A chip-bonding machine earns its place in a factory by producing usable connections at an acceptable cost, repeatedly. Extreme precision is valuable, but it is not sufficient. The commercial contest also involves clean surfaces, throughput, reliability and the time required to qualify a complete manufacturing process.
That is the useful starting point for the latest debate over BE Semiconductor Industries, or Besi. The Next Web reported on October 6 that Bank of America downgraded the company to neutral, arguing that potential ASML entry into hybrid bonding was insufficiently reflected in its shares. This is a competitive-risk assessment. It does not establish that ASML has launched a competing commercial hybrid bonder or displaced a Besi order.
Precision has value only when it produces working chips
Hybrid bonding connects surfaces using both dielectric and metal bonding, allowing dense connections between stacked semiconductor components. The manufacturing attraction is closer integration, but achieving it requires more than positioning one die above another.
Imec's die-to-wafer research describes the importance of surface preparation, avoiding particles during singulation and placement, and combining accuracy with throughput. It also explains an advantage of die-to-wafer assembly: stacking known-good dies rather than committing indiscriminately to every die on a wafer. These are process considerations, not evidence of Besi's current commercial yield.
The financial mechanism follows from them. If a costly chip fails at assembly, the economic loss can include value accumulated in earlier manufacturing stages. A customer therefore evaluates output that passes its requirements, not simply the machine's nominal speed. Conversely, a technically excellent process can still be unattractive if it slows production or costs too much relative to an adequate alternative.
This creates two uncertainties for a supplier. Customers may adopt hybrid bonding later than an optimistic technology roadmap implies, and suppliers may compete more intensely once adoption broadens. Market growth and the share captured by any one company are separate variables. A large future market does not settle either the purchase timetable or the eventual equipment margin.
Packaging lithography is not a commercial hybrid bonder
ASML already has a confirmed position in advanced packaging. Its April 2026 AGM presentation, page 16, states that it shipped the first TWINSCAN XT:260, an i-line lithography system aimed at three-dimensional applications and advanced packaging.
Lithography patterns features; a hybrid bonder performs an assembly function. The product announcement therefore demonstrates that ASML is serving this broader manufacturing area, while leaving a narrower question unresolved: what competing bonding equipment will it actually offer, qualify and sell?
ASML's expertise in precision makes the analyst concern intelligible. It does not, by itself, provide a launch date, a customer acceptance result or an installed base of competing bonders. Forecasts that jump directly from relevant capability to lost Besi revenue skip those commercial stages.
The opposite dismissal would also be too easy. A credible potential entrant can affect negotiations, research spending and customers' willingness to commit before large-scale sales appear. Those effects are plausible scenarios, not measured outcomes established by the downgrade.
Besi can defend a process, not just a placement step
Besi has a concrete response in its existing ecosystem. On October 1, Applied Materials and Besi announced an expanded partnership, including Besi joining Applied's EPIC Center as an innovation partner. Their collaboration began in 2020 and combines materials-processing capabilities with die placement and assembly.
The announcement describes a wider development agenda spanning hybrid bonding and other interconnect architectures. Its relevance is the integration of steps such as cleaning, planarization, inspection and placement. A customer purchases a working flow, even when different vendors supply its components.
The analytical implication is that competition may occur between combinations of equipment and process knowledge, rather than between isolated placement specifications. A qualified integrated flow can create switching costs because changing one step may require re-establishing the performance of others. That is an economic argument grounded in process dependence, not proof of an unassailable moat.
Partnerships also carry limits. Development objectives do not guarantee customer orders, exclusive relationships or profitable production. More demanding integration can require additional spending. Besi must still show that cooperation creates customer value it can retain, rather than only expanding the technical scope of work.
Strong quarterly orders leave the adoption timetable open
The company's July 23 results reported second-quarter revenue of €249.9 million, orders of €292.9 million and a 65.7% gross margin. Management attributed growth to several applications, including hybrid bonding, photonics and data centres. These aggregate figures provide operating evidence; they do not isolate the economics of a future hybrid-bonding franchise.
Orders are encouraging because they indicate customer commitments, but delivery timing, product mix and subsequent purchasing still matter. The same release projected third-quarter revenue growth of 10–15% sequentially and a gross margin of 63–65%. Those were management expectations, not reported third-quarter results.
The strongest counterargument to a defensive interpretation is straightforward: relevant ASML capabilities could become a qualified competing offering, while customers' adoption schedules could disappoint independently of competition. Neither risk disappears because Besi recently reported strong growth.
Evidence of a commercially qualified rival tool, repeat customer purchases, or sustained changes in Besi's orders and product mix would materially change the assessment. Until then, the valuation debate is about how much confidence to place in future factory economics. The decisive unit is productive, profitable output for customers; the technology label alone cannot supply it.