Semiconductor Chip Shortage in 2021

Hi, Al Perry here, I’ll be introducing John East in a moment, but for perspective, and to understand why there is a global semiconductor chip shortage in 2021, we have to go back and see how we got here. To do that, I’ll take some excerpts from John East’s Silicon Valley – The Way I Saw It, and that will help explain why we are where we are. It’ll be the author, John East’s words starting in the topic: Moore’s Law.”

What’s a Fab or is it a Foundry?

In the semiconductor industry, the factory where semiconductors (or “chips”) are made, is called a “Fab,” short for fabrication plant. It is also called a foundry. These Fabs are complex manufacturing plants that have extremely high initial costs to build and maintain. To build one, expect to invest billions of US dollars, with the initial cost being close to 10 billion. This is why most semiconductor design companies, do just that, design the chip. Then they send a sort of blueprint, called a mask SET (or photomask SET) to the Fab, where the photomask is projected onto a silicon wafer. There are many layers, one etched for wiring, one for millions or billions of transistors, and so on the process keeps repeating. It’s mind-boggling! Let’s go back and look at some stories from Silicon Valley – The Way I Saw It. John can explain it better than I can, and it will give us perspective of why we have a shortage today.

Moore’s Law

From Episode 4, “The Integrated Circuit.”

Here’s some perspective on Moore’s Law from 1965. When Gorden Moore first articulated his “law,” he was predicting that the number of components on a chip would double from a starting point of 32 components in 1964 to a gargantuan 64 components in 1965. The logarithmic graph that he published was so bold as to predict a time when there would be 64 thousand components on a chip. To the guys in the fab trying to make these things in the Fab, it sounded crazy! I know it sounded crazy to me when I got there in 1968. Gorden firmly believed in the 64K number, but if you had asked him about 64 million, he probably would have thought you were nuts. Today we’re doing on the order of 64 billion. Wow! That sounds like Alice in Wonderland stuff, doesn’t it?

The Foundry

From the customer’s point of view, those who wanted to purchase semiconductors, by the time 1988 rolled around they were demanding microprocessors, memories, and gate arrays. Customers could use gate arrays to design their own LSI chips. They liked that!! And virtually all systems use memories and microprocessors. The Japanese saw this coming. They coined the term “master slice” (the term they gave to gate arrays). Then they embarked on an effort to conquer the market for the three M’s: Memories, Microprocessors, and Master Slices (gate arrays). It seemed for a time that they might succeed. Basically, their plan was to dominate the business of manufacturing ICs. The good news is that they didn’t succeed.

(The bad news for the Semiconductor industry in the US is that Taiwan and South Korea eventually did.)

Coincident with this was the advent of the foundry. Morris Chang saw the future!! TSMC happened! TSMC took on the responsibility of handling the processing and manufacturing issues that were the bane of the early Silicon Valley crowd. Semiconductor companies no longer had to understand Iceo, or work functions, or mobile ions, or minority carrier lifetimes or any of the many other time-honored problems that had existed since the days that I was a hands-on engineer. Those problems would be handled by some man or woman in Taiwan whom you had never met and would never need to meet. The world had changed. In1980 the typical successful semiconductor CEO understood semiconductor physics, fab processing, and transistor level circuit design. By 1990 that set of knowledge was already rendered nearly useless.

By 1990 you needed to understand the architectures of very specialized, complex systems and the end markets of these systems. Were there exceptions besides the three M’s? Standard products that could be sold to nearly everyone? A few. Notably some analog functions and programmable logic. But not many.

