Besi: The Packaging Monopoly at the Heart of the 3D Chip Era
The advanced packaging champion is the highest beta, most AI levered name in the European semicap complex
Welcome to the 72nd investment case, 58th Resilience & Quality, 46th Reshoring & Sovereignty, 12th China, 34th AI & Technology idea a on Crack the Market (and the most comprehensive Besi investment case you will find online)! Join me as I dissect one of the purest pick and shovel monopolies in the AI build, one of the best compounders in semiconductor equipment, the company that owns the single process the three dimensional chip era cannot do without, and arguably the highest beta, most direct listed vehicle for the decade long capex supercycle driven by AI compute, chiplet logic, the logicification of HBM and the migration of assembly from the back end into the front end.
This piece sits on top of the structural work already published on Crack The Market:
For the industry context, see the two primers, on Semiconductors and Semiconductor Equipment.
For the EU Semicap basket framing, see Six Monopolies, One Supercycle: EU Semicaps, where Besi is the highest beta of the six chokepoints. Other very high quality EU Semicap names leveraged to this AI theme and complexification of chip manufacturing are ASML, ASM International and Technoprobe.
For the demand side, The European Photonics Supercycle and Optical and Networking Supercycle at the Heart of the AI Revolution explain the co-packaged optics opportunity that has become Besi’s biggest upgrade driver.
Twice a week, I will release deep dives into stocks and sectors that fit into the 6 themes that I see winning in the coming years and decades: AI & Technology, Electrification & Energy, Reshoring & Sovereignty, Healthcare & Longevity, China & Asia, Resilience & Quality.
Take advantage of this once in a generation opportunity to build long term wealth by investing in great stocks that will deliver returns for your portfolio for years to come.
After reading this article, you will understand why Besi stands out as the most levered way to own the back end revolution in semiconductors, how it is powering the shift from solder to direct copper to copper hybrid bonding, the once in a generation change in how chips are joined, and how it is capturing logic, co-packaged optics and the coming wave of HBM stacking to deliver structural outperformance versus the broader assembly market. You will understand exactly what it does, the bonders that fuse chiplets within tens of nanometers of their target in front end grade cleanrooms, and how it is hitting an inflection point where physics driven hybrid bonding necessity meets exploding AI packaging demand to unlock superior growth and margins. Most importantly, you will grasp why this Dutch advanced packaging champion has a secret weapon in the form of its Applied Materials partnership and over 90% share of die to wafer hybrid bonding, positioning it to ride the complexity supercycle, and exactly where the thesis breaks if HBM adoption slips.
In this article I go through:
Besi’s business, its history of disciplined consolidation, and how it came to own the highest value corner of semiconductor assembly
Why hybrid bonding is a physics driven chokepoint rather than a product cycle, and the precise technology, tool by tool, that makes it Besi’s
The Applied Materials partnership and the front end to back end convergence that turned a back end toolmaker into a front end supplier
How logic, HBM and co-packaged optics each independently expand the bonder market, and the unit and ASP math behind the 2030 targets
Where the thesis breaks: HBM incumbency, the wafer to wafer threat, Chinese and Korean entrants, customer and China concentration, and the cycle
Besi Investment Case
Table of content
Why now
There is a quiet revolution underway in how the most advanced chips in the world are built, and it has nothing to do with making transistors smaller.
For sixty years the entire semiconductor industry organized itself around Moore’s Law, the relentless shrinking of the transistor. That engine is sputtering. The cost of each new node is exploding, the physics is getting harder, and the incremental performance is getting smaller.
So the industry has changed strategy.
Instead of cramming ever more onto a single monolithic piece of silicon, it now breaks the chip into pieces, chiplets, and reassembles them, side by side on an interposer and stacked directly on top of one another, into dense three dimensional structures.
The value in a leading chip is migrating from the transistor to the interconnect, from the front end to the back end, from how the silicon is printed to how the pieces are joined.
Besi is the company that does the joining at the most demanding end of that shift. More precisely, it owns the single technique the three dimensional era cannot do without: hybrid bonding, the direct fusion of copper pad to copper pad with no solder in between.
This is the process behind AMD’s accelerators, Apple’s M5, TSMC’s SoIC, the next generation of high bandwidth memory and the co-packaged optics now shipping inside Nvidia and Marvell switches.
Besi holds an estimated share above 90% of the die to wafer hybrid bonding market, and it is the only supplier shipping the tool in high volume manufacturing today.
