# How Lithography Machines Are Winning The War For Moore's Law

An episode of [The Hardware Podcast with Fexingo: Chips, Devices, and Electronics Engineering Conversations](https://invisiblenetwork.ai/podcast/1896799982), 19 Sep 2026, 8:16.

Audio: https://audio.fexingo.com/business/the-hardware-podcast/episode-0191.mp3

We dig into the staggering economics of extreme ultraviolet lithography and why ASML’s monopoly isn’t just about optics, but about global supply chain fragility. With chip manufacturing costs hitting record highs in late 2026, we explore how a single machine can cost over three hundred million dollars and why the next generation of logic chips depends entirely on this fragile ecosystem. #ASML #SemiconductorManufacturing #EUV_Lithography #ChipDesign #HardwareEngineering #SupplyChainRisk #TechInfrastructure #AdvancedPackaging #MooreLaw #OpticalEngineering #GlobalTrade #FexingoBusiness #BusinessPodcast #TechnologyTrends #ManufacturingInnovation #CapitalExpenditure #TheHardwarePodcast #ChipWar Keep every episode free: buymeacoffee.com/fexingo

## Transcript

Lucas: If you look at the latest quarterly reports from the major foundries, the capital expenditure numbers are honestly staggering. We’re talking about single machines costing upwards of three hundred million dollars, and they only print one type of chip.
Luna: That is insane. I always assumed the bottleneck was just having enough factories, but it sounds like the actual tools required to build those factories are the real constraint.
Lucas: Exactly. And it’s not just the price tag. It’s the fact that there is essentially only one company on Earth that can make them. ASML dominates the extreme ultraviolet lithography market so completely that their lead time for orders stretches out two or three years.
Luna: So if ASML has a hiccup, every major tech company’s roadmap for the next few years gets delayed? That seems incredibly risky for the whole industry.
Lucas: It is. But here is the thing most people miss: it’s not just a Dutch monopoly. It’s a web of German optics, American lasers, and Japanese materials science. You cannot replicate that supply chain overnight because the knowledge is distributed across hundreds of specialized suppliers.
Luna: Right, because the mirrors alone have to be smoother than anything ever made by human hands. I remember reading that the surface irregularities are smaller than an atom.
Lucas: Precisely. Carl Zeiss makes those mirrors, and they are working with tolerances that defy conventional manufacturing logic. If you want to understand why chip prices are sticky even when demand dips, you have to look at the sheer complexity of assembling these machines.
Luna: And yet, we keep pushing for smaller nodes. Why do we need five nanometer chips when our current devices work perfectly fine?
Lucas: Because efficiency matters more than raw speed now. As we move toward edge AI and massive data centers, power consumption is the enemy. Smaller transistors mean less leakage current, which means your server doesn’t melt down or blow the electrical grid.
Luna: So it’s actually an energy play disguised as a performance war. That changes how I view the whole arms race between Intel, TSMC, and Samsung.
Lucas: It does. And speaking of TSMC, their expansion plans in Arizona and Japan are heavily dependent on getting EUV tools delivered on schedule. Any delay there ripples through to Apple, NVIDIA, and AMD before you even see a new phone launch.
Luna: It feels like we are building the foundation of the digital economy on top of a very narrow bridge. One wrong step and everyone falls off.
Lucas: That is a great way to put it. And the geopolitical angle makes it worse. Because these machines are so critical, governments are trying to restrict their export to certain regions, which complicates logistics even further.
Luna: I guess that explains why the stock hasn’t really corrected despite all the talk about tariffs. The market knows there is no substitute.
Lucas: Right. Investors treat ASML like a toll booth on the highway of innovation. You pay to get through, and you don’t have many other options. But let’s pause for a second on the listener support aspect.
Luna: Oh yeah, before we dive deeper into the optical physics, I want to mention something quick for everyone tuning in.
Lucas: If today's tech conversation gave you something usable, the way these stay ad-free is listener support. You can help us keep digging into the nuts and bolts of hardware without ads by checking out buy me a coffee dot com slash fexingo.
