Monday, April 22, 2013

Applied Materials, Process Nodes, and Future Profits

Applied Materials (AMAT) makes capital equipment for semiconductor chip manufacturing. Demand in that sector has not been robust these last couple of years, although it has come off the bottom that lagged after the recession. This article will look at AMAT as a long-term investment, not a short term trade. Given that, the first thing to note is that it pays a dividend, which is currently $0.10 per quarter, or 3.1% per year at the current $13.06 stock price.

I believe the largest factor determining future AMAT revenue and profit will be the ongoing trend towards new, small process nodes (indicated by the size of the lines used to put transistors in chips. 32 nanometer is older than 28 nm.) But let's start with where we are now.

AMAT last reported on February 13 for the first fiscal fiscal quarter of 2013, which ended January 28, 2013. Against an overall global semiconductor capital equipment spending drop of 16% in 2012, AMAT reported revenues of $1.57 billion, down 5% sequentially from $1.65 billion and down 28% from $2.19 billion in the year-earlier quarter. That is discouraging, for certain.

Applied's core semiconductor equipment business saw a Q1 y/y decline to $969 million from $1.34 billion, a decline of 28%. Its display screen segment did better, but the solar segment did worse. Display revenue dropped 17% y/y to $87 million from $104 million, which was already low by historic standards. Solar revenue dropped 77% y/y to $47 million from $206 million. There is a glut of solar supply in the market, so no turn around is expected until at least 2015. Display may see some rebound in 2013 as screen sizes start to increase in developing markets and new screen technologies are adapted.

The bulk of Applied's revenue and profit comes from the semiconductor segment. It is well known that demand for PCs has been down, and it is hard to predict where the bottom may lay. Demand for tablets and smartphones has been increasing. Overall demand is dampened because smartphones simply contain far less silicon than PC's do. They have weaker processing chips and far less memory.

Does that mean Applied and other semiconductor equipment manufacturers should be written off as dinosaurs? I think not. I think overall computational demand will continue to increase rapidly for at least the next two decades. To cram more computation into portable devices (and the computers in the cloud that serve those devices) the industry will continue to move to smaller process nodes.

Right now demand is still high at the 28 nm node. Most chips, which work in legacy, non-mobile applications, are still made at much older nodes. For high-end graphics chips from AMD and NVIDIA, 28 nm is the cutting edge. Intel is already manufacturing its newest CPUs at 20 nm, and new memory-chip production at Samsung is just started at 10 nm. Memory process nodes typically can be smaller than computing process nodes. The most advanced ARM-based chips were recently taped out at 16 nm at TSMC.

Smaller (newer) process nodes mean that more capability can be built into mobile devices (and also non-mobile devices). Transitions to 14 nm and 20 nm are almost entirely ahead of the industry. To some extent moving to these nodes may open up capacity at 28 nm, but there is a lot of technology out there that has yet to migrate to 28 nm.

The other factor is overall demand, and that depends on the global economy and the frequency of consumer upgrades. There is a lot of old equipment out there, as seen by the high percentage of PCs still running Windows XP. With the exception on Intel, Samsung, IBM, and a few others, most chip makers are now really chip designers who send their designs to foundries like TSMC and Globalfoundries for actual production. These foundries don't want to have capital equipment sitting idle, but neither do they want to lose business because of insufficient capacity.

Generally capacity has been lean since the recession, which is one reason why there has been a shortage of 28 nm capacity. The other reason 28 nm has been tight is it was harder to get it working with good yields than had been expected There is quite a bit of impatience right now among the more cutting-edge designers because of a lack of sub-28 nm capacity.

Anything under 28 nm is far more expensive to make than 28 nm. Only Intel and Samsung have had the vast capital resources to simply move to the lower nodes without concern about how much demand would be there at startup. Even Intel announced it was cutting back capital equipment spending by $1 billion in 2013 due to lower demand projections.

But what is bad for foundries is good for AMAT and other equipment manufacturers: future nodes will require far more spending on capital equipment. One reason is that some 20 nm chips will have a 3-D structure. This means a move away from lithography defined shrinking to process-defined, where with precision engineering AMAT claims a considerable advantage.

In addition to reporting revenue, AMAT reports orders for each quarter. The good news for Q1 was that orders of $2.11 billion were well above revenue and up 31% from orders in Q4, as well as up 5% y/y.

Guidance is for Q2 fiscal 2013 revenue to be up 15 to 25% sequentially. Non-GAAP EPS is expected between $0.09 and $0.15.

While Applied Materials has substantial competition in each of the types of tools that are needed for semiconductor manufacturing, it is second in overall sales revenue only to ASML in an industry where scale matters. ASML is not much bigger: it had sales of $7.9 billion in 2011, AMAT had $7.4 billion. Another American competitor is KLA-Tencor, which had $3.1 billion in sales, placing it fourth globally.

On the whole, I think it is likely that 2013 will be a year of improvement for AMAT, and with major gearing up for 20 nm in 2014, that will be a very good year. How much improvement depends on the degree of strengthening global demand for semiconductor chips.

Disclaimer: I am a long-term investor in AMAT. I also am long AMD, but do not own any other company mentioned in this article. I will not buy or sell AMAT stock for one week following this article's publication date.

See also:

My main AMAT analyst conferences page.

My Q1 2013 AMAT analyst conference notes

No comments:

Post a Comment