Chip Stocks have been Choppy but China may return

Applied Materials (AMAT) is batting clean up in a quarter that has not been pretty. Lately semi stocks seem to have been hit by not only stock specific issues but continued and increasing memory concerns coupled with more macro issues. On top of all this, China trade issues which have in the meantime taken a back burner to other issues threaten to boil over yet again.
In short the news flow has been poor as have the stocks performance and we wonder where the positive impetus will come from if any.

On the semiconductor equipment front we have gotten 2 out of 3 right by projecting a 25% drop for Lam, little to no impact for KLAC and we will see if AMAT has the 10-15% negative impact we have been projecting.

While virtually all analysts and bulls on the stocks cheered the one quarter long trough theory, reported by Lam, we have been more dubious as we have yet to see a downturn of this type last only a single quarter. Recently some other analysts have come around and supported our more cautious view.

We remain concerned that Samsung’s pullback will either get longer or deeper or both. The recent news about Samsung closing a phone factory in China adds to our concerns. With Samsung turning off display spending then the memory equipment slow down followed by a phone factory slow down it appears that Samsung is taking a much more conservative direction which usually is not a one quarter phenomenon.

We had stopped talking about China trade issues for a while because they seemed to be out of the news cycle but the tit for tat tariff war has escalated again with neither side showing any signs of blinking. The death of the Qualcomm/NXP deal seems to indicate how serious China is and how much impact they can have on US tech. We had hoped that the China trade issue would fade as the administration moved on to other things to tweet about but the opposite seems to have happened as while the administration is busy with other things the China trade issue seems to be running away, out of control, on its own, as no one is addressing it.

All things memory chip related seem to be at the front of the negative semiconductor news flow. Micron has taken a beating despite being very, very cheap. Its as if we are building in to expectations a complete collapse of the memory market or something close to it.

AMAT reports
Unless you are living on a desert island somewhere you know that the semi equipment makers are all reporting a sharp drop in CQ3. Our expectation is that Applied will see a 10-15% drop in business which is roughly halfway between Lam’s 25% drop and KLA’s near zero impact. AMAT has less exposure than Lam’s prior 84% exposure to the memory business, but still sells a lot of tools to Samsung for dep and etch.

The key of the call will be where Applied comes down in how fast business comes back. Will they be a bit vague and coy as Lam was by just saying Q3 is the trough, but not exactly how long a trough, or will they call for a “sharp snap back” like KLA did. Our guess is that they will likely be somewhere in between.

2019 CAPEX spending????
The second, and perhaps bigger question is that of what 2019 looks like as we are already trying to look through Q3 and perhaps Q4 as well and are hoping for a recovery in 2019 but there is no clarity or hard evidence as to what 2019 will look like. Will it be up, flat or down? We can easily paint scenarios that support all three directions.

The upside case is built upon memory spending resuming its prior blistering pace and China continuing to be the new spenders in the semi industry. Both of these scenarios have obvious risk associated with them.

The problem is that we need BOTH China AND memory to do well for 2019 to do well. If either one slips, we have a problem.

Chip Customer Concentration
One thing that the Samsung slowdown has brought up that many are not talking about is customer concentration.
In an semiconductor industry that used to have dozens of chip companies, many with their own fabs, we are now down to a big three; Samsung, TSMC and Intel who do the vast majority of the spending. Even Intel has been past a long time ago by Samsungs deep pockets of both memory and logic spend.

Is the industry now so dominated by the big three or Samsung alone that they can ruin a few quarters of business for the semicap industry just by slowing down or stopping a fab project? Samsung’s slow down certainly rained on Lam’s parade of increasing business.

Has customer concentration either replaced or added to what has been the industry’s previously top concern of cyclicality? One could argue that customer concentration is a bigger issue, than general cyclicality as it shifts the balance of power more sharply to the customers side which also puts pricing power on their side as well. This could be a negative impact on gross margins for companies that don’t have a unique, sector dominant positions like ASML or KLAC. AMAT and LRCX could see more pricing pressure by bigger customers, such as Samsung., who control a larger share of their revenues.

Could Samsung be able to say “give us an extra 10% discount or we are going to push out that fab project another quarter or two”?

Samsung has already played that game to a lesser degree in the past and if you are doing mainly a “turns” business in tools you may be more susceptible.

China trade
The China trade risk has risen again recently in our view. Rather than slowing down and catching their breath and de-escalating, it seems to only have accelerated of late.

Our hopes for a negotiated wind down of the trade issues and tariffs seem to be fading as neither side seems willing to make the first move and behave like adults.

We would have thought by now that China, with its larger exposure to tariffs, given the trade imbalance, would blink first. But they have employed other pressure such as the Qualcomm deal to show their strength.

The KLA Orbotech deal will be a key inflection point. China did not want or need Qualcomm to buy NXP and if anything likely preferred to keep Qualcomm less dominant and weaker without NXP. China clearly needs KLA more than it needs Qualcomm, for its fledgling start up fabs to succeed and become part of the grand “made in China 2025” plan. The Orbotech deal is obviously a rounding error as compared to the NXP deal as well. We doubt that China will block the Orbotech deal but stranger things have been happening in the trade war and we would not preclude anyone cutting off their nose to spite their face in the current show down as farmers have found out.

The Stocks
After a big pop in stock price as a sigh of relief when Lam reported their Q2 and miss in Q3, the stock has now round tripped to below where it was prior to the quarter being reported and the relief rally started. ASML is also back below where it started before the quarter. KLAC has held up the best as it has obviously been impacted the least and has less memory exposure.

Much of the news has been well baked in to AMAT’s stock price. We think that investors have already assumed that AMAT will say that CQ3 is a trough, much like Lam, and that things will get better from here.

The only wild card is the display business side of Applied. It could swing the stock one way or another as it could be more variable. So far we know of no recovery on the display side, but better display performance could give things a boost….but we wouldn’t count on it.

We had previously said that Lam could have downside to $150 and Applied into the low $40’s. KLAC seems safer for now and ASML still seems to have more weakness than we would have expected.

We see no compelling reason to buy AMAT in front of the quarter as any relief rally will likely be short lived as we have seen in other cases and the risks of downside persists.

We still think Micron is dirt cheap but its hard to fight upstream against such a negative tape.
We do like AMD but think it is a bit overdone in the near term.

We don’t think Intel has yet seen the bottom in its stock price as there are many issues still revolving around the company, least of all a new CEO.

Also Read: AMAT down 10% as expected Foundry spending slow down unexpected

Share this post via:

• 
• 
• 
•