ASML’s Bonding Patents: The Evidence Behind the Besi Competition Thesis

AHNIM STUDIO | News analysis | October 16, 2026 (KST)

ASML’s possible entry into hybrid bonding returned to market attention after an October 15 SemiAnalysis investigation and October 16 Reuters coverage. The research firm points to a portfolio of bonding patent families. We opened an anchor patent and ASML’s product documentation to separate a credible engineering direction from an announced commercial product.

Our view: this evidence warrants tracking a potential extension of ASML’s role in advanced packaging. It does not yet justify adding hybrid-bonder revenue or subtracting a quantified amount of Besi’s earnings. The important investment question is whether an integrated process can deliver qualified output at a better cost—not simply whether ASML can describe a machine.

What the patent actually establishes

WO2025201788A1, Semiconductor bonding systems and methods, lists ASML Netherlands BV as the original assignee. The record separates a March 27, 2024 priority date, a February 27, 2025 PCT filing and an October 2, 2025 publication. The document was therefore public before this week’s reporting; the new event is renewed analysis of that evidence, not a patent filed on October 16.

The claims describe independently controlled substrate holders moving along a track through processing stations, including bonding and alignment. One claim specifies dies as the first substrates and a semiconductor wafer as the second. That is a real architectural disclosure relevant to die-to-wafer bonding. It is not a customer acceptance test, a production specification or an announcement of an order. [1]

SemiAnalysis counts 27 bonding patent families and argues that ASML’s alignment and motion expertise could support a die-to-wafer tool. That count and the competitive forecast are the research firm’s work, not an independently reconstructed AHNIM patent census. We reviewed the publicly accessible portion; its later discussion is subscriber-only. [2]

An advanced-packaging scanner is not a hybrid bonder

ASML already sells equipment serving advanced packaging. Its official TWINSCAN XT:260 page describes an i-line lithography scanner; its October 15, 2025 results release reported the first shipment. A scanner exposes patterns. The bonding architecture in the patent concerns joining substrates. The existing scanner is not proof that the proposed bonder is shipping. [3] [4]

This distinction also prevents a misleading performance comparison. Scanner wafers-per-hour, lithographic resolution, bonding placements-per-hour and placement accuracy are different measures. A scanner’s advertised productivity cannot be relabeled as the throughput of an unannounced bonding product. Nor does a claimed patent objective demonstrate an achieved manufacturing result.

Evidence	Verified scope	Still required
Published ASML bonding application	Track-based substrate handling and bonding concepts	Working performance, reliability and customer validation
XT:260 product and first-shipment disclosure	Advanced-packaging lithography business	Separate evidence for a commercial hybrid bonder
Research report and press coverage	A competitive thesis attracting attention	Company product terms, orders and financial contribution

AHNIM evidence review, October 16, 2026. Technical concepts and commercial milestones are not interchangeable.

How a successful entry could change the economics

ASML could benefit if tighter coordination between positioning, measurement and bonding lowers the cost of producing a usable chip. The economic gain would come from less rework, fewer lost dies, higher utilization or a more productive process. Customers would compare those gains with equipment price, maintenance, consumables, footprint and the disruption of changing an established process. These are AHNIM analytical conditions, not reported ASML performance.

Integration could also matter even without displacing every incumbent tool. A vendor that improves the whole process may capture value through software, metrology, service or complementary equipment. Conversely, a technically impressive prototype may remain commercially unattractive if its cost, uptime or process compatibility falls short. The relevant market is not simply the number of chips that could theoretically be stacked.

Before revenue, a development program can require research expenditure, engineering staff and evaluation systems. Customer qualification can extend the interval between that spending and cash receipts. Early shipments can also have different margins from a mature installed base. We therefore would not value hypothetical sales at ASML’s existing group margin or assume that its lithography franchise automatically transfers to bonding.

Besi has an existing ecosystem, not an empty position

Applied Materials and Besi announced an expanded partnership on October 1, with Besi joining Applied’s EPIC Center as an Innovation Partner. Their release links the existing collaboration to the Kinex integrated die-to-wafer system and describes further development across bonding and interconnect technologies. These are the partners’ statements and forward-looking research plans where indicated. [5]

For Besi, a credible rival could affect expected market share, pricing or the research spending needed to defend its position before it affects reported sales. Yet competition can coexist with market expansion and complementary process roles. Customers may prefer multiple suppliers or retain a qualified process for longer than investors expect. A forecast of lost share needs evidence about the particular application and customer, not an automatic transfer of the entire opportunity to ASML.

Our valuation decision and the broker chronology

Reuters reported that ASML did not confirm the development report and declined to comment on it. This limits what we can call company-confirmed news. [6] The Next Web’s October 6 article had already described a Bank of America downgrade of Besi to Neutral over possible ASML competition. That coverage predates the October 15 research report; we did not establish its complete old-to-new target pair or prior rating from the opened article and do not manufacture those details. [7]

We did not verify a new ASML broker-target change specifically attributable to this patent report. Other earnings-related calls must not be presented as reactions to it. AHNIM has not established an ASML point target in this article, so a comparable broker-average-versus-AHNIM Base calculation is unavailable.

Our ASML Q3 analysis remains the operating baseline: equipment demand must convert into deliverable capacity, recognized earnings and cash. The bonding evidence broadens the strategic opportunity to monitor but does not supply a verified new earnings input. This is a new research branch, not an overwrite of the earlier analysis.

What would turn the thesis into financial evidence

We would look for a named product, demonstrated performance under disclosed conditions, customer evaluation or qualification, an order and shipment schedule, and ultimately a separable financial contribution. For Besi, customer decisions and realized pricing would help distinguish competitive pressure from an expanding market. Until those milestones appear, the disciplined conclusion is a credible direction of travel with substantial execution uncertainty—not a completed market-share shift.