Over the same period of time the company has shifted its stance from that of an analog and mixed-signal IC company that does sensors to a company with a platform approach to sensor subsystems. That also implies a change from ASICs to application-specific standard products (ASSPs) and the offer of hardware and software and application and use case knowledge.
And the change has been made since Alexander Everke took over the CEO position on March 1, 2016. In a notable spurt of growth and investment the Austrian company is now rapidly building up a manufacturing base in Singapore with the help of the Economic Development Board there. It is already employing thousands of people in Singapore while manufacturing capacity is only just being installed and ramped up at two new sites.
This is in keeping with the dynamism and ambition of Everke. As he said to a recent press tour visit to AMS headquarters: “Either you go big or you go home.”
Part of the problem is that at €600 million in annual revenue (about $700 million) AMS is neither small nor big on a global scale and as a public company a potential takeover target itself. Everke came to the CEO’s position promising growth. In 2016 he said the company would achieve a 30 percent compound annual growth rate (CAGR) over the next three years. At the recent press tour Everke upped the rhetoric: “We will grow at more than 40 percent per year. We are comfortable we can deliver that.”
That is a big promise even in the context of a semiconductor marker that is expected to grow nearly 20 percent in 2017. In the second quarter of 2017 AMS achieved nearly 37 percent year-on-year revenue growth and the company is forecasting 50 percent annual growth in the second half of 2017.
As he was coming in to the company back in 2016, Everke helped it conduct a strategic review that concluded that there was not enough growth or differentiation in the IC sector. Everke and his colleagues decided the company needed to move its center of gravity up the value chain to hardware-software platforms based on key sensor technologies that it either had, or that it could acquire. Everke then dropped a legacy plan to open a mixed-signal wafer fab in New York (see AMS pulls out of $2 billion wafer fab project). That and subsequent investment in Singapore means that the AMS center of gravity has also moved East.
Everke emphasizes that AMS is not dropping IC manufacturing. But wherever IC manufacturing can be outsourced it should be, he said. Internal resources such as those used for ICs are used where AMS has differentiating technology.
Next: Where is AMS hiring?
Where are these employees located? Some are at the series of small technology-based sensor business acquisitions that AMS has been making. They include the environmental sensors business from NXP, specialty image sensor firm CMOSIS, Mazet for spectral sensing and Cambridge CMOS Sensors for gas sensing.
eeNews Europe has already listed these in an interview conducted with Everke at the Electronica exhibition in November 2016. We also listed the four sensing sectors AMS is pursuing: optical, imaging, environmental and audio. But it is notable that in all cases AMS is avoiding the commodity sensor business and is looking to operate where sensors are set to grow next (see CEO interview: Everke of AMS on four pillars of sensing). Second-guessing technology take-off is a risky business but getting it right is a key to success.
How AMS sees its sensor platforms. Source: AMS
The series of acquisitions was followed by the larger acquisition of the optical packaging company Heptagon Pte. Ltd. headquartered in Singapore (see AMS to pay up to $850 million for optical business). And in March 2017 AMS acquired the next piece of the jigsaw; Princeton Optronics Inc. (see AMS buys Princeton VCSEL provider).
And the fitting together of Hexagon and Princeton is almost a blue print of Everke’s vision for AMS as a sensor solutions company.
Hexagon had time-of-flight depth measuring courtesy of its 2014 acquisition of Mesa Imaging AG (Zurich, Switzerland) in 2014. Princeton Optronics was already a supplier to Hexagon and VCSELs (vertical cavity surface emitting lasers) are becoming of strategic interest for use in structured light, time-of-flight depth sensing applications both in smartphones and in lidar for autonomous automobiles. It should come as no surprise that at the time of closing Hexagon had a plan to spend $250 million expanding its production on the strength of a confirmed customer commitment to additional capacity.
No names are given but it is possible that this is an optical module for a significant smartphone manufacturer. Hexagon brought nearly 1,000 additional staff to AMS but even Hexagon and all the tuck in acquisitions do not count for the rise in AMS headcount.
Next: AMS still on acquisition trail
And then it was announced that AMS is moving into a 150mm wafer fab suitable for the high-volume manufacture of VCSELs and sensors in the Tampines region of Singapore. This is on top of the announcement in May 2017 of a new manufacturing facility in the Ang Mo Kio suburb of Singapore that was scheduled to open in July 2017.
So it becomes clear that once Everke got hold of the reins of the company he did a 180-degree turn changing focus from ICs to sensor subsystems and moving the next manufacturing location from New York to Singapore, but importantly he also did it quickly.
Alexander Everke, CEO of AMS.
“Yes, the Singapore build out replaces the fab in New York, which would have been CMOS, and which would not have been differentiating. Singapore is an optical play which is highly differentiating and differentiating at a lower price point,” Everke told eeNews Europe, on the side-lines of the press meeting.
“At the same time we have integrated six acquisitions and we have not stopped. We continue to look at which companies are available and which could strengthen our core business.”
Everke continued: “We want to own the value chain but we don’t have to own the whole thing.” He gives the example of a bio-sensor that AMS offers that includes an embedded microcontroller. AMS uses a Cortex-M4 processor core licensed from ARM. Everke said in the same manner when and where machine learning becomes important in sensor subsystems AMS could license in a hardware accelerator and either develop code in house or outsource that as well.
AMS is also involved in MEMS manufacture to a limited degree although its biggest success in this area is as a developer of the mixed-signal interface ASIC for MEMS microphones.
A fragmented application sector such as that served by MEMS sensors would seem at first glance to be suitable ground for AMS and its sensor subsystem philosophy. “There’s no plan for MEMS,” said Everke. “Knowing what not to do is as important as knowing what to do. We see so many opportunities to strengthen the portfolio. But that also means that we see scope for a continuous transformation of AMS over the next ten years.”
Related links and articles: