Management is tripling advanced packaging capacity by Q2 2024 while guiding Q4 revenue down 8% during what should be peak smartphone season.
Thesis: Amkor is the unglamorous spillover bucket for the AI trade. The Street treats them as a cyclical smartphone OSAT, missing the structural pivot. TSMC cannot handle all CoWoS volume; Amkor is the designated pressure valve. They are tripling 2.5D capacity by Q2 2024 with customer commitments. The 'fungibility' of their equipment—cited by bears as a lack of moat—is actually a hedge that protects the balance sheet if AI cools. Margins expanded 400bps QoQ, proving they have operating leverage when volume hits.
Verdict: LONG — Conviction: MEDIUM
Catalyst: Q2 2024 earnings, when the tripled 2.5D capacity comes online and utilization rates confirm the AI spillover thesis.
Key Risk: Utilization collapse. If hyperscalers pause ordering or TSMC brings more CoWoS-L in-house, that new capacity sits idle.
The Tell: The CEO's response to the 2.5D profitability question. Instead of claiming high margins on the new tech, he emphasized equipment 'fungibility' with standard bumping lines. He admitted, perhaps unintentionally, that the moat is low, but the safety net is high.
Friction Level: HIGH_FRICTION — The value of the 2.5D expansion. Bulls see a structural AI bottleneck release valve; Bears see commoditized assembly where TSMC owns the high-margin interposer.
Report not found
The report data is no longer available. Please return to the archive.