As we have moved to virtual briefings, I have increasingly been excerpting short video segments (with permission) as part of my Analyst Cam series.
This time it is Jeff Howell, Global Vice President, Head of High-Tech Industry Business Unit at SAP.
We have all heard about the geopolitics and risks of continuing to be dependent on SE Asia for a steady supply of chips. We have heard about the chaos in the auto sector with the inadequate supply of that increasingly important component as vehicles get smarter and electrified.
Jeff takes us through a much more nuanced perspective on the complexity, the long lead times and capital intensity of the semiconductor sector. I was first introduced to him as I was helping SAP with research on the recently released Business as Unusual book (available on Amazon – click on badge on left)
Here are just a few excerpts from the conversation in the video below:
- The semiconductor industry amounts to 0.3% of the U.S. GDP, yet semiconductors are required to produce 12% of the U.S. national output. It is even more pronounced in Japan - with impact on 30% of their national output
- If you and I decide to go into the chip making business we would need to come up with $10 to $15 billion and spend about two to three years building that production capability
- Over 400 chemicals are used in a combination of solids and gases to make those chips. Over a dozen noble gases on the periodic table are critical supplies
- Roughly 30% of the neon that's needed in the semiconductor industry comes from Ukraine
- 2 to 4 million gallons of ultra-pure water are required every day during the chip production process. Almost half of all the world's chips are manufactured in Southeast Asia. Well, a couple years ago, there were droughts in Southeast Asia
- Every wafer produced in a typical fab generates two to three gigabytes of data and you multiply that by 24 wafers in a lot and 1000s of lots per day, it's very easy to come up with terabytes, and generate even petabytes of data every day.
- The little known fact is more than 80% of that data is never even looked at - not enough process engineers who can examine that data
He spends a fair amount of time talking about networks like Catena-X which is auto industry-centric but heavily depends on semiconductors. He cites use cases at companies like NVIDIA and Wolfspeed (a critical supplier of silicon carbide) and partners like PDF which helps companies analyze that mountain of operational data for better yields and other outcomes. He also discusses how SAP is blending financial and operational data for the sector.
It is a master class on critical components for so much of the economy and the unique systems the industry needs. Very nicely done in about half an hour – he could easily have gone much longer