[SMHN] From Fixing Microscopes to Building AI: The Strange Story of SÜSS MicroTec
While everyone's watching trillion-dollar tech giants, this small German engineering firm quietly became essential to manufacturing AI chips
Origins
In 1949 Munich, while Germany was still clearing wartime rubble, Karl Süss started selling microscopes from his garage. He represented Leitz, the prestigious optical manufacturer (ever heard of Leica? That's Leitz Camera). Karl Süss built a reputation among research labs for his precision maintenance work. His sons Ekkehard and Winfried joined early - they were building more than a business, they were part of Germany's economic miracle, one repaired lens at a time. Japanese and American firms might have been racing ahead in electronics, but Germany still owned precision engineering.
The story shifted in 1962, when Siemens came asking for help. They needed someone who understood optical precision to build specialized bonding machines for transistors. The choice revealed how the semiconductor revolution actually happened: while big companies grabbed headlines, small specialized suppliers solved the hard engineering problems. Süss delivered.
A year later, they created their first mask aligner, the MJB3. Their repair shop expertise in optics had accidentally positioned them at the intersection of old-world precision and new-world semiconductors.
(A quick nerdy word about mask aligners, since they'll matter here: They're the machines that help print microscopic patterns onto silicon wafers - think of them as incredibly precise photographic enlargers, but in reverse and working at the scale of micrometers. If your alignment is off by even a fraction of a hair's width, your entire chip is useless. This is why optical precision became so crucial to semiconductors. I digress.)
The transformation played out from their base in Garching, near Munich, where they still operate today. Each new product - diamond scribers, testing systems, increasingly sophisticated mask aligners - emerged from that same foundation of precision engineering.
The 1975 launch of the MJB55, their first mass-production mask aligner, marked their arrival as serious players. Yet they kept their repair shop DNA: solve the customer's problem first, then figure out how to turn it into a business. When Karl and his eldest son Ekkehard passed away in 1994, they left behind more than a company - they had shown how traditional German engineering skills could evolve to serve an entirely new industry.
How do they make money
At its core, SÜSS MicroTec sells specialized equipment for making semiconductors. Think of it as selling precision manufacturing tools, but for an industry where precision means controlling things at the atomic scale. Their machines handle critical steps in semiconductor production - aligning microscopic patterns, applying coatings, and connecting different chips together. A single machine can cost millions.
These aren't commodity products. When semiconductor manufacturers like TSMC or Samsung qualify a piece of equipment for their production line, they're essentially committing to it. The qualification process takes months, sometimes years. Once qualified, switching to another supplier means requalifying their entire manufacturing process. This shows up in SÜSS's numbers: gross margins around 40%, and customers paying substantial deposits upfront to secure production slots.
The company operates through two main segments:
Advanced Backend Solutions (70% of sales) handles what happens after the basic chip is made - especially connecting different chips together.
Photomask Solutions (30%) maintains the ultra-precise stencils used to create chip patterns.
Both segments are seeing strong growth in Asia, where most advanced semiconductor manufacturing happens. Taiwan, China, and Korea generate over 85% of sales. The concentration makes sense: that's where the customers are.
The business is evolving with semiconductor technology. As chips get more complex, SÜSS's specialty in advanced packaging - basically, sophisticated ways of connecting chips together - becomes more valuable. This is particularly relevant for AI chips, which need massive amounts of memory connected at high speeds. SÜSS holds over 50% market share in temporary bonding equipment among leading memory manufacturers. They're maintaining this position through significant R&D investment (10% of sales) and strategic partnerships with semiconductor research institutions.
The company is also shifting its manufacturing model, building relationships with specialized suppliers to handle entire modules rather than individual components. This suggests they're preparing for sustained growth while protecting their technical edge.
Who are their customers?
Looking through the filings, it's an interesting puzzle. The company doesn't explicitly name their customers (standard practice in this industry - everyone's secretive), but we can piece it together:
For memory chips and advanced packaging (their biggest segment), the universe of potential customers is actually quite small. Only a handful of companies make high-bandwidth memory: SK Hynix, Samsung, and Micron. When they mention having "more than 50% market share" in temporary bonding equipment among leading memory manufacturers, they're talking about these players.
On the Photomask side, they dropped an interesting detail: they're the only qualified supplier for 3nm photomask cleaning. That points straight to TSMC, since they're the only foundry currently running 3nm production. Intel and Samsung are getting there, but TSMC is the leader.
