BE Semiconductor Industries NV (BESI.AMS) AI OPINIONS & ADVISOR ANALYSIS

The bull case activates when physics stops cooperating with solder. If HBM5-class pad pitches force memory back to copper-to-copper bonding around 2028 and Besi holds its accuracy lead through qualification, the deferred order cohort returns on top of an already-ramping logic base. Operating leverage then compounds the effect — opex near 19% of revenue converts revenue upside almost entirely into profit — and management finally dates the EUR 2.2bn target. Earnings and multiple expand together, delivering returns well beyond the projected path.

The bear case activates when AI capex intent and competitive parity arrive in the same year. Hyperscaler pacing slows bookings in 2028 just as Hanwha or ASMPT clears qualification at a lead customer; Besi's order book, which sells into intentions rather than shipments, contracts first and hardest. Reverse operating leverage pushes opex back above 25% of revenue, earnings halve, and a stock still capitalizing an untimed target model loses both numerator and multiple simultaneously.

The sell-side treats hybrid bonding as inevitable: fifteen buys, zero sells, an average target of EUR 297.64 whose *lowest* estimate still sits 18% above spot [25]. The crowd anchors on record Q2 orders of EUR 292.9m [3] and reads the 43% drawdown from EUR 328 as an AI-pacing scare rather than a cycle tell. Memory's deferral is barely discussed.

The shares look modestly underpriced, though not for the reason the bulls give. Consensus anchors to a trailing 72x multiple and to memory adoption that JEDEC's height relief has pushed out a generation [12]; both distort the picture. Against a plausible 2027 earnings base built on logic, photonics and co-packaged optics — H1-26 already delivered EUR 434.7m revenue and EUR 140.6m net income [6] — the anchor price capitalizes roughly thirty times forward earnings for a sole-source franchise. Two consensus errors partially cancel; the residual favours the buyer.

The October 22 Q3 print is the test: management guided 10-15% sequential growth in a seasonally declining quarter [7]. A fourth consecutive record order intake above EUR 300m, with the hybrid-bonding customer count past 25, would prove run-rate rather than pull-forward and start the re-rating.