Semico is a semiconductor marketing & consulting research company located in Phoenix, Arizona. We offer custom consulting, portfolio packages, individual market research studies and premier industry conferences.
If there is one truism in the semiconductor market, it is that rising costs will impact unit demand at some point if they continue long enough. The subject of this blog deals not with device ASPs; but rather with rising SoC design costs, and their effect on the number of designs at the advanced nodes. Even though the mechanism governing each set of numbers is different (device ASPs vs. design costs), the overall impact can be similar. In this case, the number of design starts is impacted by the climate of rising design costs.
The SoC market has become the largest segment of the worldwide semiconductor market and is expected to reach $169.8B by 2021, a 7.7% CAGR. The adoption of the SoC design methodology has allowed the creation of very complex silicon solutions with rich feature sets and functionality. However, the predicted end of 'Moore's Law' is looming over the industry as improvements, once taken for granted through process geometry gains, start to become much harder to achieve at each new process node.
No question, 2017 is expected to be a good year for the semiconductor industry. Semiconductor revenues for 2017 are expected to increase over 9% this year. A 6% increase in unit sales, as well as higher average selling prices for memory products, will help drive the revenue growth rate to its highest level since 2010. Wafer demand is forecast to grow by almost 8%. The higher revenue growth compared to units and wafer demand is a welcome change compared to the last two years. But there are a couple clouds on the horizon.
The strong unit growth over the past several years has been at the expense of falling average selling prices. New MEMS and sensor products, the driving forces behind IoT, have experienced steep declines in ASPs. The industry is very familiar with the declines in DRAM cost per bit and how that drives increased applications and demand for memory. Comparing MEMS and sensor ASP declines to that of DRAM, there is a close correlation between the two. In fact, between 2010 and 2016 sensor ASPs fell faster than DRAM cost per bit over the same timeframe.
The Wafer Demand Summary and Assumptions is a quarterly publication. It includes an excel spreadsheet with annual wafer demand by product by technology from 2010-2021. Product categories include DRAM, SRAM, NAND, NOR, Other Non-volatile, MPU, MCU, DSP, Computing Micro Logic, Communications, Other Micro Logic, Programmable Logic, Standard Cell, Gate Array, Analog, Discrete, Optoelectronics, Sensors and Digital Bipolar. In addition, there is a summary write-up providing the major assumptions behind the forecast and changes from the previous quarter.
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.
published by Morry Marshall on Sat, 2017-04-15 00:13
Semico Research has just released a mature technology market research study. Wait! Mature technologies? Aren’t those fabs trailing-edge technology, old hat, passé? They may use older technology, but there’s a lot of action there now.
For many years, semiconductor manufacturing has tended to migrate from older fabs to newer fabs in a predictable manner. Leading-edge semiconductors such as processors and memory migrated to leading-edge fabs. ASICs and other integrated circuits migrated to the second-generation fabs just vacated by the leading-edge parts. Discretes and other trailing-edge devices migrated to the third-generation fabs. Older fabs were decommissioned. That pattern ended several generations ago. The reasons are complex. It involves economics, diverging memory and logic technologies, new applications which require low power, and market dynamics which include company consolidation.
From 2015 to 2016, just three semiconductor segments showed increased revenues: sensors, discretes and linear. What these products have in common is that they use older technologies. Products which utilize advanced technologies such as Microprocessors, NAND and DRAM require a significant amount of investment dollars from the standpoint of both capital as well as R&D expenditures. This is beginning to limit the number of companies that can participate in high-end, advanced technology manufacturing.
Approximately twenty years ago, ASIC vendors were busy assimilating 0.5mm process capabilities and creating large libraries of functional blocks to be utilized in their Standard Cell product lines. While continual advances on the process side of the semiconductor industry were well documented and even expected, the design side of the product creation equation lagged farther and farther behind process capabilities. The semiconductor market evolved to create the SoC approach as a design methodology used the world over to allow silicon solutions of amazing complexity and functionality while
Senior Analyst Rich Wawrzyniak was featured in an article on ChipEstimate.com this week: "EDA Tools, Global Growth and Trade Sanctions Affect IP," by John Blyler, Editor, Semi-IP Systems. Rich shared insights on IP vs. EDA tool markets, IP growth in China and potential effects of trade sanctions on IP.
There have been improvements in many areas of the SoC market from better, more useful Semiconductor Intellectual Property (SIP) to better EDA tools. These improvements become masked by the fact that just as SoC designers master the technicalities and quirks of the current generation of process technologies, a new generation is introduced to the industry and the process resumes again.
