Ray Zinn Reflects On 37 Years As CEO Of Micrel

What managing a company for nearly four decades taught him could fill a book, but he’s far from done.


Believed to be the longest serving CEO of any company to have existed in Silicon Valley, Ray Zinn does more in retirement than many people would accomplish as full time employees. While his name may not be at the top of most influential people in the valley, the industry may not be the same today had it not been for his contributions. Semiconductor Engineering spoke with him about his past, the present and the future.


SE: You were quite the inventor in your early days. Perhaps you could highlight a few that stand out.

Zinn: My first invention was when I was in my twenties. This was a method for casting solid rocket motors while I was working for United Technologies. Then I had an idea in 1974 while working at Electromask. It was for a tool that is used to image a mask onto a wafer, something that today is called the wafer stepper. The stepper was a little before its time. It was a disruptive technology and involved a lot of groundbreaking steps. At the time we were using 3-inch wafers rather than 12 or 18 inches of today, and that was an issue because the stepper is more valuable as the wafer gets bigger. The feature size was 435nm, which was way behind what people are currently using. When you are in the dark ages and creating something, people don’t immediately see the need for it. It was a challenge for Electromask to see the value of creating such a product. I did convince them to go ahead with it, and it basically sunk the company.

SE: When you did you decide to start your own company?

Zinn: I quit Electromask in 1976 and decided to start my own company. I didn’t know what I was going to start. I tried four or five different things and Micrel was one of them. I am a semiconductor guy and got more enthralled with that than the other businesses, so I let the other ones fade away. Micrel was born in November 1978. I didn’t have any products. So how do you run a semiconductor company with no products? That meant doing services, so I bought a tester and did test services for people doing reliability testing, productivity testing, wafer sorting, etc. I did that for about seven years until I had enough cash flow that I could start developing my own products.

SE: That is a little different than what Micrel finished up looking like.

Zinn: I bought a fab in 1981. I paid $900,000 for it. It was a 3-inch fab and only did bipolar. I wanted to develop a CMOS process, so I needed a product to do that with. I contracted with a couple of people to develop some CMOS parts, standard logic, digital logic, interface types of devices. They helped me to develop the process. I expected that I would be able to sell them as well as use them as a test vehicle. It did allow me to build the process, but the products didn’t sell very well. I didn’t have a product plan or migration from there to some place in the future. Another consultant suggested that I look at power products because there were far fewer people doing that. So that meant developing a new technology and some new products. That got me into Bipolar CMOS DMOS (BCD). That is how we got into analog, and it put us ahead of the game because most other companies didn’t go there until the late ’80s.

SE: If you were starting again today, would you do the same again?

Zinn: I still think analog is the place to be, but today I wouldn’t consider getting into the semiconductor business unless I had something very disruptive. A process differentiation is more important than a product differentiation. Right now, the area with most traction is stuff above 20V. I would look at GaN even though it is difficult to do, but I like things like that. I would look at silicon carbide or processes that offer a very unique capability. I would not touch anything that the Chinese are going after, which is low-cost consumer. I want to see higher margins and something that is not easy for other people to copy. It is difficult with the fabless model. How are you going to get anyone to help you develop process? So you would need to have a fab to do this.

I would invest in a company if there is a need in the market for a disruptive technology that a competent team is developing. I believe customers will pay you top dollars if the technology being developed solves a real problem that the consumers are facing. We need to keep our eyes open and our minds sharp to find those opportunities. We need to stay alert to make sure we are not falling in love with our own ideas, and stay objective in assessing what is a real market need and what is a pipe dream.

SE: How long do you think it will be before China becomes a semiconductor powerhouse?

Zinn: First of all, when we think about the advancement of China in the semiconductor field, we have to differentiate between Taiwan and mainland China. Taiwan is already a semiconductor powerhouse. In fact, one can argue that Taiwan is in many respects ahead of the United States, especially on the semiconductor manufacturing prowess.

As for mainland China, the Chinese government has declared the semiconductor field to be a strategic technology focus. As a result, we have seen many attempts by the Chinese investment community to acquire American semiconductor companies. We just witnessed the attempted acquisition of Fairchild by CEC, which is a government funded technology umbrella of companies. In addition, many Taiwanese companies have announced large investments in mainland China. For example, just a few months ago, TSMC announced that they will build their latest fab in Nanjing, China.

Therefore, I believe much of the advancement of technology in mainland China will not be organic but rather through acquisitions. And we have already seen an accelerated pace on this front. I expect the Chinese will target larger, more dominant companies with a significant manufacturing base already in China. This is a multi-year effort which I think will take 4 to 5 years of focus by the Chinese government and investment communities.

By the way, it is somewhat unfortunate that the U.S. government in many respects has given up on the semiconductor field. This technology is important for U.S. manufacturing and job creation, but also for our national security. Apparently the Chinese have realized this, but the U.S. government has given up on it. The U.S. semiconductor manufacturing industry is smaller than the California wine industry.

SE: You retired in July of 2015 when Microchip bought Micrel. What are you doing these days?

Zinn: I consider myself to be an author and venture capitalist. The title of my book is Tough Things First. I have invested in one company already, a software startup (OnRule), and I am talking to other companies that are interested in using the concepts contained within the book. I am also involved with Silicon Catalyst. It is not for the financials, but in validating the principles that I learned at Micrel.

Tough Things First is intended to be an educational book. We are trying to get it adopted within Universities. I have taught at Draper University, a venture aimed towards Silicon Valley entrepreneurs. We are also working on another book, tentatively titled the Zen of Zinn, but Tough Things First is a standalone book that covers how to run a sustainable business.

I create two blogs a day for my website, I write at least two articles a month, two podcasts a week, lecture twice a month, and then work with companies where I think I can help. This afternoon I fly to my ranch in Montana where I ride horses, enjoy riding ATVs, doing a little target shooting and fishing.

SE: Any final thoughts?

Zinn: Semiconductors is a great business and we are ruining it because greed gets in the way. We should be doing things for mankind, not for ourselves.

Related Stories
What Will China Do Next
IC M&A activity hinges in the balance between the yuan devaluation and U.S. interest rate hikes.
What China Is Planning
With about $120 billion in funding at its disposal, China is looking to buy key technologies it doesn’t already have.

Leave a Reply

(Note: This name will be displayed publicly)