ASML--the critical link in the chain--new position

 Understandable business.

Industry Overview

The semiconductor industry players' businesses are incredibly interconnected, with each business within the industry playing a part in the whole. The complexity of individual parts, together with the need to coordinate the development of each activity, makes the chain a unique construct and interdependent. The chain comprises some important segments.

The chain (put simply) is, designer software (CDNS, SNPS), chip designers, (AMD, NVDA), FAB operators (TSMC, Samsung, INTC which is integrated), and equipment suppliers that go into the FABs (ASML, KLAC, AMAT, ENTG, LRCX) and the end users, a large group comprising the likes of AAPL, the hyperscalers, consumer electronics and industrial uses. The industry requires an elevated level of cooperation between these players for strong and smooth growth with no bottlenecks emerging.

Having said that the industry has a reasonable level of cyclicality driven by consumer electronics demand and product cycles. The overall trend is for growth and this is expected to continue.

ASML’s history is interesting with investments by INTC, Samsung and TSMC as a venture to solve the lithography issues and a competitor to the Japanese consortium of Nikon and Canon. In this light, ASML can be viewed as a critical part of the industry with a specific mandate, which it has achieved after spending some 25 years and e10b to become the leader in lithography.

Estimates are that ASML lithography machines make up 20-25% of a total FAB build, so are a very big part.  

Pricing in the industry appears to attempt to reward investment and innovation, while all players watch the returns each segment earns. There is a quid quo pro approach in pricing, not gouging. ASML states broadly that pricing is shared 50/50 between customers and shareholders.

ASML business model

In this chain, ASML plays an important step providing DUV and EUV machines. These lithography machines are critical for the manufacture of chips. ASML has some competition in the less complex DUV market from Canon and Nikon but at this stage has the more complex EUV industry to itself.

ASML customers are the FAB operators. Chief amongst these is TSMC. Customer concentration is apparent with the top 2 being approx. 60% of revenues, the top 3 maybe 65-70%. There is a tug-of-war between the cost of the lithography machines, which are enormously expensive and the need for the machines. Customers desire good productivity gains for a reasonable price. The order book has proven to be lumpy. Over the long term, the growth in ASML revenues however should mirror the growth in FABs. The push by different governments to establish FABs in their own countries, the US, Japan, and Germany, will help demand as well as ongoing chip demand. The low absolute number of systems sold, especially EUV make the numbers lumpy.

At ASML’s investor day in 2022, they pointed out that the semiconductor revenues are expected to double from 2021 to 2023 (9%pa). ASML is due for another investor day on Nov 24, and the market growth will be higher with the advent of AI. Strong organic growth is expected. At that stage, ASML disclosed capacity expansions across their various products as per below. ASML also projects some detailed LT financials despite the difficulty in accurately forecasting these.

ASMl revenues are typically 75% new machines and 25% upgrades and maintenance. Customers are concentrated and geographically typically comprise 40% Taiwan, 30% South Korea and 15% China (DUV). Machines can last +30 years. As can be seen below the number of machines are not large with 600 DUV, 90 Low EUV, and 20 high EUV (e350m each!) starting in 2027/8 expected.

ASML outsource 80% of the components to construct a machine. That is a large reliance on a fragmented supplier base and is one of the significant risks if suppliers do not innovate enough to keep up with technological change. ASML is willing to acquire the most critical suppliers.

The main risk is that ASMl cannot keep up with customers' expectations for technological advancement in time or price and alternatives are investigated. That happened in the NAND space with the stacking of transistors on the chip instead of denser spread which is of lower lithography intensity and is a more etching and deposition-intensive chip.

Risks 1. Customer and geographical concentration, 2. Suppliers appropriately investing and innovating, 3. ASML must deliver productivity advancements at an acceptable price otherwise disruptive alternatives may occur.

Operating History.

In summary, the financials show a highly profitable company with strong LT growth but with above-average volatility. Some acceleration of growth in the latter years is evident.

Revenues have shown growth of 8-19%pa on 5Y rolls. Acceleration in the latter years.

EPS growth 5Y rolls have been between 9-35%, with a 5Y roll average of 21%. There is also a skew towards better latter years as the more expensive equipment becomes a larger part of the mix.

ROE 10Y average is 31.6%, while the 5Y ave is 45.3% and rising as large share buybacks impact. GM has averaged 47.2% over the 10Y and 49.6% over 5Y. Again this shows an improvement in latter years and exhibits above average volatility.

Total assets have increased 7-16%pa on 5Y rolls, with an average of 12%. Good growth in assets.  

NPM 10Y average has been 25.3%, with a range of 21.9-31.6%. Highly profitable business.

The reinvestment is reasonably low with share buybacks being large skewing the data. The business does not appear capital intensive, which is a bit of a surprise but indicates the outsourcing model while investment in R&D is quite large (above the line) and is where the real investments are made.

Cash flow reconciliation is good. 10Y is 110% and 5Y 116%. However, there are negative cashflows in the poor years as ASML continues to invest capital through the cycle. The volatility is a product of this as well as some semi-industry cyclicality.

Note below the recent above-average earnings, possibly due to the pull forward of Chinese demand as the customers attempt to front-run possible US restrictions. That will inevitably lead to a lower level of demand as Chinese orders normalise.

 

Management

Management has been stable and experienced. The SBC is low at 2% (pretax on NPAT).

Management has given long-term forecasts but has had to readjust those, shorter-term forecasts have usually been ok.

Management is assessed as above average.

Balance Sheet

The balance sheet is net cash and has easily funded activity through the weak demand periods.

Other

Unsurprisingly the PE has traded in a large range over the long term with a low of 22X, a high of 59X and a median of 39X historic earnings.

Earnings growth is expected to be very strong over the LT but punctuated by negative years due to the cyclicality of the industry and the strategy of continuing to invest through the cycle.

The market is concerned about the reduction in revenue forecasts, the decline in Chinese orders and possible export restrictions, and the delay of Samsung/INTC orders. ASML is seeing the impact of slower conditions and caution in the older parts of the semi-market without as yet seeing the AI boom. Note the non-Chinese part of the business is expected to grow strongly into 2025.

ASML continues its strategy to build through the cycle to not become the bottleneck when conditions improve. There appear no structural issues with the demand profile or any significant obsolescence risk.

CONCLUSION - VALUATION

ASML fits my international investing strategy, which is to acquire leaders with good qualitative stories at reasonable valuations. ASML is likely to be my only equipment supplier in the semi-space; I don’t want to over-diversify and expect to play this theme through the best alternative, imo.

The interesting feature of the investment case here is that, despite outstanding investment metrics, there is higher-than-average earnings volatility, which is explained through their business model. Earnings calls are littered with analysts attempting to accurately forecast ST earnings. This exercise is fraught with difficulty and will inevitably lead to over- and underreactions in the SP, IMO. Therein lies the opportunity.

At US$680 I see 9% compound for the next 5 years, assuming a 28X exit multiple and 12% eps growth (which could be light). If the SP falls towards $600 I will substantially increase the exposure.

Note ASML has a market day in a few weeks.

 

 

 

 

 

 

Please note the disclaimer.

I have heavily relied upon company results and various investment podcasts for this report.