Management is betting on physics: the market leader is doubling 2.5D logic capacity, implying a 4x surge in HBM demand that hasn't actually been ordered yet.
Thesis: Onto Innovation is a derivative call on TSMC's CoWoS expansion bottleneck. The thesis relies on physical constraints: you cannot double 2.5D logic capacity without quadrupling HBM density. The market is pricing in the current 'muted' HBM order book, ignoring the lead-lag effect. With gross margins already hitting 55% on efficiency alone, the eventual volume inflection creates non-linear earnings power. The risk is timing, not structure.
Verdict: LONG — Conviction: MEDIUM
Catalyst: Resumption of HBM orders to match the committed 2.5D logic capacity expansion, likely in 1H 2025.
Key Risk: Nvidia successfully qualifying Samsung HBM faster than expected, potentially altering the yield/inspection intensity assumptions embedded in the current SK Hynix/Micron dominance.
The Tell: When pressed on the HBM capacity model, CEO Plisinski retreated to theory: 'If the markets were a perfect equilibrium... That's just if everything was in equilibrium.' It's an admission that the current order book defies their internal logic models.
Friction Level: MODERATE_FRICTION — The timing of HBM orders. Bulls see mathematical inevitability based on CoWoS expansion; Bears see a 'Say/Do' gap where orders fail to materialize despite the narrative.
Report not found
The report data is no longer available. Please return to the archive.