AlphaLens
Research
研报5月12日 · Morgan Stanley

Greater China Semiconductors: Raising RF semi PTs on new business, share gains

中文EN⚠ quality lint: see notes

Raising PTs for WIN Semi and AWSC on Diversification, but Underweight on Limited AI Upside

Price targets for WIN Semi and AWSC are raised on stronger-than-smartphone revenue growth in 2Q26, driven by new businesses and market share gains respectively. However, we maintain Underweight ratings: investor expectations for WIN Semi’s InP business appear excessive, and AWSC’s long-term growth is capped by heavy smartphone concentration and limited AI exposure. The key swing factor is China’s InP substrate export ban, which could cap photodiode shipments in 2H26.

Why the Upgrades — and Why We Stay Underweight

WIN Semi: short-term momentum from AI datacenter and satellite, but InP hype is premature.
Revenue in 2Q26 could rise by teens % Q/Q despite China smartphone weakness (contrast with declines at MediaTek and OmniVision). Utilization rate has lifted to 60–70% in 1H26 from ~50% six months ago, driven by AI datacenter (mid-single-digit % of 2026 revenue) and satellite (mid-teens %). Capex in 2026 is guided above NT$2bn for InP and GaN capacity expansion.

Yet the InP business remains a long-dated story. Photodiode (PD) production for Broadcom starts small in 2Q26, ramps in 2H26, but InP substrate supply from China export controls could cap volume. Laser diode (LD) revenue is not meaningful until 2027/28 due to extended reliability testing. 6-inch InP production is unlikely before 2028–29. AI datacenter revenue still accounts for only mid-single digits of 2026 sales — the base case does not reflect mass optical outsourcing soon.

AWSC: share gains in Android smartphones drive near-term growth, but the runway is narrowing.
AWSC 2Q26 revenue should grow 10–15% Q/Q (vs. Q/Q declines at MediaTek and OmniVision), driven by share wins against WIN Semi and Sanan in China and Samsung. 2026 revenue is now forecast to grow >20% Y/Y, with UTR at 75% or above. Gross margin expands to 34% from 26.5% in 2025.

However, smartphone still represents >70% of revenue. The ability to push UTR above 80% is constrained by a shortage of engineers. Optical/AI exposure is low, and the total revenue CAGR over the next few years is limited compared to purer AI proxies. The stock at 3.0x 2027e book value is not cheap relative to those alternatives.

Key Risks and Where We Could Be Wrong

WIN Semi upside triggers:

  • Qualification with Lumentum or a major Japanese IDM for CW laser and EML.
  • Securing additional InP substrate supply, enabling PD shipments to exceed expectations.
    We would turn more constructive only on evidence of qualification and meaningful orders.

AWSC upside triggers:

  • Development of non-PA businesses (LEO satellite, 400G/800G datacom) growing faster than expected with capacity expansion.
  • A smartphone replacement cycle driven by Agentic AI/inferencing demand in 2027/28, given AWSC’s high share in China Android.

Downside risks:

  • China InP substrate export ban remains the single biggest swing factor for WIN Semi’s optical ramp.
  • Pricing competition from Chinese peers and Sanan could erode margins.
  • Smartphone demand could weaken further, affecting both companies.

Valuation Implications

WIN Semi PT rises to NT$300 (from NT$130), based on a residual income model with cost of equity 8.0%, intermediate growth 12%, terminal growth 3%. The PT implies 33x 2027e EPS. Bull case NT$650 (73x 2027e EPS) assumes faster optical adoption; bear case NT$175 (20x).

AWSC PT increases to NT$130 (from NT$85), residual income model with cost of equity 9.2%, intermediate growth 12%, terminal growth 2.5%, and dividend payout raised to 30%. The PT is 2.5x 2027e book value. Bull case NT$210 (4.0x); bear case NT$65 (1.2x).

Despite higher PTs, the risk-reward is skewed to the downside relative to purer AI plays within coverage. We remain Underweight on both names.