Reading between the lines of their geographic revenue (Taiwan dominance) and their strategic moves (new facility in Zhubei, Taiwan - conveniently close to TSMC's facilities), TSMC appears to be a major customer. The company's focus on building out Taiwan capacity suggests they're positioning for TSMC's massive AI-driven expansion.
You can also play detective with their order patterns. When they mention receiving "orders worth around €130 million in direct connection with AI," that aligns perfectly with the timing of TSMC's major HBM (high-bandwidth memory) expansion for Nvidia's AI chips. Though they'll never say it directly, that's probably who those orders were for.
Numbers
Start with the basics: The company made €304 million in revenue in 2023. That's about what a mid-sized regional bank makes. Or roughly a third of what Taylor Swift earned on tour last year. Scale is important - this is a relatively small company making very specialized equipment.
Their profit margins tell us more. They keep about 39% of what they sell as gross profit. Think of gross profit as what's left after you pay for materials and direct manufacturing costs. For a company making complex machinery, that's solid. It means they have pricing power - customers pay up for their technology.
Now here's a number that makes semiconductor people sit up straight: Their order book is €452.5 million. That's 1.5 times their annual revenue, sitting in confirmed orders. In most industries, companies can barely predict next quarter. Here's a company that knows what it's building for the next 18 months.
Speaking of building things - they're almost entirely focused on Asia. As we said, 86.3% of their sales come from Taiwan, China, and Korea. This isn't surprising if you know semiconductors. Most advanced chips are made in Asia. But it's a concentration worth noting. When Taiwan Semiconductor Manufacturing (TSMC) sneezes, companies like SUSS catch a cold.
And their special sauce: High-Bandwidth Memory (HBM). They have over 50% market share in machines that help make HBM chips.
Here's why this number matters so much: Let's talk about memory chips and AI. Modern AI systems are like extremely hungry data processors - they need to access massive amounts of information incredibly quickly. That's where HBM (High-Bandwidth Memory) comes in - by stacking multiple memory chips on top of each other, you create something like a microscopic apartment building of memory, which can feed data to AI processors much faster than traditional flat memory chips. SUSS owns half the market for the machines that handle these delicate chips during manufacturing - specifically, their temporary bonders hold and position these chips during processing. We're talking about a market worth billions as every major tech company races to build AI infrastructure. The really clever part? Once a chip manufacturer starts using SUSS machines, they rarely switch - qualifying new equipment in semiconductor manufacturing takes years and millions of dollars. And while competitors like ASM International are strong in permanent bonding, SUSS's dominance in temporary bonding for HBM gives them a critical position in the AI supply chain. Every time someone wants to build more AI processing capacity (think ChatGPT, cloud computing, autonomous vehicles), they need more HBM chips, which means they need more SUSS machines.
Their R&D spending shows they're serious about staying ahead: 9.6% of revenue goes to research. For comparison, Apple spends about 7% of revenue on R&D. In specialized manufacturing equipment, you innovate or die.
The balance sheet looks strong. €122.3 million in net cash. That's like having a year's worth of operating expenses in your checking account. Zero debt worth mentioning. They can weather storms or grab opportunities.
A subtle but important metric: Customer prepayments of €98.7 million. When customers pay you before you deliver, you have leverage. It means they really need what you're selling. It also means you're running the business on their money, not yours.
Some numbers raise eyebrows: Inventory has grown to €216.8 million, up from €166.7 million last year. That's a lot of parts and machines sitting around. The company says it's for expected growth. Maybe. But inventory is like milk - it has an expiration date.
Finally, a telling constraint: Only 18% of their suppliers are in Asia, despite 86% of sales being there. Supply chains are complex in normal times. They're even trickier when your suppliers and customers are on different continents.
People
SUSS MicroTec started fixing microscopes in Munich. Now it builds the machines that make AI chips in Taiwan. That pivot needs specific people - here's who they are.
Starting with ownership. SUSS MicroTec has no controlling shareholder - instead, it's owned primarily by European investment firms. The largest, Van Lanschot Kempen from the Netherlands, owns just under 10%. Traditional European institutional investors own about 45% of the shares, with the rest spread among smaller holders. This matters because it gives management significant freedom to operate, while maintaining solid European corporate governance standards.
