Management blamed a 'timing pause' for the TCB decline, then admitted their biggest growth engine is entirely dependent on customer roadmaps they don't control.
Thesis: ASMPT is an option on TCB adoption, not a fundamental owner of the trend. They are a capex outlet for oligopolies (TSMC, HBM makers) who dictate pricing and timing. The 'pause' in TCB revenue amidst an AI boom reveals their lack of leverage. While SMT provides a floor, the upside depends entirely on technical decisions made in Hsinchu and Suwon that ASMPT cannot influence.
Verdict: HOLD — Conviction: MEDIUM
Catalyst: Confirmation of HBM4 volume production orders or a definitive logic chip-to-wafer win (or loss) in 2H 2026.
Key Risk: Hybrid bonding acceleration displacing TCB at the 20-high stack generation before TCB fully scales.
The Tell: When asked about the substrate cancellation, Cher Ng interrupted to insist analysts 'De-link these 2 pieces... This cancellation has nothing to do with AI.' The defensive urgency suggests fear of contagion to the AI narrative.
Friction Level: HIGH_FRICTION — The timing and magnitude of the TCB ramp: Bull sees 2026 supercycle, Bear sees structural delay and hybrid bonding displacement.
Report not found
The report data is no longer available. Please return to the archive.