Applied Materials, the largest equipment company in the world, thought enough of that position to buy roughly 9% of the company and to co develop the integrated production line with it.
Besi was the most explicit upcycle story I heard when meeting with Semicap companies last month (which I describe in detail in my EU Semicap piece).
At its June 18 Investor Day, management did not reset the targets down, it pushed them up.
The catch, and there is always a catch, is that the market knows. The stock sits near an all time high around €320, having compounded 155% over twelve months, and trades at one of the richest multiples in global equipment.
Three things make this the moment to write the case in my view rather than a year ago or a year from now.
First, the cycle has turned: the new upcycle began in the second half of 2025, the industry has moved from surplus to deficit, and orders, revenue and profitability have accelerated for several quarters, with first quarter 2026 orders up more than 100% year on year.
Second, the Investor Day was a genuine positive catalyst (though largely anticipated), raising the long term model, with co-packaged optics emerging as a brand new growth vector that did not exist in the model two years ago.
Third, the proof points are accumulating in the real world (the most important factor in my view): AMD, Intel and Apple are in production, TSMC’s COUPE is in high volume manufacturing for CPO, and the memory makers are publicly committing to hybrid bonding for HBM4e and HBM5. The thesis has moved from promise to evidence, which is precisely the phase in which a high quality compounder commands its richest multiple (which it now does).
The bear case is equally clean, and it is not foolish:
You are paying a peak multiple for a company whose largest future market, high bandwidth memory, is still dominated by lower cost thermo compression incumbents deeply embedded with the memory makers. Adoption timing in memory is genuinely uncertain and has slipped before.
Chinese and Korean challengers are circling, and management itself counts more than ten announced hybrid bonders.
There is a parallel wafer to wafer path that bypasses Besi’s die to wafer toolset entirely.
Revenue is concentrated in a handful of customers and heavily skewed to Asia, with a large and politically exposed China share, in an industry that has been violently cyclical for fifty years.
This is the highest beta, most AI levered name in the European semicap complex. The question is no longer whether hybrid bonding is real. It is how much of a decade of adoption you are being asked to pay for today, and where exactly the thesis breaks. The honest answer, developed across this deep dive, is that the entire debate compresses into one variable, the pace of hybrid bonding adoption in memory, and that this variable moves both the estimates and the multiple at the same time.
Business Description
BE Semiconductor Industries (Besi) is a Dutch headquartered supplier of semiconductor assembly equipment, the tools that take individual chips off the wafer and connect, protect and package them into finished devices. The company’s machines perform a deceptively critical function. They join a die to other electronic components, a substrate, an interposer or another die, while shielding it from the outside world, and they do so with a precision that increasingly determines whether the finished chip works at all.
Founded in 1995, headquartered in Duiven in the Netherlands and listed on Euronext Amsterdam under the ticker BESI, Besi employs roughly 2,200 people split about 75% across Asia and 25% across Europe and North America. Product development sits mainly in Europe, in Switzerland, Austria and the Netherlands, and in Singapore, while manufacturing runs out of Malaysia, with new capacity being built in Vietnam and an existing presence in China and the Philippines.
To place the company correctly, it helps to walk the full flow of building a leading chip.
The front end fabricates the transistors on the wafer through hundreds of lithography, deposition, etch and cleaning steps.
The wafer is then probed and tested, diced into individual chips, and handed to the back end, where the chips are sorted, placed, bonded, encapsulated, plated and tested again before shipping.
If you picture the process as constructing a building, lithography draws the plans, deposition and etch build the floors, test inspects the work, and assembly is the step that joins the finished modules and wires them together.
Besi sells the machines that do the joining and wiring at the most precise, highest value end of that assembly step.
The single most important strategic fact about the company, which the rest of this note develops, is that for the most advanced devices this back end joining step is migrating into the front end, blurring a boundary that has held for decades and pulling Besi up the value chain with it.
History - four acquisitions and one founder:
Besi’s history is a study in patient, disciplined consolidation around a single idea: own the die placement step and move it steadily up market toward higher precision and higher value. The company began in 1995 and was built out through four targeted acquisitions over the following fifteen years, each adding a specific capability rather than buying scale for its own sake. Through all of it, one person, the charismatic founder and chief executive Richard Blickman, has run the company since inception. The deals that built today’s portfolio are set out below.
The four acquisitions that built Besi
Source: company reports, Crack The Market.
