We sit down with Jean-Marc Chery in an effort to find out more about his path to CEO-dom
SHENZHEN, China — A little over five months ago, Jean-Marc Chery became the new CEO of STMicroelectronics. He succeeded Carlo Bozotti, whose latter part of a 13-year tenure at the Franco-Italian company was one of the most turbulent periods in the company history.
Although new to his position, Chery is one guy who doesn’t need a crib sheet to catch up on the Bozotti ST playbook. A little-known fact is that Chery has long been in lock step with Bozotti as his close confidant since ST initiated a drastic restructuring program in 2012.
It was a brutal time. ST, and by extension, Bozotti took a lot of heat. It followed ST-Ericsson’s spectacular failure in the smartphone apps processor business. In the public eye, ST had not only shed most of its past glory but had also lost its soul.
ST still had been known for the set-top and digital TV chipset businesses — even though these consumer markets were losing steam. But even companies close to ST were questioning its future, wondering what would be left in the end.
ST’s top management went to the mattresses and stopped giving interviews. Between 2012 and 2015, ST made no effort to reveal its turnaround strategy, or even say if there was one.
ST’s decision not to pursue FinFET at that time seemed yet another sign of surrender by substantially weakened ST.
EE Times’ first one-on-one interview with Chery last week in Shenzhen, at the Global CEO summit, reveals Chery’s behind-the-scenes role during that period and how Chery undertook a plan that some might have called impossible.
Jean-Marc Chery, President and CEO at STMicroelectronics at the Global CEO Summit last week.
(Photo: AspenCore Media)
It was Chery who proposed that Bozotti steer the company’s commodity imager module business toward development of advanced imagers. To develop “structured light,” ST worked closely with “an important customer” on a full custom solution. ST has scrupulously kept secret the customer’s name, but an iPhone X teardown later revealed Apple as ST’s partner on this project.
For time-of-flight (ToF) proximity sensing, ST first took a standard product to develop proximity sensors for smartphones, but the company also devised a broad set of derivatives such as automotive lidars.
Aware of the potential danger in dedicating the whole team and its resources to develop technology for only one very powerful customer (Apple), ST simultaneously proceeded on other imager product lines with ToF sensors — one based on SPAD and another on photodiode. ST did this specifically to address the Android world.
While demurring credit for a miraculous turnaround that became evident in the last two years, Chery is pretty much the man behind ST’s newly devised imaging strategy. During the interview, Chery spoke straightforwardly — without gesture or drama. Asked if his elevation to the top was resulted from his revamped imaging strategy, he said humbly, “No… I participated, consistent with my level of responsibility. It was my duty.”
Following are excerpts from our transcribed conversation:
EE Times: Some people aspire to become CEO as soon as they join a company. Others never expect it. Did you always want to become a CEO?
Jean-Marc Chery: No, no, no. It was absolutely, not my case.
EE Times: So, this [becoming ST’s CEO] was a move you hadn’t expected?
Chery: It was in 2012 when, with Carlo Bozotti, we reset the strategy of our company, after the disengagement from ST-Ericsson… the company we built with Ericsson to address the chipset for smartphones. We decided to disengage [from the JV], and we reset [ST]. Basically, it’s the strategy — or a part of the strategy — that we are still following today.
At that time, Carlo decided to change a bit the organizational arrangement of our company. Up to 2012, ST had the classic organization of a semiconductor company. It was based on a matrix of operations between manufacturing, technology, product, sales and marketing and corporate staff functions, with very strong financial, treasury, logistic, supply chain, IT, corporate operation, strategy, quality control, etc.
Carlo decided to split into two big business units – embedded processing solutions (EPS) and sensor power and analog (SPNA). These two blocks covered manufacturing, technology and product, with only sales and marketing remained as central corporate operations.
And Carlo proposed me to manage the Embedded Processing Solution segment.
EE Times: This, in essence, became the foundation of ST’s restructuring plan. And this involved what?
