Interview w Microchip CEO Steve Sanghi „China steals our technologies to take over the market“

Microchips CEO Steve Sanghi mit DESIGN&ELEKTRONIK-Chefredakteur Frank Riemenschneider.
Microchips CEO Steve Sanghi with DESIGN&ELEKTRONIK Editor-in-Chief Frank Riemenschneider.

Microchip's CEO Steve Sanghi usually rejects interview requests. For DESIGN&ELEKTRONIK, he made an exception and explained Microchip's rise to 18 billion, the biggest challenges in 27 years of CEO activity and why he finds Trump's punitive tariffs against China good.

Steve, while most of your competitors have significantly shorter CEO terms, you have been Microchip's CEO for more than 27 years since October 1991. Wouldn't you be tempted to look again for a new challenge, e.g. with a larger competitor?

No, absolutely not. Apart from the fact that Microchip itself has grown considerably through all the acquisitions, every colleague of mine in the management team has worked for Microchip for at least 18 years. It's our special culture that makes the company so attractive: Every August, our shareholders meet to elect a new supervisory board. We in the management team - and I myself have introduced this system - are resigning unanimously and the new board can appoint anyone else as new managers. But this has never happened before. I am still full of energy for this company and am not looking for any new challenges - except at Microchip itself.

When you look back on more than 27 years of CEO activity, what were the most challenging years and what challenges did you face?

The biggest challenge was right from the start: Microchip had only $16 million in revenue and incurred losses. It was about finding investors, optimizing our processes and ultimately making the company profitable so that we could go public. The next challenge was our huge growth of up to 50% per year, which they had to manage first. And can the economic crises came in 1996, 2000/2001 and 2008. The last 8-9 years have been about integrating all the acquisitions into our company. These are cultural challenges, trainings for new employees and a lot of other things. No, we don't get bored (laughs).

Microchip has been profitable for 112 consecutive quarters. How did you manage to remain profitable even during the major economic crises, while your competitors generated losses running into billions?

While our competitors achieved a gross margin (a business ratio representing the difference between sales revenue and cost of goods sold) of 32%, we had 60%. While the competition generated an operating result in the single-digit percentage range, we had 30 %. Mind you, in non-crisis times. In 2009, these figures - the worst case for us - fell by 18%. If you had 8% before, you will logically be negative and make losses. If you had 30% before, you will logically still be profitable. That is one thing. The other thing is that growth and profitability together are our most important financial indicators. Every employee knows that. When we go through difficult times, we all make sacrifices together to save money and keep the company profitable. That is our fundamental common principle for all we do.

In your annual report, you are praised exuberantly in several places, specifically Microchips' growth, profitability and strong positioning in key markets are attributed to your management skills. Is Microchip Inc. really a Steve Sanghi Inc. by now?

(laughs) No, fortunately not, because that would mean that if the CEO left, the company would go up against the wall. The opposite is the case: I have implemented a culture and processes that Microchip could continue very well without me.

With almost 2% of all shares, you are the fourth largest shareholder in Microchip and thus hold more than 200 times more shares than your fellow board members. Is your commitment to Microchip through a share purchase so strong or that of your colleagues so weak and why is this so?

(laughs) I didn't know, you have to ask my wife. The reason for this discrepancy is that I was the only one on board at the IPO and then as CEO, and my colleagues gradually took on their roles in the management team only later.

After you had already made two major acquisitions with Micrel and Atmel in 2015 and 2016, it was Microsemi's turn this year. To finance the purchase, you took out new loans worth 8 billion dollars. You yourself list numerous risks that can occur during takeovers, e.g. the loss of important employees, customers, business partners or problems with different corporate cultures. The acquisition of EnergyMicro by SiliconLabs is a warning example of cultural differences. Why do you still think Microsemi is worth the high price and risk?

