Earnings call transcript: ASMPT posts record bookings in Q3 2026 on AI demand

ASMPT reported Q3 2026 revenue of $669.2 million, up 48.7% year over year, record bookings of $1.13 billion, and adjusted EPS of HK$1.65, up 489.3%; management said revenue and adjusted EPS exceeded consensus. Adjusted gross margin rose 517 basis points year over year to 42.9%, and Q4 revenue guidance of $670 million–$730 million exceeds current market consensus, with the midpoint implying 4.6% sequential growth. AI infrastructure demand drove orders, while management expects Q4 SEMI revenue to be flat and bookings below the Q3 record; shares were unchanged at $121.2.

Analysis

ASMPT’s key asymmetry is the gap between order momentum and revenue realization: advanced-packaging lead times defer conversion, so the record backlog is evidence of demand, not near-term earnings certainty. Over the next 1–3 months, watch whether Q4 delivery and gross margin hold up; over 2–3 quarters, bookings need to ship without cancellations or slippage to validate the earnings step-up. Mix and utilization can support margins, but SMT’s unusually strong orders include bulk business and pull-ins, making extrapolation risky.

The less obvious competitive implication is that logic-led TCB demand may partly offset delayed HBM investment, while also broadening the contest with established bonding-equipment suppliers such as Besi and Kulicke & Soffa. Hybrid-bonding qualification could create a second-source opportunity, but it is not yet volume revenue. Near-term photonics growth is more plausibly tied to pluggables/NPO than CPO; treating CPO as an imminent earnings driver would overstate the catalyst. SMT’s momentum also weakens the case for assuming a near-term divestiture, though it may increase strategic optionality.

Contrarian read: a high book-to-bill can reflect lead-time-driven ordering and timing shifts as well as end-demand strength. Management’s 2027 confidence is a catalyst, not independent proof. Thesis weakens if bookings retreat beyond the expected normalization, Q4 guidance is missed, advanced-packaging shipments fail to convert within the stated delivery window, or margin gains reverse. Consumer weakness and HBM timing remain downside offsets.

Key Decisions for Investors

• Build a measured long in ASMPT rather than chase the order print; add only if Q4 execution confirms the guide and subsequent SEMI revenue begins converting the advanced-packaging backlog. Keep size modest until order quality and delivery timing are clearer.
• For a relative-value expression, consider long ASMPT against a diversified semiconductor-equipment exposure, not an unhedged AI-equipment basket. The thesis is its TCB and photonics mix; the hedge limits broad cycle risk. Reassess if bookings normalize sharply or margins contract.
• Track quarterly SEMI bookings-to-revenue conversion, cancellations/push-outs, and gross margin, plus qualification milestones for hybrid bonding and customer ramps for NPO. Treat panel-level packaging and CPO as longer-dated optionality, not a basis for near-term earnings estimates.
• Do not underwrite a near-term SMT sale or assign a premium to management’s 2027 outlook without evidence. A guide miss, deterioration in SMT volume leverage, or continued HBM delays would invalidate the constructive view.