The problem I have with semiconductor analysts and media today is that they rarely have depth in what they are talking about. Some because they have never actually worked in the industry and others because they have not worked in the industry since the 1970s. One famed analyst even repeated the mythical “Fabs cost $10B” generalization to spread his self-serving doom and gloom of future semiconductor success. What kind of Fab is that, logic or memory? Where is it built and how many wafers per month does it produce? All fabs are not created equal and for the record: TSMC builds logic fabs in Taiwan for $5B and the new one in China is budgeted at $3B but I digress…
The thing I like about Weston Twigg, Director of Capital Markets Research at Pacific Crest Securities, is that he is an actual semiconductor professional having worked at both Samsung and Intel at the process level (he has an MS degree in Chemical Engineering to go with his MBA). Weston thoroughly understands and respects the foundries which very few analysts do and his presentation reflects that:
Bulls, Bears, Bits: Investor Views on Changing Semiconductor Industry Dynamics
Investors are struggling to keep up with changes in the semiconductor industry. Semiconductor economics are becoming challenged as process complexity increases; compute functions are shifting from local compute to the cloud; PC, tablet, and smartphone demand is decelerating or declining; and signs of maturation are emerging as M&A becomes a dominant theme. With all of the uncertainty, however, investors generally remain upbeat with two prevailing investment themes: story-driven stocks and M&A beneficiaries.
In our never ending quest to find “The Next BIG Semiconductor Thing” Weston and I share a similar view: From the demand side Internet of Things (IoT) and cloud are ramping up but if demand for smartphones decelerates (which it has) what else will drive big die size chips? Weston highlights three potentially big trends which I agree with, absolutely:
Investors are contemplating disruptive shifts:
- Semiconductor economics are becoming challenged as process complexity increases
- Compute functions are shifting from local compute to the cloud
- PC, tablet and smartphone demand is decelerating or declining
- Signs of maturation are emerging as M&A becomes a dominant theme
Questions that investors are working on:
- Can technology-driven (leading-edge) companies be successful if node transitions slow?
- Which trends could drive more leading-edge demand?
- Are there companies positioned at n-x nodes that might perform well?
- Which segments should remain relatively high growth?
Despite the uncertainty, investors generally remain upbeat over the long-term on semiconductors
My personal view of the next big thing is along the same lines. The products we have today will need to be increasingly smarter which will result in much larger die sizes and more wafers at the leading edge. It will also lead to increased design complexity and stricter power requirements which will benefit the fabless semiconductor ecosystem as well.
In my humble opinion, IoT will go through a similar transformation. Today’s IoT chips are pretty dumb if you think about it, which I have. Before you know it the “IoT SoC Revolution” will begin and it will be the lather, rinse, repeat shampoo cycle yet again.