The brands survive inside the product lines today: Datacon and Esec in die attach, Fico in packaging, Meco in plating.
The acquisitions were deliberately capability led rather than empire building, which is why Besi has remained a focused specialist rather than a diversified equipment conglomerate.
It is engaged, in its own words, in one line of business: the development, manufacturing, sale and service of semiconductor assembly equipment.
The criteria it applies to acquisitions, increasing process technology leadership, profitably increasing presence in the highest potential assembly markets, improving the efficiency of its Asian manufacturing, and growing the less cyclical service and spares revenue, are exactly the criteria that have shaped the current franchise. M&A has been rare and small, and the company has grown overwhelmingly organically since 2009, funding research from cash flow and returning the rest to shareholders.
Besi reports across three product groups:
The dominant one is Die Attach, at roughly 80% of revenue, which covers the full range of die bonding processes: epoxy, flip chip, soft solder and multi module die bonding for the mainstream, plus the advanced techniques that matter most for this thesis, hybrid bonding, thermo compression bonding, embedded bridge attach, die lid attach and fan out wafer level packaging. \
Packaging, under the Fico brand, is about 17% and covers conventional, ultra thin and wafer level molding, trim and form, and singulation. Plating, under the Meco brand, is about 3% and covers tin, copper, precious metal and solar plating with related process chemicals.
Spares and services run at roughly 15% of revenue and are a less cyclical, structurally growing layer that scales with the installed base rather than with the capex cycle, a mix feature that becomes more valuable as the hybrid bonding installed base expands.
Besi has grown in steps through successive cycles, holding gross margin above 60% since 2021.
The four principal end markets are computing, mobile, automotive and industrial, which together made up roughly 85% of 2025 revenue, with spares and service the remainder.
Computing has become the largest and fastest growing of the four, driven directly by AI accelerators, custom ASICs and the data center build, and it now sits at around half of revenue.
Mobile, historically the largest swing factor in the up cycles, has shrunk as a share but is reviving on AI enabled phones and rising packaging complexity per device.
Automotive and industrial are the more classically cyclical, mature node exposed markets, and they are the laggards now showing early signs of recovery.
The breadth matters: it means Besi is not a pure AI story bolted onto a declining base, it is an AI story on top of a broad mainstream franchise that is itself turning up.
Customers - the AI supply chain, plus the mainstream:
The customer list reads like the semiconductor & AI supply chain itself: TSMC, Intel, Nvidia, Micron, STMicroelectronics, Texas Instruments, Amkor, ASE, Fabrinet, InnoLight, LG Innotek, Infineon and the major Korean and Chinese assembly houses. Customers fall into two structural groups.
Integrated device manufacturers, such as Intel, Infineon and STMicroelectronics, buy Besi’s equipment for internal use in their own fabs.
Foundries and subcontractors, such as TSMC, ASE and Amkor, buy it to assemble packages for third parties on a contract basis.
In many cases Besi’s tools represent a significant share of a customer’s installed assembly base, which creates switching costs and a recurring service relationship.
The table below maps Besi’s main die attach systems to the customers and applications they serve, and it is worth studying because it shows how broad the franchise is beneath the hybrid bonding headline.
Besi’s die attach systems mapped to customers and applications
Source: company and industry disclosures, Crack The Market. Representative, not exhaustive.
A small market that Besi dominates at the top:
The assembly equipment market is small, roughly $5bn in 2025, around 4% of total semiconductor manufacturing equipment spend, which itself is heading toward $250bn.
That headline understates the prize, because Besi has deliberately concentrated on the highest value sliver of it rather than the whole. Within the roughly $1.3bn it considers addressable, the company holds about 43% of the die attach market, around 70% of advanced die placement below 7 micron accuracy, an estimated share above 90% in die to wafer hybrid bonding, and roughly 60% in the small plating segment.
Approximately 70% of 2025 revenue came from advanced packaging, about 60% from leading edge devices below 7 micron placement, and management estimates around half of 2025 orders were for AI applications.
This is the strategy in a sentence: do not chase volume in a commodity, own the structural growth segment of a small market and ride its mix relentlessly upward. The share data shows the strategy working. Over 2019 to 2024 Besi lifted its share of its addressable market from roughly 26% to 38%, with die attach share rising from 31% to 43%, even as the addressable market itself grew.
Besi’s addressable market and share by segment
Source: company disclosures, industry data, Crack The Market. Shares are approximate.
Besi has steadily taken share of its addressable market, led by die attach.
