When AEM reported earnings for its 1QFY2021 ended March 31, market observers couldn’t help but wonder if the company’s growth momentum has faltered. However, according to CEO Chandran Nair, the blip is akin to changing tyres on a moving vehicle.

“What you see right now is about making space for our next-generation platforms by keeping the current platform going. You can imagine it as kind of like changing the wheels of a moving car without stopping or slowing down as the change happens and then ramping up again,” says Nair in an interview with The Edge Singapore.

In FY2020, AEM had a stellar year, reporting an 85% increase in earnings to $97.6 million compared to FY2019. Revenue in the same period rose 60.6% to $519 million. In contrast, revenue for 1QFY2021 was down by nearly half to just $80.2 million from $146.8 million a year ago. Earnings dropped 63% y-o-y to $13.3 million. However, when compared to 1QFY2019, AEM’s 1QFY2021 earnings and revenue were still 50.5% and 34% higher respectively. AEM also highlighted that the 1QFY2021 revenue figure was the second-highest in its history of first-quarter results.

As at June 15, AEM is trading at $3.65 or 10.2 times earnings and 4.6 times book value, and gives the company a market cap of $1.03 billion. Year to date, the stock is up 2.46%.

AEM points out that FY2020 was an “exceptional” year because its biggest customer was ramping up demand on additional equipment and consumables throughout the year. The customer, whom the company did not name, is understood to be Intel Corp. Furthermore, AEM is riding the long-term digitalisation trend which should increase demand for services provided by the company.

Based on its orders on hand and business outlook, AEM is guiding for FY2021 revenue to be between $460 million and $520 million. While 1HFY2021 results are expected to be relatively muted, the second half of the year is likely to see a “strong recovery” lasting into 2022 as next-generation tools are phased into Intel’s high volume manufacturing sites globally. “A good way to think about this is when you build the next-generation platform, it’s the trade between the mix of the current platform and the new platform,” says Nair.

AEM says that the new tools it would be delivered soon and throughout 2022 are “highly differentiated tools for the high-performance computing segment”.

AEM does not intend to put everything into one Intel basket too. “In addition to our largest customer, we’re also engaged in deep technical engagement with 10 out of the top 20 semiconductor manufacturing companies so we expect to see meaningful revenues in 2022,” says Nair.

Changing face of semiconductors

AEM has been actively expanding its capabilities via acquisitions. In April, it announced the US$3.8 million ($6 million) acquisition of a 26.59% stake in Ateco Inc, a South Korean company that specialises in the design and development of memory test handler solutions for the memory chip market. In December 2019, AEM acquired France’s Mu-Test for EUR7.53 million ($11.4 million).

The most recent and most significant acquisition was the $99.7 million takeover of listed peer CEI, which focuses on making printed circuit boards. According to Nair, integration with CEI has started, with AEM driving cost-efficiency improvements and leveraging customer relationships to increase its revenue opportunities. “We are looking to a very strong sustainable business for the coming years,” says Nair.

Just to be sure, despite reporting growth rates many other companies would be envious of, AEM isn’t entirely immune to the business fallout from the pandemic, which has forced the company to adjust the way it operates.

For example, instead of having their overseas staff flying in for training, the familiarisation was done using augmented reality. “We have done the entire training process using things like the HoloLens and we will continue to use these training and remote deployment technologies, even after the pandemic is over,” says Nair.

Big Data trend favourable to AEM

AEM is set to benefit from an explosion in Big Data generated increasingly by machines, which is expected to eclipse the Big Data generated by humans through their social network postings, Instagramming and video-making. Citing industry studies, Nair notes that over the next four years, humans will account for only 1% of the total data generated around the world.

The Big Data generated by machines will also have to be processed and analysed. This will fuel the growth of artificial intelligence and machine learning, which, in turn, increases the need for more powerful computers with even more complex semiconductors and components.

And there will be a growing need for testing and handling services, says Nair, similar to those provided by AEM, to ensure that the components have a high level of reliability.

Among them are what he calls “mission-critical” components, like the sensors in an aircraft, where if an application program fails, lives can be at risk. “These components are going into the automotive industry, going into the energy industry. We can’t say, oops, this failed,” he adds.

“It is going from traditional semiconductor 1.0 to meeting the requirements of semiconductor 2.0.” He elaborates by saying this is where the devices manufactured will be far more complex and involve so-called “heterogeneous packages”.

The term refers to the integration of separately manufactured components into a higher-level assembly (also known as a system-in-package) that provides enhanced functionality and improved operating characteristics.

“What happens now is the effectiveness of conventional ways of testing using conventional automated test equipment breaks down ... The heterogeneous packages result in devices that are much more difficult to test than the traditional means. And that’s why approaches like system-level tests are taking so much precedence now.” he says.

Multiple sites

In addition to these major changes in technology, Nair believes that the current geopolitical reality, along with the pandemic, has “driven a rethink of the entire global supply chain for technology”.

According to a study by advisery firm Forrester and reported by CNBC in May, the global chip shortage could last until 2023. But Nair believes that AEM is in the “right geography and right time” to benefit from these changes, pointing out that the semiconductor industry is a “global ecosystem”.

A 2020 study by Accenture on the semiconductor ecosystem showed the semiconductor value chain involves an average of 25 countries in its direct supply and a product can cross international borders as many as 70 times before it finally reaches the end customer.

This is likely to benefit AEM with its diversified supply sites — or what Nair calls “distributed architecture” in manufacturing. Including Singapore, AEM has sites in places like the US, China and even Finland. “We have made sure that our supply chain is designed around each plant in case we have countrywide shutdowns,” he says.

In an earlier interview with The Edge Singapore, AEM’s chairman Loke Wai San made a similar point, saying the new requirement in the manufacturing landscape is “availability”, which means that regardless of where the customer is, suppliers must be able to meet their demands, 24/7. “You’ve got to compete with others onsite. When the customer wakes up at nine, you’re there to take his call. Singapore companies that just compete on quality and costs will have no value proposition because they fail to address availability and technology,” says Loke.

With all these developments, Nair says AEM is in a good position to serve global markets and says the company will continue to invest in automation to increase throughput as this will “enable semiconductor manufacturers to meet the quick ramping up of demand and changes in high-volume manufacturing over the next few years”.