We have now completed 20 acquisitions and all have been successful. Why? Point 1: We fully integrate the companies into our organization. That means for example to completely transfer their Oracle- or SAP-based systems into ours, there will only be one customer relationship from the organization, one kind of remuneration system for the employees, one distribution structure and in 2-3 years nobody can recognize any more, this is Microchip and this is Atmel or Micrel or Microsemi after we have integrated the old business areas into our organization. We have done a survey at the old Atmel sites 2.5 years after the acquisition, the results e.g. on employee satisfaction are very similar to those we see at traditional microchip sites. The secret of our success is full integration: financial, cultural, pricing, business units, sales, operations and factories. This allows us to control all the risks. As far as Microsemi is concerned, the biggest risk was the high inventory level of 4 months. We have now sold it off for 180 million dollars to the extent that it is at the same level as our own, namely only for 2.6 months. Most of the potential risks still mentioned in the annual report are already behind us.

After all, Microsemi has good customers who are working with the government and are now to serve as door openers for Microchip products as well. Microchip was always rather weak in this environment, why actually? You manufacture in the USA, have created a lot of jobs here, should politicians like that?

That is true, we have inherited many good contacts in the government organisation. Until now we were not dependent on political support because of our orientation, you only need it if you are in the military sector and/or in aviation. We were simply too busy expanding our business in industry and the automobile, for example, to have had the capacity to build a completely new relationship with the government. From Atmel we have already inherited a comparatively small military and aviation business with a volume of 35 million dollars and the best contacts e.g. to the European Aviation Authority ESA. The goal is, of course, to sell other microchip products in these markets as well, but this is a long-term process.

Not only have you inherited great products from Atmel, but you have also inherited various legal disputes which, to my knowledge, have still not all been settled. Are you afraid that this could happen again with Microsemi?

As far as Atmel is concerned, only one legal dispute is still pending. And at Microsemi we're talking about one or two litigations and that in a history of 28 years - that's nothing, please compare that with Intel or Tesla or Amazon, every successful company will be sued, you and I won't be able to change that. Our market value has risen from the original 10 million dollars to 25 billion dollars, which of course arouses desire.

Let's move on to manufacturing. You say that a high proportion of in-house manufacturing is part of your strategy to achieve the lowest costs and at least partially achieve the margin retained by the foundries yourself. On the other hand, you write that a growing proportion of your products are manufactured in foundries, in 2018 it was already 42% and with the purchase of Microsemi the share will continue to increase. How does this fit in with your in-house strategy?

It's relatively simple: what we can produce in-house, we can produce in-house. As you know, we are limited here to 200 mm wafers. A 300 mm production for 90 nm structures and below is extremely expensive and requires very high investments. With our in-house production we can go up to 110 nm. With ever smaller production geometries, the share of our foundry production is constantly increasing, by approx. 1 % per year from the current 42 %. Our competition is the same, so we have no competitive disadvantage.

While your competitors have microcontrollers manufactured in 40nm or 28nm processes in the future, their most advanced process today is 110nm in Fab 4. Regardless of other drawbacks such as higher power consumption in active mode, you also have higher material costs over a larger chip area. How does this fit in with your low cost strategy?

We have foundries manufacturing in 55 and 40 nm, and we're also going in the 28 nm direction. In this respect, there is no disadvantage compared to the competition with regard to the parameters you describe. As far as low costs are concerned, with these small production geometries you no longer have a cost advantage with in-house production. Why is this so? Because they would need extremely high volumes to achieve this advantage. With the exception of Intel and Samsung, these volumes can no longer be achieved by a single manufacturer, but only by a foundry.

Unlike your competitors, you manufacture controllers and memory in a single process, which keeps costs low, but makes specific optimizations for controllers on the one hand and memory on the other impossible. Why are your products still competitive? Is it just the lower price?

Our storage business now accounts for only about 5% of our sales, with a downward trend. We only develop new technologies for microcontrollers, i.e. the connection between controller and memory described by you no longer exists.

After the acquisitions, I can no longer specify exactly what this means - you have 6,000-7,000 microcontrollers in your portfolio, which generate by far the largest share of your sales, namely 55 %. How has the share of 8-, 16- and 32-bit controllers shifted in recent years?

As you know, we do not report exact sales figures for the individual segments. There are market researchers like Gartner or DataQuest who publish figures that we never confirm or call wrong. We are the world's No. 1 in 8-bit, No. 4 or 5 in 16-bit, and No. 4 or 5 in 32-bit.

