x.

Risks

Key downside risks to our target price include: 1) a longer-than-expected demand recovery; 2) more competition from peers in 28nm; 3) worse-than-expected developments from the company's China investment; 4) a significant decline in smartphone end-demand; 5) poor contribution from Novatek's iPhone OLED business; and 6) fiercer ASP competition from Chinese peers.

UMC (2303.TW)

Analyst: Laura (Chia Yi) Chen

Date

Appendix A-1

2725 - 0ct-2314:44:18

1

*63.00

49.05

30-0ct-24 10:24:38