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!
I listened to both Intel's and TSMC's earnings conference calls. Here are some simple comparison that's not really related to the earnings itself:
Intel 7/15/2015, last 53 minutes, Q&A answered by CEO and CFO
TSMC 7/16/2015, last one hour 46 minutes, Q&A answered by Chairman, two CO-CEO, and CFO
It could be just my personal feeling that TSMC's executives in the conference call show more willingness and happiness to answer questions. It probably explains why TSMC's conference spent about twice amount of time than Intel's. From time to time they exchanged questions and opinions with analysts and created laughs among attendees.
On the other hand, Intel's conference is kind of depressed. I don't feel that Intel's CEO, CFO, and even investment analysts were happy or eager to be there.
There are several information mentioned by the TSMC's executives are contradicted to the common Internet rumors/news/opinions. For example:
1. TSMC said they started 16nm volume production several months (or last month) and will generate revenue in 4Q2015. So is TSMC really late for Apple's A9? Because the rapid ramp, who else will be TSMC's 16nm customer?
2. TSMC mentioned that second half of 2015 "Growth is expected to come from industrial and automotive segments, as well as from new iPhone launches and several launches of Android-based high-end phones". Note, they mentioned "new iPhone launch" will impact TSMC's revenue significantly. Does that mean Apple does have two A9 (not A9x for iPad) manufacturers?
3. TSMC manages to improve the tools portability among different nodes: "As far as we can see now from 20 to 16, 16 to 10, we can manage migration more than 95% of the tool can be reusable". Also, they claimed they minimized the conversion loss from migrating from one node to another.
4. TSMC's 2015 revenue will have a double-digit growth. You probably have heard the rumors/news report that many TSMC's major customers (Qualcomm, Apple, AMD, Nvidia, etc.) are switching from TSMC to Samsung/GlobalFoundries or scale down their relationship/orders with TSMC significantly. Then how can TSMC still manages to have double-digit growth? Either those rumors are not totally true or TSMC found more customers to place new orders and generates much more revenue than those "loss".
"(Global) Semiconductor growth revised from 4% to 3%. Foundry growth from 10% to 6%. For TSMC, we are still confident to outperform the foundry segment and still target double-digit growth rates this year."
It could be just my personal feeling that TSMC's executives in the conference call show more willingness and happiness to answer questions. It probably explains why TSMC's conference spent about twice amount of time than Intel's. From time to time they exchanged questions and opinions with analysts and created laughs among attendees.
On the other hand, Intel's conference is kind of depressed. I don't feel that Intel's CEO, CFO, and even investment analysts were happy or eager to be there.
I have had the same feeling about all of Intel's conference calls, except that I would say that the analysts are not so much depressed as too cowed to ask even the most obvious questions. For example the only specific benefit of buying Altera mentioned was the importance of FPGA being first to node. At 10nm they look like being at least 12 months behind Xilinx/TSMC. Add that to the statement that they are better at implementing ARM designs!
Looks to me that they have pushed out 10 nanometer because they do not have the volumes or market share to commit to the capex needed to stay in front of the race. Buying Altera in that context does not impress.
Contrast this with Morris Chang's answers. His only stumble this time was when he was boxed in to talking about which of his customers might be going into the server and notebook markets.
In regards to Intel I agree with this completely:
"Looks to me that they have pushed out 10 nanometer because they do not have the volumes or market share to commit to the capex needed to stay in front of the race."