What is your 32-bit strategy? After your PIC32 with MIPS CPU, you have now inherited Cortex M CPUs from Atmel Arms.

We will continue both lines, as there are a lot of customers for both today, even though in our 32-bit business today the arm portion we inherited from Atmel is larger than the MIPS portion. We develop identical peripherals and identical manufacturing processes for both lines.

After the bankruptcy of Imagination, the MIPS roadmap looks very bad. What have you promised your customers here?

The support is still good. What we need, we get. As you know, we develop our own development tools and debuggers, there is no dependency. What we needed were the cores, and we got them. We don't need another MIPS core.

You've introduced many controllers with proprietary analog IP in recent years. Is this the way to stop the ongoing price erosion in the MCU industry by developing very specific products for all kinds of applications instead of general purpose controllers?

The primary reason is a higher integration density in order to be able to offer the customer an inexpensive solution. Of course, it is also a more differentiating product, whose price stability is higher than with pure commodity products. I don't mean to say that the price remains constant for all eternity, but at least at a level where we can earn money (laughs). It would be great if prices stayed stable forever.

Another important building block in your strategy is easy-to-use and scalable development tools. Can a developer today program everything from the PIC8 to the 32-bit arm MCUs purchased from Atmel with identical tools?

This is true for all Microchip products and also for some of the products inherited from Atmel, but not for all Atmel products. Many of our customers now also want to use arm products and are familiar with the microchip tools, former Atmel customers usually want to stay with the Atmel tools and some even use both. We give the customer the choice by continuing to offer both tool lines. For future products there will be a uniform tool support, which you can no longer program with the existing tools (note: Atmel-Studio).

If you look at the sales development of your microcontrollers and your analog business, your storage business is growing at a disproportionately low rate. What are the reasons and what is your strategy for this business area?

Our storage business is not strategic and only serves to complete our customers' solutions. We don't want to abandon our storage business, but not to grow or gain market share. Nor did we make any acquisitions because of the storage business, but only to drive our microcontroller or analog business forward. In this respect, we can live well with the below-average growth in the storage business.

You say that the technical expertise of your salespeople and strong technical customer support are important for your sales success. Given the shortage of qualified embedded engineers in the labor market, where do you get these specialists in sufficient numbers anyway?

This is indeed a big challenge, as it is for our competitors.  When we were a small company with only a few microcontroller families on offer, we had a global distributor for all products that was active all over the world. Today with more than 6,000 microcontrollers plus analog products plus wireless products plus network products plus Ethernet plus USB plus microsemi products this is no longer possible. We have changed to train specialists for individual functionalities such as engine control, digital power, touch, automotive and so on. We are not able to have these experts anywhere in the world. We therefore have experts in Taiwan who cover all of Asia via remote support or travel, or experts in Germany and England who cover Europe. With this approach, fewer specialists can cover larger regions.

You make 30% of your turnover in China and Hong Kong alone. In these regions, economic growth is slowing down, will Microchip also be affected and to what extent?

We are already affected, our current quarterly forecast for September to December is 7.5% lower sales and the main reason for this is China. The trade tensions between the USA and China are very negative, currently Chinese products imported into the USA are subject to a duty of 10%, possibly even 25% after the 90-day grace period expires on 1 January 2019. As a result, Chinese manufacturers are already reducing their inventories. Texas Instruments even expects sales to fall by 11%, Cypress by between 7.5% and 8% and Maxim by a similar amount.

Do you and your competitors in the USA talk to politicians about how you are negatively affected by customs duties?

No, I am not talking to politicians because I believe that I can have more and better influence by managing our company. Apart from that, I do not take a fundamentally critical view of customs duties. China has benefited from the US economy for decades, they copy our products and IP - 90 % of all Windows installations in China are pirated, and that is only a passing fad - force us into joint ventures with local companies that steal our technologies and then take over the entire market. Apart from that, our products destined for China are subject to a 16% tariff. Trump has done what we are suffering from in the short term, but in the long term trade relations must be fair. China will not fulfil its fair trade obligations until there are limits.

You say that you are significantly affected by product piracy in China, i.e. your products are being copied. Can you estimate your damage and how can you defend yourself against it?