Chery: This involved massively transferring people from set-top box, digital TV and microcontrollers to imagers — to really transform the imager business.
EE Times: What do you mean by “imagers” in this context?
Chery: We were making standard commodity modules for imagers. We set out a strategy for the division to pursue more specialized imagers.
It was around this time when Carlo asked me, “Jean-Marc, I feel if we are successful together in the transformation of ST in terms of business, I would like to push you to become my successor.” And it was only at this period that I started to think [about becoming a CEO].
EE Times: How did you respond to Carlo?
Chery: I said, “Well, Carlo, thank you very much.” [laugh] Then, I said, “OK. As usual, I would do my best. First, I work for you. You can rely on me — forever.” And we will see [where this will take us.]
In 2016, we finally become successful in turning around EPS. Then we decided together to bring the organization back to a more classic model. So, ST is back being one company with operations, technology, products, sales and marketing, etc. During the two years from 2016 up to early January 2017, the succession process has been done.
EE Times: So, you were already selected as CEO in 2012.
Chery: Carlo said in 2012, he “will propose” that I become CEO, if we are successful in transformation… Carlo is not like in a kingdom, so he can’t say, “my son will become the next King.” [laugh]
EE Times: So, Carlo’s decision was to pick a successor internally?
Chery: He said, “The transformation of ST will certainly be a long journey, and we have a lot of challenges still. [Bozzotti felt that the] distraction of an external candidate or “somebody who is not well balanced like you would make this process difficult.”
EE Times: What makes you “well balanced”?
Chery: For the sake of transparency, basically, I’ve done all kinds of jobs at ST — customer service, marketing, production, engineering, R&D, product, business development, industrial programs, and business programs. I have a wide set of experiences.
EE Times: When you took charge of Embedded Processing Solutions, who was responsible for sensor power and analog (SPNA)?
Chery: A very young manager called Carlo Bozzotti himself. He was acting head of the division.
EE Times: So, the decision to go with more specialized imagers was made by Carlos?
Chery: No, that was a decision I made in 2012. In 2013, I proposed it to Carlos.
EE Times: So, by “imager,” you weren’t talking about Mobileye’s business?
Chery: Correct. Our business with Mobileye had started since10 years before. That business belonged to the other unit, SPNA.
3 families of imager products
EE Times: Then, what exactly do you mean by advanced special imager products?
Chery: Our specialized imagers basically cover three kinds of products.
We started with a technology called SPAD (single photon avalanche diode). We have enabled with this technology a product called time of flight (ToF). This ToF based on SPAD became a product we introduced first as a proximity sensor in a very efficient, very small module — which we introduced with a very important customer.
EE Times: You mean the big customer you can’t name.
iPhone X cluster of sensors (Source: Apple)
Chery: Right. Second, we developed a set of products — one a CMOS-based global shutter camera sensor. It is an IR camera sensor. We also developed what we call an IR flood illuminator, a module encompassing both ToF and VCSEL (vertical-cavity surface-emitting laser). Then, an analog driver. These three are key elements in a bill of materials for a technology called structured light.
Structured light, up to now, has demonstrated its ability to enable true 3D sensing for face recognition. This was the second, important, block of products.
We decided ToF proximity sensing [could be] a standard product to address a very wide set of applications. Smartphone first, but we also planned for some derivatives like lidar to address the automotive market.
In contrast, for structured light, our approach was a full custom solution, dedicated to an important customer, developed by a team in very tight cooperation with our customer. We took this step so that we can design a strong roadmap to anticipate our customers’ needs and develop an entry barrier…
EE Times: So that competitors can’t catch up?
Chery: They can’t catch up or it would be difficult for them to catch up. In semiconductors, the competition can always catch up. It’s a question of money.
Then, we developed a third family of products. It is also ToF but using different types of technologies. It is either SPAD, but much more complex, which we call direct ToF, or it is indirect ToF based on a photo diode. Such ToF — either indirect or direct — will address the camera for smartphones mainly for AR, VR or these types of advanced applications.