Today I’m spending quite a bit of my time with semiconductor start-ups. Over the past four years, I’ve been working with Silicon Catalyst. Silicon Catalyst is an incubator who incubates only semiconductor companies. That has allowed me to get a good look at dozens of start-ups in the field. My take? Successful semiconductor CEOs today often know little or nothing about semi- conductors per se. Moreover, they don’t even care!! Nor should they!!! They know about the market they’re trying to serve. They understand the hardware, firmware, software and applications related to their chosen market. Building blocks of ever-increasing complexity went out the window a long time ago. Today the call is for a complete solution to an existing problem. Let TSMC worry about how to make the things. The change in the attitudes of customers caused a corresponding change in the strategies of the traditional IC companies. In the case of AMD, “Building blocks of ever-increasing complexity,” morphed into “High performance computing is transforming our lives.” It was a big change. Companies that don’t embrace change wither away. In the case of AMD, they made the changes well. The stock market now pegs their value at north of $30 billion!!!

TSMC and Morris Chang

Morris Chang was born in China and moved to the United States when he was in his teens. He attended Harvard and MIT before eventually getting a PhD from Stanford. His first job was to develop a line of transistors at Sylvania Semiconductor, but most of his career prior to TSMC was spent at Texas Instruments where he eventually ran TI’s entire semiconductor business. TI was by far the largest manufacturer of semiconductor products at that time. Morris was the person who put TI’s forward pricing model in place. The forward pricing model involved quoting what seemed like ridiculously low prices for distant future deliveries of their products. He saw that, because of the progress being made, a low price today would be a reasonable price tomorrow so TI would, in fact, be able to be profitable at those prices.

The benefit came when competitors saw the low price and decided to stay away from that market because of the crazy, seemingly unprofitable pricing. The strategy worked for a while but eventually many, many companies entered the TTL market. Still, TI was able to maintain dominance in the market.

From TI, Chang moved on to become CEO of General Instruments for two years. Then he returned to Taiwan. The start-up craze was still alive and well in Silicon Valley. Morris saw that there were countless companies with ideas for nice semiconductor products – These products had the potential for sales of, maybe, 50 or 100 million dollars a year. He also saw what was happening to the cost of a top-notch wafer fab – the area where ICs were manufactured. It was headed north. The cost was soon to pass through a billion dollars on its way to ten billion dollars – roughly where it sits today. A young company with a $50 million per year product couldn’t afford a fab to make that product in. Morris saw the potential for a semiconductor foundry – a huge fab that served the needs of all the smallish semi-conductor companies. TSMC was born. TSMC (Taiwan Semiconductor Manufacturing Company) is now a $50 billion per year company who serves IC companies all over the world.

Conclusion – The Shift of power

In the early days every semiconductor company had their own wafer fab. The capabilities of their fab determined, to a large extent, the success of the company.

Today, though, only the very largest semiconductor companies can afford their own wafer fab (Intel, Samsung, and Micron for example). All the rest use semiconductor foundries — third party wafer fabs that offer manufacturing services to the rank and file of semiconductor companies. Morris Chang saw this sea- change coming and started TSMC – Taiwan Semiconductor Manufacturing Company. TSMC soon grew into a huge company with tremendous influence in the semiconductor market.

In those early days, the number one worry of a semiconductor CEO was his or her fab and the processes that ran in the fab. At times that was an all-consuming set of worries because manufacturing a semiconductor is an extremely complicated process. Once TSMC relieved semiconductor companies of those worries, the CEOs were free to spend their time worrying about products, customers, and markets — the things that really should matter. Today a semiconductor company designs a product and sends a mask-set (Roughly speaking a set of blueprints) to the foundry and the foundry takes it from there. It sounds great, but there is a problem — the center of excellence for integrated circuit manufacturing has moved from Silicon Valley (as well as Texas where Texas Instruments had their fabs and Vermont where IBM had theirs) to Taiwan. Given the importance of semiconductors in our everyday life as well as in the defense of our nation, the situation today can be a bit troubling.

When any kind of global trouble makes the waters murky, like a pandemic, the supply of a vital resource is now across a vast ocean. With the rising cost of fuel, and a shortage of labor, the industry has lost control of its prized innovation. As consumers, we pay higher prices and wait for any product that uses a semiconductor, that’s just about everything.