You are currently viewing SemiWiki as a guest which gives you limited access to the site. To view blog comments and experience other SemiWiki features you must be a registered member. Registration is fast, simple, and absolutely free so please, join our community today!

After a strong bounce off Q4 bottom Q1 may slow Weak macro & semi sales do not foster

Robert Maire

Moderator
After a strong bounce off Q4 bottom Q1 may slow Weak macro & semi sales do not foster

While we expect an improving 2016 through the year we will not likely see another big "bounce" as we did off of a very low Q4 bottom. This means that the sequential increase is slowing and investors may have overblown expectations given the outperformance of the stocks over the last 90 days versus the broader market.

Stocks usually trade on the second order derivative...

As we have seen in many others stocks, especially Apple's, the question is not wether we are growing or not, its rather is the rate of grow increasing or slowing?

The semiconductor equipment industry hit a pretty sharp bottom at the end of 2015 which made the predicted uptick in Q1 look even better than it actually was given the depth of the bottom. If we now look at the likely rate of change of Q2 over Q1 as predicted in the upcoming conference calls, we think most management will not be able to guide to a similar percentage increase in business in Q2 even though Q1 will be good. This is obviously the law of small numbers as the industry comes off a downturn.

Its hard for capex to increase while end user semiconductor markets are weak.....
There are two reasons to buy new semiconductor equipment tools:

  1. Technology buys- to keep up with Moore's Law
  2. Capacity buys- which are driven by the volume of end user demand

Right now, technology is the main driver as we continue to need ever more multipatterning with ever more layers of 3D NAND. This is despite the fact that Intels cadence of Moore's Law has slowed from 2 to 3 years.

The very simple math of Intel is that two years of spending is now spread over three leading to lower capex intensity.

NAND is on a headlong race to the bottom of the cost curve fueled by aggressive spending. Given that NAND is a price elastic market that sees increased demand as prices drop (think end user applications such as SSD's that have almost limitless demand) spending will continue with bigger players using outsize spending to keep ahead on the price curve versus less well off competitors.

The DRAM market is not as price elastic and as such has seen more of a drop in spend after a long strong period.

The foundry market has been less than great given weak PC and tablet/smartphone sales.

Stocks may be defying reality....

If we look at stock performance over the last 90 days since the beginning of the year, most of the large cap semi equipment stocks have had strong gains while the S&P and NASDAQ are only up a couple of percent.

Of the large caps, AMAT and ASML are both up about 20%, with LRCX up about 12.5%. AMAT was certainly more upbeat than Lam but the price of the stock is based on a lot of positive future performance despite some of the headwinds in the industry.

We would feel more comfortable with AMAT's stock price below $20 as we still are wondering what Applied does for long term growth. We also remain concerned about AMAT getting aggressive on price to try to gain share.

ASML at $100 is even more outdone than Applied. It probably belongs more like $90.

Though immersion sales have been good, investors expectations for EUV are way overly optimistic. SPIE wasn't all that positive for EUV. Though source power is improving, the infrastructure is still lacking.

We have it from several different industry sources that TSMC has pushed EUV out even further than previously thought and that ASML would be lucky to see EUV at 5nm at TSMC (if at all). Multi patterning just works....it may be ugly. & costly ..but it gets the industry from here to there. There are just too many moving parts left in EUV to bet the ranch on it. European analysts who are less familiar with the EUV issues just regurgitate the party line.

The back end test companies always have danced to a different drummer and while Teradyne is up a solid 8% we think there could be some upside room left on the stock especially relative to other large caps which have all eclipsed TER. We would remind investors of our note "Teradyne = Targetdyne" as we believe their business could be attractive to others.

Mid and smaller cap stocks
...
The stock thats up the most in the whole space is AEIS, at a 30% gain over the last 3 months. Most of what we have seen at AE is the strong financial performance of a pure play company without the losses associated with the horrible solar business. However , that story is lot older than 90 days and should have baked into the stock a long time ago. Even with the strong financial performance its a little hard to justify being up 30% when your biggest customers, AMAT & LRCX are up 20% and 12.5% respectively. Although we continue to love the company and management we would not want to be too greedy and consider taking some money off the table.

At the other end of the spectrum, Formfactor is down about 15% after a combination of self inflicted as well as market related damage. FORM has done a good job of recovering from these "trip & fall" scenarios, and we would not hesitate about putting money to work here as we are certain they will recover as they have in the past.

While VECO's has been dead flat for the last 3 months, AIXG has been among the most volatile stocks but is now up 15% over 90 days. Even though the LED market is a nuclear wasteland, AIXG likely still has upside as a takeover target as it still trades at a fraction of its prior high whens LEDs were hot. Wether the buyer is VECO, AMAT, ASMI or some other company the company has value.

Ultratech is up about 10% in the first quarter but we don't see the recovery in laser anneal that would support that move. While inspection has potential, in our view its still too early to assign significant value to it. Lithography is not knocking the cover off the ball either. Unless we start to see real traction in the promised land of laser investors may get tired of waiting.

ACLS has done OK with the stock up about 7%. Though we think there is upside from here, it may take a while to materialize. The company continues to execute well, and much better than historically, but the markets are not cooperating 100% especially in memory which is important to ACLS.

End user demand is needed to support the stocks from here...
What the industry really needs is some solid end user demand in phones or tablets. IOT is not a good driver as it utilizes trailing edge fab capacity. 28nm continues to be a great node for the industry. The IPhone SE looks like a yawner, so we really need a big Iphone 7 splash in the fall to get the industry rolling again.

We don't see DRAM picking up again in 2016 although NAND continues to roll along.

We think that TSMC is likely in the best position of the top three chip makers and will continue in that position for likely most of the year.

Management shake ups at Intel seem to underscore the view that things are not going all that well and we think that senior management at Intel does not want to stick its neck out and double down on spending to keep the lead in technology and would rather slow spending to improve near term financials.

We are hoping to see upside on XPoint memory spending but so far have been disappointed.
While Samsungs spending on memory seems a certainty, the logic side may be a little less so without Apple's business.

Summary...

As a whole we would want to lighten up on some of the stocks, especially some of the larger cap names, going into Q1 reporting season as theres not that much upside on the horizon to support the uptick many of the stocks have seen.

We would look for more sustainable stories for the balance of the year. Given the lack of large cap analyst meetings at Semicon West, we don't expect a positive "kicker" coming out of that show after Q2.

There are no significant technology shifts on the horizon either as Multi patterning is still the name of the game while EUV continues to slip. 10nm is not that much different than 14nm and 7nm does not look like its going to see any significant changes that would alter market share on a technology basis. Materials and transistor design are not expected to see significant change that could drive new equipment or process changes.
 
Last edited:
Top