Luna: It really helps us cover the expensive research needed to break down these complex supply chains for you. Just a quick check-in if you found value in today’s deep dive.
Lucas: Thanks for keeping the show running. Now, back to the mirrors. Since light behaves weirdly at that scale, how do they even handle the heat generated by the plasma source?
Luna: I read that they use liquid tin droplets for the target material, right? It turns the metal into plasma instantly to create the UV light.
Lucas: Correct. They shoot microscopic drops of molten tin at high velocity and hit them with a CO2 laser. The resulting plasma emits extreme ultraviolet light, which is then collected by those impossible to make mirrors.
Luna: That sounds like something out of a sci-fi movie. Shooting liquid metal with lasers to carve circuits. I never thought about the physical violence involved in making a microprocessor.
Lucas: It is violent engineering. And because EUV light is absorbed by air, the entire process happens in a vacuum chamber. So you are dealing with ultra-high vacuum, extreme heat, and precise mechanical alignment all at once.
Luna: No wonder these machines take weeks to assemble and test. It’s basically building a particle accelerator inside a clean room.
Lucas: A perfect analogy. And the yield rates are still improving, but they are not perfect. Every defect in that lithography step ruins the entire wafer, which is why yield management is such a huge cost center for fabs.
Luna: So if a fab runs at ninety percent yield instead of ninety-five percent, that’s a massive amount of wasted money. It really highlights why experienced engineers are so valuable.
Lucas: Absolutely. You can buy the machine, but you cannot easily buy the know-how to run it efficiently. That institutional memory is what keeps the older fabs competitive against newer ones.
Luna: It reminds me of how difficult it was to scale up solid state battery production earlier this decade. The tooling is half the battle, the process control is the rest.
Lucas: Similar dynamic. And with chiplets, we are seeing a shift where instead of making one giant monolithic die, we are connecting smaller dies together using advanced packaging techniques.
Luna: Does that reduce the reliance on EUV lithography? Or does it just change where the bottlenecks are?
Lucas: It shifts the bottleneck. You still need EUV for the logic cores, but the interconnects between chiplets might use older, cheaper nodes. It’s a hybrid approach to stretch the value of each expensive machine.
Luna: So we are optimizing around the scarcity rather than trying to eliminate it. That seems like a pragmatic response to the constraints we discussed earlier.
Lucas: It is. And as we look toward two thousand twenty-seven, the focus is shifting to higher numerical aperture EUV systems. These will allow even smaller features, but they add another layer of complexity to the optics.
Luna: Which probably means even longer lead times and higher prices. It feels like we are entering a phase where only the biggest players can afford to participate in the leading edge.
Lucas: That is a very real concern. Mid-tier manufacturers are getting squeezed out of the premium segment, which could consolidate the industry further over the next five years.
Luna: It will be interesting to see if sovereign wealth funds try to invest directly in local capacity to bypass these global dependencies. The politics of silicon are going to get much louder soon.
Lucas: They already are. But until those alternative supply chains mature, the world remains dependent on a handful of facilities in the Netherlands, Germany, and the United States.
Luna: Which brings us back to the fragility of the system. It is impressive technology, but it is also a single point of failure for the entire global economy.
Lucas: Exactly. And that is why understanding the hardware behind the software is so important. You cannot appreciate the app store without appreciating the factory floor that makes the phone possible.
Luna: Well said. I think listeners should start looking at their gadgets with a bit more respect for the machinery that created them. Next week we might tackle the thermal challenges in quantum computing.
Lucas: That sounds like a hot topic. Literally. We will see how close we can get to absolute zero without breaking the bank.

Shows are listed from Apple Podcasts' charts and directory, and Podcast Index where it is used. Every show and episode belongs to its publisher and plays from the publisher's own feed; this site hosts none of them.