The real story lives in who runs the company. CEO Burkhardt Frick joined in 2023 from ASML, the world's dominant semiconductor equipment maker. At ASML, he managed strategic sourcing and supply chains - exactly the expertise SUSS needs as it expands in Asia. His presence, along with former ASML executive Jan Smits on the supervisory board, brings deep semiconductor industry knowledge right when the company pushes into AI-related chip equipment. You might say SUSS hired the people who built the playbook.
The management team combines semiconductor veterans with fresh perspectives. The CFO comes from automotive supply chains, while the COO brings optical technology expertise from Carl Zeiss. Think of them as architects who've built different parts of the house before.
Look at the workforce numbers: SUSS grew from 937 to 1,414 employees in five years. The German operations maintain their engineering core - average tenure 9 years, deeply technical workforce. Meanwhile, rapid growth happens in Taiwan, where a new production facility in Zhubei shows where the company sees its future. For investors, this geographic split creates a useful natural hedge: European stability meets Asian growth.
This creates an interesting dynamic: European ownership backing a predominantly Asian business - 86% of sales come from Asia. It's like a German precision timepiece assembled in Taiwan.
Of roughly 1,400 employees, 856 work in production and development. In Germany, where we have detailed data, the average age is 42 and about 22% are women - typical for German engineering firms. The company keeps its technical DNA while growing globally.
Those new production facilities in Taiwan will test if this carefully assembled team can deliver on their global ambitions. After all, in semiconductor equipment, you're only as good as your next machine.
Competition & the Moat(?)
Competition in semiconductor equipment feels straightforward at first glance. Company A makes a tool, Company B makes a similar tool, they compete. Simple enough.
Except it's not. Süss operates in specific niches of semiconductor manufacturing, where competition takes different forms. Let me explain.
First, there are the direct equipment competitors. EV Group and Tokyo Electron make similar tools for temporary bonding - the process of sticking chips together temporarily during manufacturing. Think of it as high-tech masking tape, except it costs millions and requires perfect precision. These companies fight for market share in memory manufacturing, where Süss currently holds over 50%!
Then we have the photomask cleaning competitors like Lasertec and Screen Semiconductor. A photomask is like a stencil for printing chip designs. At advanced manufacturing nodes (3 nanometers and below), any tiny particle can ruin the whole process. Süss is currently the only qualified supplier at this level.
But there's also a different kind of competition: alternative technologies. TSMC, the world's largest chip manufacturer, develops its own packaging methods like CoWoS. And we have Intel with fluid alignment. Sometimes your competition isn't another company - it's a different way of solving the problem entirely.
Let me explain.
CoWoS (Chip-on-Wafer-on-Substrate) is fundamentally an architecture - a way of stacking and connecting multiple chips. Think of it as the blueprint for a building. SÜSS's temporary bonding equipment is more like the scaffolding needed during construction - it holds ultra-thin wafers in place during critical processing steps.
Current CoWoS manufacturing requires temporary bonding. When you're handling silicon wafers thinned down to 50 microns or less (about half the thickness of a human hair), you need extremely precise handling equipment. That's SÜSS's core technology.
The existential risk to SÜSS comes from radically different approaches to wafer handling. Intel's fluid self-alignment technology represents the most serious potential disruption. Instead of physically bonding wafers during processing, this technique uses surface tension forces to position and hold chips - imagine how water droplets naturally align themselves.
Key Implications:
CoWoS expansion is actually positive for SÜSS in the near term
Fluid self-alignment is still experimental but represents a legitimate long-term threat
Watch Intel's packaging technology announcements carefully
And zooming out from the nanometer world, there's geographic competition. Chinese equipment makers want to build domestic alternatives. U.S. companies want to establish local supply chains. Korean manufacturers might develop internal capabilities.
So what about Süss's moats?
The first moat is process power in advanced nodes. Getting qualified as a supplier for 3-nanometer chip manufacturing takes years of development, testing, and validation. One mistake can cost customers millions. Once you're in, you're in. Customers don't switch suppliers easily at these levels.
The second moat is switching costs in temporary bonding. When a memory manufacturer integrates your tools into their production line, they also integrate your processes. Their engineers learn your systems. Their quality control validates your output. Changing suppliers means revalidating everything. That's expensive and risky.