Our ambition is to use two ToF technologies to address more of the Android world… The goal is to position ST as a strong competitor of ams, Sony… and Infineon.
EE Times: It sounds like this specialized imager strategy became a critical element to turn around ST’s Embedded Processing Solution (EPS).
Chery: This was a major decision we took around 2012, 2013 — to maintain an imager division with more than 500 people.
At the same time, we also made another important decision: We said we want to stop advanced CMOS technology development after 28nm and allocate very skilled people in our fab in Crolles to support basically two important technology families: imagers – the three products I just described to you – and embedded flash for microcontrollers.
At the end we have been able to set up vertical activities on imagers with a division of more than 500 people, and a team for technology development and design enablement, which is close to 200, 250 people.
It was very important work for us, totally dedicated, totally focused, under the leadership of a strong person, with strong sponsorship.
EE Times: What happened to the people who worked in this imager that belonged to EPS, after ST decided to bring the organization back to a standard model?
‘ST was never a bottleneck’
Chery: The imaging division continued to report to me, even after I became responsible for manufacturing and technology. After the reorganization, I continued to manage the imaging team for two years, up to the completion of the project with an important customer.
The project was completed — which means contract done, development done, manufacturing ramp-up achieved — last year.
During this period, ST was never a bottleneck for this customer. ST was instrumental in the ramp-up. We considered the team and the technology mature enough to go back into the organization, and we decided to put this specialized imaging sensor division inside our analog and sensor group.
EE Times: This is the first time I’ve heard this story, especially how ST proceeded with its transformation around advanced imaging sensors between 2012 and 2016. This makes you the man behind strengthening the advanced imaging sensor business at ST.
Chery: I participated, consistently with my level of responsibility. It was my duty.
EE Times: Speaking of cutting-edge technologies like SPAD, which ST developed in its imaging division, was this something you already had, or had seen some seeds of the development, before you began working with your customer?
Chery: You have to remember, ST, up to 2014 and 2015, had developed a very wide technology roadmap and portfolio. We developed pure CMOS technology on FD-SOI, in bulk. We developed analog and mixed signal technology — BiCMOS. We developed image sensor technology, CMOS-based. We developed embedded flash technology. We developed RF CMOS-based technology. On another side, we developed BiCMOS, DMOS, BiPolar-CMOS-DMOS technology — so called BCD. We developed power technology on silicon. We developed power technology on silicon carbide (SiC). We started to develop on Gallium Nitride (GaN). We developed MEMS and a set of MEMS.
Decision not to do FinFET
EE Times: In other words, you’ve pretty much covered all the ground of processing technologies.
Chery: Then, we have people really leveraging these technology portfolios, to adapt SPAD technology, which is basically CMOS-based. At that time, it was CMOS 65nm in terms of complexity. The new generation is 40nm technology, done by engineers very familiar with these CMOS technologies. They know all the devices, they know different flavors you can use in process technologies. But beyond that I can’t say much because it’s secret.
The important thing for you to know is that ST in terms of technology development has a very wide spread of capabilities. We decided not to go into FinFET, because we anticipated that FinFET will be a technology for only advanced pure digital devices with high-frequency, high-performance real-time processors, sensor fusion for ADAS. It will not be a technology that will support massively our product portfolio and our application-focused strategy.
While we can always go to TSMC or others when necessary to pursue advanced CMOS, we decided it’s better for us to concentrate on “More than Moore” — with capabilities to mix digital, power, analog. We have a fantastic technology called galvanic isolation to build SoCs in which we have digital chips, analog chips, or digital and power. So, we have all these capabilities.
But we’ve been courageous enough to stop digital development before going to FinFET. And believe me, it was not an easy decision. Especially for me, because I was CTO then.
— Junko Yoshida, Global Co-Editor-In-Chief, AspenCore Media, Chief International Correspondent, EE Times