Where is the Semiconductor Manufacturing Sweet Spot?

[Daniel Nenni]

Where is the semiconductor manufacturing sweet spot? Two recent Semico Research Corp. studies provide the information to not only determine the overall sweet spot but to dig even further to find which products and technologies are the driving forces behind the growth or decline.

Chart 1, below, was developed from data in the fab database study. It shows the number of fabs operating and planned by wafer size.

View attachment 19625
Chart 1. Number of Fabs, Operating and Planned, by Wafer Size


Surprisingly, there are still more than 350 fabs operating at 150mm or smaller. Many manufacture trailing edge devices, especially discretes, while others manufacture relatively new devices, including analog, power, MEMS and LED devices. These semiconductors are not leading-edge devices, but they are essential to many electronics industry end-use products. The die size for many is quite small, and they do not scale well to smaller geometries. Many may continue to be manufactured in 150mm fabs for the immediate future. Others will undoubtedly migrate to 200mm fabs.

The sensor category is one attracting a lot of interest from the IoT, medical, and automotive markets. Sensors are produced on 150mm and 200mm wafers, but even 300mm will be impacted by this growing market. TSMC reported last year that they were moving some sensors to 300mm production.

There were more than 150 fabs manufacturing devices on 200mm wafers in 2016. Many are producing MOS Logic semiconductors, but there is a sprinkling of 200mm fabs producing a wide variety of other semiconductor product types including analog semiconductors, MEMS devices, power semiconductors, and SRAM. Nineteen new 200mm fabs are planned or under construction.

There were more than 100 fabs operating at 300mm in 2016. These fabs are focused on high-volume devices manufactured at leading-edge technology nodes, including MPUs, DRAM, and NAND Flash. Following the usual pattern, many semiconductor product types will migrate from 200mm to 300mm fabs. More than 40 additional 300mm fabs are planned or under construction.

Chart 2, below, developed from data in the wafer demand study, shows wafer demand by technology node.

View attachment 19626
Chart 2. Wafer Demand by Technology Node

The first category, greater than or equal to 800nm, was chosen because it represents most manufacturing in 150mm or smaller fabs. The second category, 500nm through 130nm, was chosen because it represents most manufacturing in 200mm fabs. The CAGR for 2016 through 2021 for this category is 4.6%. Most 300mm fabs are at technology nodes beyond 100nm, the third category. The CAGR for 2016 through 2021 for this category is 9.8%.

So, where is the semiconductor manufacturing sweet spot? It depends. The sweet spot for new fabs is 300mm, with 200mm fabs close behind. The sweet spot for wafer demand is 300mm wafers. Wherever your sweet spot is, Semico's fab database and wafer demand studies have the data to help you find it. These studies are:

Semico Fab Database: Update Summary, Second Half 2016
Semico Wafer Demand: Q4 2016 Highlights

Contact Rick Vogelei at rickv@semico.com for more information.

 

[Arthur Hanson]

Dan, do you see TSM's Info packaging giving them a significant advantage in this area of mems/sensors? Also will their capabilities of integrated photonics give them an even larger lead in the large node area?

 

[Daniel Nenni]

Dan, do you see TSM's Info packaging giving them a significant advantage in this area of mems/sensors? Also will their capabilities of integrated photonics give them an even larger lead in the large node area?

Coincidentally I had quite a few conversations about InFO packaging last week in Taiwan. TSMC bringing packaging in house accomplishes two big things:

1) It locks big customers into TSMC
2) It allows TSMC and big customers to work in secret on new packaging technologies.

It is not a big margin business so this is not a profit play and it does alienate packaging partners but point #1 and #2 by far outweigh the negatives. And yes MEMs and sensors are a focus for TSMC now more than ever before. I remember meeting you years ago and you yakking about MEMS and Sensors and me yawning. So yes, I was wrong and you were right.

 

[Arthur Hanson]

Economies of scale if handled properly should allow those that take advantage of it to outpace everyone else. This is also where I see TSM as the leader and if handled properly the advantage will grow and the gap will become ever larger. This will lead to either consolidation or concentration depending on financial and talent acquisition strategies.
Pushing technology from the lowest level to the highest level in design, manufacturing, packaging at all levels, all the time is key, for they must make all others at all levels be followers and not leaders.

 

[wgt46]

So is this move for TSMC to 300m sensor production responding to current IoT demand or is it in advance of it hoping to catch the wave as it crests?

 

[count]

Economies of scale if handled properly should allow those that take advantage of it to outpace everyone else. This is also where I see TSM as the leader and if handled properly the advantage will grow and the gap will become ever larger. This will lead to either consolidation or concentration depending on financial and talent acquisition strategies.
Pushing technology from the lowest level to the highest level in design, manufacturing, packaging at all levels, all the time is key, for they must make all others at all levels be followers and not leaders.

100% agree. Semiconductors are a scale driven industry, those that have the scale and the growth can outpace everyone else, especially on the leading nodes. If TSMC plays their cards right, they will run away with the market.