Let me tell you the following story: A manufacturer of household appliances such as thermostats in Shanghai used our microcontrollers because they work very quietly in environments with high EMC exposure. They then set up a subsidiary for the design and manufacture of semiconductors and transferred all the know-how they had from us to this company and started copying our products with very similar technology. We sued them as well as many other companies in Taiwan and China, but it is extremely difficult to get justice in China. It is always delayed and postponed and in this case after 10 years there is still no verdict.

And the damage?

Well, an advantage of our industry is that the development is very fast and that the copies refer quite fast to older products. For example, if the new products are more energy efficient than the copies, they will of course be used. This means that the damage is still manageable. It's like a race between hare and hedgehog.

Around 54% of your turnover is achieved through distribution. The latter now increasingly wants to appear not only as a salesperson but also as a consultant to the customer and to tap into a higher part of the value chain. Are you afraid of losing influence on the end customer and becoming, so to speak, the plaything of distributors?

No, I'm not that afraid. No distributor can provide technical support for all our products and those of our competitors, so virtually all major customers turn to us directly when they need support.

I was a little surprised at the sentence that you sign contracts with extremely high risks for winning designs, such as unlimited liability for product defects and their consequences. What kind of customers from what business areas have such great power to force you into such contracts, especially since you don't, like other competitors, have large customers who alone account for 10 or more percent of your total sales?

We did not sign such contracts with unlimited liability. Not even with our major customers such as data centers, PC manufacturers or automobile manufacturers, even if they ask for it.

Oh, that surprises me now after reading...[1]

Maybe we did that 15 years ago. As a rule, small manufacturers do this because they have less to lose. Competition with large competitors is difficult and if they want to win a contract, they may agree to such contract terms. When we bought Micrel, there were some such contracts, we ended them all.

Why did you suddenly have to pay so much more tax this year than in previous years, which ruined your profits?

That was a unique effect based on the new tax legislation. Previously, you had to pay 35% tax on all profits that you generated abroad and transferred to the USA. With the new laws, profits could be taxed whether you brought them to the US or not. Assets abroad were subject to a one-off special tax of between 8 and 15.5 percent. Future foreign profits are to be punished with about 10 percent. As a result, many companies have returned their foreign profits to the USA. It was a one-off effect.

Steve, thank you very much for your time and your open answers!

 

M.Sc. Steve Sanghi

is chairman and CEO of Microchip Technology Inc. He is also the company's founder and first president.

Sanghi was named President of Microchip in August 1990, Chief Executive Officer in October 1991, and Chairman of the Board in October 1993. Prior to joining the Company, he was Vice President of Operations at Waferscale Integration, Inc. another semiconductor company from 1988 to 1990. From 1978 to 1988, Sanghi was with Intel, where he held various management and engineering positions, most recently as general manager of programmable memories.

He holds a Master of Science degree in Electrical Engineering and Computer Science from the University of Massachusetts Amherst and a Bachelor of Science degree in Electronics and Communications from Punjab Engineering College, Chandigarh, India, which he received in 1975.

Together with his wife Maria T. Sanghi, Steve Sanghi is Microchip's fourth largest shareholder and his more than 4,600,000 shares [2] represent nearly 2% of the company's shares and were valued at nearly $350 million at the time this article went to print.

 

References:

[1]: Quote from the Annual Report of July 12, 2018 on page 77 under the item „Our contractual relationships with our customers expose us to risks and liabilities“ (https://www.microchip.com/sec/annual/FY18/2018%20Proxy%20Statement%20and%20Annual%20Report.pdf):

„In order to win important designs, avoid losing business to competitors, maintain existing business, or be permitted to bid on new business, we have been, and may in the future be, forced to agree to uncapped liability for such items as intellectual property infringement, product failure, or confidentiality. Such provisions expose us to risk of liability far exceeding the purchase price of the products we sell under such contracts, the lifetime revenues we receive from such products, or various forms of potential consequential damages.“

[2]: Annual Report of 12 July 2018, page 22: https://www.microchip.com/sec/annual/FY18/2018%20Proxy%20Statement%20and%20Annual%20Report.pdf