The third moat is customer captivity in specific ecosystems. Süss is deeply embedded in Taiwan's semiconductor manufacturing ecosystem, particularly with TSMC. They understand the local requirements, speak the language (literally and figuratively), and have built relationships over decades.
These moats aren't company-wide though - they're segment-specific. In hybrid bonding, where chips are permanently connected, Süss is actually playing catch-up. No moat there yet. In imaging and coating, they face more commoditized competition.
The strength of these moats shows in the numbers. Despite capacity constraints, margins are improving. The order backlog stands at €450 million - about 18 months of revenue. Customers wait in line for their tools.
A word about geographic concentration though: when 88.8% of your sales come from Asia, particularly Taiwan, your moats are geographically concentrated too. In the semiconductor industry, that's both a feature and a bug.
Mr. Market
Sometimes stock charts tell you a story. SUSS MicroTec's chart tells you several.
In early 2023, you could buy shares for around €20.
By late 2024, they touched €70.
Today, they trade at €41.
That's quite a ride. The rise made sense - SUSS makes machines that help produce AI chips. When the AI boom hit, companies rushed to secure manufacturing capacity. SUSS's order book swelled to €450 million, more than a full year of sales. Their expected earnings jumped from €4.7 million in 2023 to €104.7 million in 2024.
Think of it as being the company that makes the picks and shovels during a gold rush. Everyone needed their equipment, fast.
Then reality set in. The stock dropped 40% from its peak. What changed? The market started asking harder questions. SUSS's customers focused on installing the equipment they already ordered rather than placing new ones. Analysts began worrying about 2025. Even NVIDIA, the company powering much of the AI excitement, announced delays.
Where does this leave us today? At €41, SUSS trades at ~15x expected 2025 earnings. That's neither cheap nor expensive for a semiconductor equipment company. The market seems unsure what to make of it.
Look closer though, and some interesting patterns emerge. The company's gross margins have steadily improved, from 29.9% in 2019 to over 39% today. They're expanding production capacity in Taiwan, close to key customers. And that €450 million order book provides solid visibility well into 2025.
The market is wrestling with two competing stories. One says SUSS is just another cyclical semiconductor equipment maker, riding a temporary wave. The other sees a company transforming itself, grabbing a crucial position in the AI supply chain.
The truth? Probably somewhere in between. The market got ahead of itself at €70, pricing in perfect execution. At €41, it might be overthinking the risks.
Sometimes Mr. Market's indecision creates the most interesting opportunities.
Bear Thesis
The bear case for SUSS MicroTec centres on concentration. This is a company that commands over 50% market share in temporary bonding equipment - machines critical for making memory chips thin enough for AI applications. That's an enviable position right now. It might also be a trap.
Let's talk technology risk first. SUSS's bonding equipment has physical limitations. Their tools handle down to 10 microns. The industry needs sub-1 micron precision. Meanwhile, Intel and others are developing "fluidic self-alignment" techniques as we mentioned, promising 10x throughput improvements. When your core technology faces both physical limits and potential disruption, you pay attention.
The scale-up story raises eyebrows too. In November 2023, management announced they were expanding production in Taiwan, hiring aggressively, and leaning more on external manufacturing partners. All sensible moves when your order book hits €450 million. All incredibly complex in an industry where, as management notes, "training of people takes more time than on other tools."
Geography adds another layer. That new production facility? It's in Zhubei, Taiwan. SUSS generates 88.8% of sales from Asia Pacific already. They're doubling down on regional concentration just as U.S. politicians debate repealing the CHIPS Act and reshoring semiconductor production. Bold move.
Here's an interesting wrinkle: some companies are working to sidestep the whole issue. Microsoft-backed d-Matrix just launched AI processors that don't use traditional high-bandwidth memory at all. Early days, sure. But in semiconductors, architectural shifts tend to reshape the equipment landscape quickly.
The market prices SUSS at €42 per share, about 16x next year's earnings. That's actually a discount to peers. The question is whether it's enough of a discount given the concentration of technological, operational, and geographic risk. The bears think not, seeing fair value closer to €25-30.
In semiconductor equipment, dominance in one generation of technology doesn't guarantee relevance in the next. Just ask the companies that dominated equipment for DRAM production in the 1990s. Markets have a way of remembering this lesson at inconvenient times.
