This week it’s time for one of my since long largest holdings, ASML, to report. I’ve previously discussed the company, but to give some brief context it sells high-tech equipment (machines) to chip manufacturers such as Samsung, TSMC, Intel and Micron Technology $MU. (1/17)

As you can see in the attached image, ASML has a solid track-record in terms of growth and profitability. However, the reason for picking ASML is, imo, their increasing moat. I would argue that it’s hard to find a company with the same level of moat as ASML currently has.

So what does their moat consist of? The short and tl;dr answer is: the launch and ramp-up of an extremely advanced and highly anticipated technology, which took decades to develop. I’m referring to their EUV technology, or Extreme Ultraviolet Lithography.

You guys probably enjoy your new smartphones that contain more GB of RAM, more GB of storage, a better screen, more powerful CPU and longer-lasting battery. But have you ever thought of how it is possible to achieve this combo for devices that remain in pretty much the same size?

Moore’s Law and thus higher transistor density is the simple answer. The more transistors it fits on a circuit of the same size, the more powerful, energy effieicent and cheaper (at least compared to its performance) it is. Here you can see how Moore's Law works.

Today, ASML deliver machines (DUV, Deep Ultraviolet Litography, and EUV) to chip manufacturers. But the older DUV technology has run into a dead end, as today’s production processes have reached 7nm and even 5nm. Up until now, production processes have been on >7nm nodes.

Luckily for us EUV has arrived! It utilizes an extremely short wavelength and allows exposure of fine circuit patterns that are 0.1 micron in width (approx. 1/1000th the width of an average human hair). EUV is thus pretty much required on some 7nm processes, on all 5nm and below.

ASML is the only hope for consumers who want Nvidia, Intel, Qualcomm, AMD and others to continue release faster, more powerful and energy efficient devices in the future. The company is a PURE pick-and-shovel bet. A monopolistic one.

On the market for DUV machines there are some competition from Nikon and Canon. But ASML currently holds around 85% of the DUV machine market. For EUV on the other hand there is NO competition at all. And to be frank I doubt that there ever will be. The CAPEX needed is too high.

Returning to DUV, its viability actually ran into "trouble" a decade ago on the 20nm node. But supplementing techniques enabled DUV to progress to 7nm. Double patterning (litho-etch-litho-etch), e.g, requires several separate lithography and etch steps to define a single layer.

Thus far, EUV machines enable manufacturers to use single patterning technique for their production processes on 7nm. Single patterning is less complex, requires fewer steps and is cheaper than multi (double, triple etc) patterning techniques. It’s also a lot more precise.

Here are two examples of a standard logic cell which has been exposed with DUV triple patterning technique and EUV single patterning on the 10nm node. As you can see on the image to the right, EUV is much more precise even with just the single patterning technique.

Here you can see mask counts and production cycle time for DUV (KrF, ArF, ArFi) and EUV on various nodes. If anticipating 1.5 days between masking steps in multiple patterning processes, the cycle time if utilizing EUV can be reduced significantly on 7nm and lower nodes.

Going forward, EUV will require multi patterning at some point. However, ASML hopes to ship a next-gen EUV called high-NA (0.55 NA) in 2022. High-NA will be an expensive system targeted for 3nm or 2nm and below. It will extend manufacturers' use of the single patterning approach.

On 7nm 5nm and 3nm, manufacturers will be in less need of equipment from companies such as Lam Research due to the fact that EUV single patterning reduces the amount of production process steps. In Lam Research’s case, the etch steps. It's worth having in mind if investing in Lam

Because of this, ASML will most likely be in a sweet spot for at least 5 years. When high-NA EUV hopefully hits the market in 2022 or 2023 I believe that ASML’s margins will increase in a higher pace than now. Mainly due to the company’s pricing power and product efficiency.

Key for ASML going forward is to continue increase EUV’s yield on the 7nm node and below, in order to speed up the manufacturers’ transition to the technology. But from now on, I at least expect almost all high-end devices to contain chips that have be produced in EUV machines.