It is true that semiconductor spending will likely reach an all-time high in 2017, but context is needed.
First, and most basically, the current level of spending represents only modest growth relative to both
previous peaks and mid-cycle periods:
The growth rates, considered over time, look undemanding in a category that should, on a normalized
basis, grow well in excess of GDP. Furthermore, we are in the early stages of a technological inflection in
the composition of equipment spending that benefits AMAT. For the last 20+ years, more transistors
were crammed onto each chip using the relatively simple process of “planar shrink”: chopping the wafer
into ever smaller cells using lithography to create ever tighter patterns. A few years ago, this strategy
reached a physical limitation, which first became an issue in memory chips. If the transistor cells were
made any smaller through planar shrink, they would no longer fit enough electrons to accurately hold
the “on” or “off” position. Therefore, in order to make progress in adding transistors, chip makers began
to focus on building vertically. This vertical building is more materials and equipment intensive, and it
relies heavily on etch (cutting/removing material) and deposition (adding material).
The verticalization trend in chipmaking has driven general growth in equipment spending, and AMAT
has been a disproportionate beneficiary given their high market share in dep and etch. AMAT has no
exposure to lithography, the previously dominant application.
China is another bolster to semiconductor equipment demand. While China has been a minor player in
semiconductor production for decades, in the past year the government has launched a high profile
campaign to support investment in domestic fabs. This may result in only modest incremental demand if
the industry rationally shifts spending to China from currently dominant geographies like Korea.
However, it could also lead to an extended period of over-building, of which AMAT would be a key
beneficiary. Given that most WFE projections do not contemplate significant Chinese WFE spending
growth, China’s entry into the semiconductor industry introduces attractive demand growth optionality.
The underlying demand drivers for semiconductor equipment have also changed. For most of the last 20
years, the PC cycle was the key driver of demand. PC demand was economically sensitive, and
dependent on product cycles dictated primarily by INTC and MSFT. Today, we are in the upswing of
structural growth in demand for semiconductor materials that is driven by a wide range of consumer
and enterprise products, including increasing smart phone penetration in emerging markets, general
“Internet of Things” applications, and escalating demand for data storage. End market diversification
should also contribute to a smoother demand cycle relative to history.
In thinking about the growth trajectory, it is also worth mentioning Display. AMAT sells tools, mainly
related to deposition applications, to producers of LCD and OLED screens. The segment amounts to
~10% of AMAT’s earnings power, and is benefitting from similar trends to those in semiconductor. There
is a technological inflection underway as display companies build out OLED capacity, which is much
more capital intensive than LCD. While we are certainly in a display upcycle at the moment, our work
has suggested that elevated display activity should persist for a number of years. Further, AMAT is using
the inflection to develop new tools and gain share. China is also a factor, as it seeks to replace Korea as
the primary provider of displays.