Bull Thesis
The bull case for SUSS MicroTec rests on a surprisingly simple chain of logic: AI needs HBM memory chips, and SUSS makes the machines that help build them. When Nvidia announces another AI breakthrough, somewhere a SUSS sales rep smiles.
The numbers tell the story. Sales jumped 46% while margins expanded from 29.4% to 39.0%. That kind of operational leverage makes CFOs write poetry. Their €430.8M order book provides visibility through 2025.
Memory makers understand the stakes. They're planning billions in HBM investments. Micron alone is dropping $7B on HBM capacity. The total market could hit $100B by 2030, larger than today's entire DRAM industry.
SUSS holds over 50% market share in their critical bonding technology. Getting qualified as a supplier takes months, sometimes years. It's the semiconductor equivalent of becoming a made man - once you're in, you're in.
They're doubling production capacity in Taiwan, reinforcing their position near the epicenter of advanced chip manufacturing. Smart move when your equipment needs more babysitting than a room full of toddlers.
For the quantitatively inclined: We're looking at a company with clear market leadership in a crucial semiconductor niche, demonstrating pricing power and operational leverage during rapid growth, trading at 15x forward earnings. The order book alone equals 54% of market cap. Sometimes the market hands you an asymmetric bet.
So what do we make of all this?
Here's what we've learned after spending way too much time thinking about a company that makes machines that make machines that make AI chips: The world of advanced semiconductor manufacturing comes down to solving incredibly specific problems with incredibly specific tools. SUSS MicroTec basically owns one of those problems - they're the go-to company when you need to temporarily stick memory chips together without breaking them.
If you're bullish on SUSS, you love this story: A German engineering company that parlayed its obsession with optical precision into a 50% market share in a critical piece of AI chip manufacturing. Their customers literally pay deposits to get in line. Memory manufacturers spend years qualifying their equipment, then become effectively locked in. The kind of investor who gets excited about this sees a company with deep engineering DNA that's positioned exactly where the AI boom needs them to be.
The bears look at the same facts and see something else entirely: A company betting everything on one specific technology (temporary bonding) in one specific place (Taiwan) for one specific application (high-bandwidth memory for AI). The kind of investor who worries about this remembers that semiconductor manufacturing has a habit of making today's critical technology tomorrow's footnote. Intel and others are already working on completely different approaches to chip stacking.
There's this great Peter Lynch quote that fits perfectly here:
"Know what you own, and know why you own it."
With SUSS MicroTec, you own a company that turned German optical precision into a crucial piece of the AI supply chain. You either believe that's a sustainable advantage, or you think it's a temporary position that technology will eventually disrupt. Both views are perfectly reasonable, which is what makes markets work.
If you're diving into this stock in the coming months, here's what actually matters:
Watch the order patterns from memory manufacturers. Not just the headline numbers, but the timing. Are customers still paying deposits to secure production slots? That tells you more about market position than any investor presentation.
Keep an eye on Intel's packaging technology announcements. They're spending billions to catch up in advanced packaging. When they talk about "breakthrough bonding approaches," they're talking about potentially disrupting SUSS's core market. The day Intel announces a major advance in fluid self-alignment is the day this thesis gets interesting.
And maybe most importantly: Watch what TSMC does with their new Arizona fabs. If they start qualifying SUSS equipment there, suddenly the Taiwan concentration story looks different.
We could say the best sign isn't the thing you're worried about - it's the thing that makes you stop worrying about it.
LEGAL DISCLAIMER AND DISCLOSURE
© 2025 Silba. All rights reserved.
This analysis is for informational purposes only. The content represents the views and opinions of the author, based on publicly available information. While we strive for accuracy, we make no representations or warranties regarding the completeness or accuracy of the information presented.
This analysis is not investment advice. It is not a recommendation to buy, sell, or hold any security. The author may hold positions in securities mentioned. Future performance of any investment discussed cannot be guaranteed.
All financial figures, unless otherwise noted, are approximations based on publicly available information. Readers should verify all claims, statistics, and analytical processes independently. The information contained herein is not intended to be a complete analysis of any security, company, or industry mentioned.
Market conditions change. Past performance is not indicative of future results. Investments mentioned may not be suitable for all investors. Readers should conduct their own due diligence and consult with professional financial, legal, and tax advisors before making any investment decisions.
By accessing this analysis, you agree that neither Silba nor its affiliates bear any responsibility for investment decisions you make based on this information.