ST’s Bozotti on ‘back-to-silicon’ differentiation
“The market is on an upturn, despite an IMF revision taking down global GDP forecast,” said Bozotti. “We saw a good trend in bookings in Q3 and this continued in October and the upturn is broad.”
But an examination of the numbers shows ST’s main business group revenues were still marginally down in Q3 compared with a year before. The overall revenue increase was largely down to success with a time-of-flight ranging image sensor that, for now, is part of the “others” product group (see ToF sensor success lifts ST’s third quarter).
“In Q3 distribution sales were up 12.2 percent and the book-to-bill ratio was well above one. In the second half all the product groups will contribute to year-on-year sales growth. The year-on-year is expected to be up 11.2 percent fourth quarter over fourth quarter,” said Bozotti attributing this to strong smartphone, automotive and industrial demand.
“MCUs, automotive, speciality image sensors are all contributing to growth,” he said. But what about MEMS, which has been a stalwart part of the ST product portfolio during the ST-Ericsson transition? It would appear that ST’s strategy of extending from success with MEMS in smartphones and consumer applications to automotive and industrial applications is proving more difficult to execute than market leader Bosch’s migration in the other direction; from automotive to consumer.
Bozotti answered: “We are certainly growing in MEMS. In automotive we need to do more; both in sales and design wins. So yes consumer is the biggest part [of MEMS sales] but it is much more variegated than before. Before it was one customer one product. Now it is multiple sensor types spread over many products.”
And Bozotti was also positive about ST’s success with microcontrollers. “We were the first to enter the distribution market with ARM microcontrollers. We have 700 different microcontroller types and 40,000 customers. We want to do more but a very strong ecosystem has developed. The Nucleo boards provide a lego approach to supporting other products through a standard physical interface and form factors.”
Bozotti added: “Silicon technology is an imperative differentiator for MCUs. We are on 40nm and the next step is FDSOI at 28nm with some form of embedded flash.” FDSOI stands for fully depleted silicon-on-insulator, a manufacturing process said to provide advantages in terms of low power and simpler manufacturing over FinFET technologies but at the expense of requiring an engineered wafer containing a buried oxide layer.
However, Bozotti declined to comment on whether ST would go beyond its ARM license with the addition of support for machine learning, a particularly hot topic at present.
Next: Pushing automotive
Bozotti said automotive, with the surge in interest in ADAS and autonomy, was a particularly fertile ground for ST where it is already strong. Bozotti said ST could capitalize on two trends in automotive electronics; digitalization and electricification.
In 2015 ST ranked as about the fourth largest automotive chip vendor behind NXP, Infineon and Renesas and about 8 percent market share, according to Semicast Research. “We want to keep going in the traditional areas where we are strong – smartpower for the power train and car body and also in infotainment.”
In car infotainment ST is a chip partner for SiriusXM satellite radio, it also provides tuners for radio, GPS, telematics and mid-range processing. “But not the application processor,” he said. “We also want to push our silicon-carbide power semiconductor technology. 2017 will be the first important year for silicon-carbide. Silicon-carbide replacing IGBTs will be an important trend.”
“And then there is advanced safety with the Mobileye generation three [EyeQ3] and generation four [EyeQ4] in FDSOI coming next year. And then the EyeQ5 in 7nm FinFET technology which will have 10 application cores and perform sensor fusion computing.”
A second key partner in automotive is Autotalks Ltd. (Kfar Netter, Israel), which provides vehicle-to-anything (V2X) communications chipsets (see Denso selects Autotalks V2X chipset). “And then there is MEMS and sensors for automotive,” said Bozotti indicating that this rounded out a broad offering.
Next: What about Qualcomm/NXP?
But lying fourth in an automotive chip vendor ranking and with Qualcomm expected to takeover NXP Semiconductors it could be perceived that ST’s participation is threatened.
“They would be a very important player but our customers do not always like bundles. Qualcomm does not bring much market share so the shares will stay the same,” Bozotti countered.
Qualcomm does bring wireless technologies such as LTE and 5G modems and high-end Snapdragon processors as well as automotive wireless charging, which some could see as the future of automotive digitalization and electrification.
“We will not be in the modem business or in the high-end processors. But what we are doing is technology-based and very difficult to compete against. Some 30 percent of ST is automotive and silicon-carbide is extremely strategic. ST has R&D spend of $1.4 billion per year.
There may be strong R&D spending but in the area of FDSOI it has been argued that years of investment has yet to produce a return in terms of sales. With regard to FDSOI and the fact that there appears to be a lack of breakthrough applications and customers Bozotti said: “We have some important projects. In communications infrastructure we have business. And space is important because of the radiation hardness of FDSOI. And the EyeQ4 from Mobileye in FDSOI is a tremendous success. It is in many high-end cars in 2017.”
He added: “The combination of FDSOI with some sort of novel flash memory is the way to go from 40nm to 28nm MCUs,” Bozotti argued adding that this was both for general purpose and automotive MCUs. “It takes time,” he said.
But time at ST could be something that Bozotti does not have in abundance.
Next: Management transition?
According to reports that appeared in April 2016, ST’s board of directors had already started reviewing candidates for a replacement CEO who could be brought in prior to the expiration of Bozotti’s contract in May 2017. This would provide for an overlap and smooth transition. But Bozotti declined to confirm or deny that, or expand on the situation.
“No comment. The focus of the company is on growth. The exit from ST-Ericsson is done. And now we want to grow,” Bozotti said.
But has ST’s time working through its ST-Ericsson problems left the company weaker and smaller and has it been distracted while others have been participating in industry consolidation?
“We have the financial power to make moves in M&A if necessary. We have a $2 billion net cash position,” answered Bozotti. “The priority today is organic growth but also to reinforce the lines of product we have,” said Bozotti referencing ST’s recent acquisition of NFC and RFID assets from Austria’s AMS AG as an example that more moves might follow.
“In the semiconductor world, manufacturing need no longer be a barrier to being in the business. Any small company with good ideas, good IPs, can compete using a silicon foundry,” said Bozotti emphasizing why ST is turning to areas where manufacturing can be proprietary and differentiating. “With MEMS, with time-of-flight sensors, with FDSOI and silicon-carbide we have differentiated technology. This we make at home. This differentiation and technology convergence means we can develop products for automotive, for industry and IoT. The R&D is there. We will be a leader. We are on course,” Bozotti concluded.
Related links and articles: