AMAT LRCX and EUV Economics

All of this coming out of an industry with anemic, single digit, overall spending growth, and much consolidated , more powerful customers.

“Somethings gotta give…”
Unless Applied and Lam are trashing every other competitor in the market and capturing 100% of new business its hard to view both as being right, someone has to lose…it just doesn’t add up.

Applied’s “in line” quarter…
Applied reported a quarter more of less in line with guidance and gave guidance above a very conservative and downward , current quarter expectation (with the “whisper” expectation probably being much higher).

As expected, display business is cut in half but Applied still expects to see more sustainable higher growth in display going forward based on Gen 10.5 and OLED.

Applied’s guidance underwhelmed investors and the stock traded flat after the earnings announcement in a relatively strong market. Investors are dubious of the “sustainability” argument and level of business over the next several quarters.

Applied has done a great job over the last year or two and we have seen the stock do well but it feels like there may be a bit of a pause with investors moving into “show me” mode

Lam- Consolation prizes & Ignoring the dead body….
Lam held a lightly attended analyst event in New York City (as compared to crowded AMAT & ASML recent meetings). Lam management completely avoided any mention of the failed merger or KLA in any prepared remarks in an event that was originally intended to be a marriage celebration.

Much as Applied did after TEL failed, we did not hear about a new strategy or alternatives , all we heard was that we were going to get more growth, more SAM and more share but no real explanation of how or what will change, other than to say we did it in the past and we will do it in the future. Short on details, long on promises…

We did get a very good new target model (much as Applied did post TEL) and as an added bonus we got the promise of a super strong March quarter as a kick starter.

The whipped cream with cherry on top was a billion dollar stock buy back and a 50% increase in dividend to 45 cents.

Investors obviously liked the consolation prize and near term guide and bid the stock up a bit but most remain skeptical on the sustainability and longer term promises especially now that large, inorganic growth has been blocked.

Cyclicality and Sustainability, Cycles are smaller than they used to be, but so is growth….
In Applied’s earning conference call, Bob Halliday, the CFO went to great mathematical lengths to say that the cyclicality and therefore volatility of the industry has been reduced over the last ten years as compared to past feast and famine cycles.

While we certainly agree that the volatility has been reduced, we are still in a cyclical industry where the drivers of the cycles have changed. It is still very important to understand the cyclical drivers as they will cause the quarterly and yearly variations.

In addition not only are the cycles dampened but so has growth dampened, as compared to the “bad old days”.

“Seacyclical” behavior…
We had coined the word “seacyclical” a number of years ago to describe an industry that has gone from multi year cycles to “seasonal cyclicality”. The industry moved from multi year cycles, driven by new DRAM hungry, Microsoft Windows versions to annual Apple Iphone renewal cycles and holiday electronic sales.

Technology & Capacity cycles…
We view the industry as still being driven by two primary cyclical spending factors; technology & capacity. Each has many underlying types. In technology we have the Moore’s Law technology node cycles that used to be 2 years in length but are now longer or shorter, depending on wether you are Intel or TSMC, going from 14NM to 10NM and now 7NM etc;. Technology cycles also include going from planar to 3D NAND, planar to FinFET, XPoint, etc;.

Capacity cycles come in different varieties and most usually impact “commodity like” markets with DRAM being the poster child for capacity driven cycles. Foundry capacity also comes in waves at different nodes, such as the continued lack of capacity at the hugely popular 28NM node.

Granular throttling of capacity…
One of the reason’s for reduced volatility is that the industry no longer buys capacity in “fab sized” bites. We can now incrementally add capacity to existing facilities without starting a new “greenfield” fab.

Consolidation has also helped in that we no longer have 7 DRAM makers in the U.S. each building capacity assuming they will win a 50% share resulting in 350% more capacity than the industry needs (that was a long time ago but it did happen…)

Less cyclical but less growth too….
The wafer fab equipment industry has never matched its previous peak and for the last three years has remain in the low $30B range without its past huge growth spurts, now looking at low single digit growth.

In lieu of growth what we had seen is shifting of spending mix away from litho to wafer processing equipment over the last few years as EUV stalled out while the industry had to keep moving and thus refocused its spend on wafer processing to make up for litho issues thus driving wafer processing companies. We will likely see that mix shift back towards litho…..

EUV economics…
EUV economics are relatively simple…..EUV tools cost twice as much as regular litho tools but save spending on other processes (read that as etch & dep) by cutting those costs at least in half (otherwise the economics wouldn’t work and no fab would buy a tool or risk having their capital costs go through the roof).

There was one slide in the recent ASML analyst day that said it all……

After looking at this slide its hard to suggest that EUV will have a minimal, single digit percentage, impact on dep and etch sales. Implying otherwise flies in the face of economic decisions at fabs. While Lam and AMAT may be in denial, there will be a shift in spending mix back towards litho away from wafer processing, reversing the prior trend.

To be fair, this reversal hasn’t even started yet as EUV is not in production and both Lam and AMAT have several more years of outsized growth in dep and etch at lithos expense.

Near term still looks good…
The next few quarters continue to look good. Spending on 3D NAND continues unabated and it appears that we will have a couple of quarters (at least) of renewed DRAM spend, all of this is in addition to TSMC’s spending on 10NM and now 7NM.

Memory is still driving most of the spending and historically that has been Lam’s wheelhouse. TSMC has had a historically strong relationship with Applied.

The key question is how “sustainable” the near term is into the future. The elasticity of non volatile memory demand seems to be endless and probably one of the better bets for sustainability. Logic remains soft and DRAM is still limited by supply demand imbalances.