AEM CEO Chandran Nair Discusses Trends and Opportunities for Semiconductor Test Industry

Article By : Stephen Las Marias

Chandran Nair, CEO of AEM, speaks about his new role, the challenges and opportunities in the semiconductor test market, and the role of data analytics and machine learning in reducing the cost of test.

AEM Holdings Ltd is a global semiconductor and electronics test and handling company headquartered in Singapore. Publicly listed on the Singapore Exchange, the company provides the most comprehensive semiconductor and electronics test and handling solutions for markets such as advanced computing, 5G, and automotive electronics, to name a few.

AEM is a pioneer in System Level Test and has over 1,000 systems running on production floor and more than 35,000 test cells deployed worldwide.

Chandran Nair, CEO of AEM, speaks with EE Times Asia about his new role, the challenges and opportunities he’s seeing in the semiconductor and test market, and the role of data analytics and machine learning in reducing the cost of test while at the same time increasing coverage.

EE Times Asia: One year on, what can you say about your role at AEM, and what strategies have you implemented to help the company navigate its way through the COVID-19 pandemic?

Chandran Nair: I joined AEM in January 2020. In recent years, AEM has been tremendously successful under the stewardship of the chairman of the board, Loke Wai San, and the former CEO Chok Yean Hung. As Chok decided to retire, the transition plan between him and me ensured continuity, with Chok moving on to become a board member, and Loke Wai San continuing as chairman. So, we continue to have strong strategic leadership, with the previous leadership team being on the board. I have been very, very fortunate to inherit a strong executive team to take the company forward.

The pandemic situation last year brought the global teams together quickly. Of prime importance was that we ensure the health and safety of our team members first. Along with our suppliers, we met our customers’ increased demand, ramped up production, and deployed our tools successfully. We ensured site resiliency to overcome the various lockdowns worldwide by working closely with our customers and suppliers in true partnership mode to overcome things like shortages, border restrictions between countries, and so on. We conducted training and deployment of our tools by quickly adopting technologies like augmented reality [AR] to help train our field engineers in various countries to deploy new tools at our customers’ factories.

Since we could not travel, we used tools like the HoloLens and other AR/VR technologies. In many ways, we will continue to use the training and remote deployment methods we’ve developed over the pandemic period moving forward, even when the world comes back to normal because it’s extremely efficient. We’ve reduced travel, carbon footprint, increased our effectiveness, and decreased our time to respond to customers’ issues.

We’ve learned to use many of these technologies effectively over the past year. We are also using AR/VR to provide a look-and-feel for our remote team members to continue working effectively on the deployment side.

EE Times Asia: What opportunities are you seeing for AEM in this environment?

Nair: The increasing complexity of devices driven by the user applications of the 5G economy, which includes things like the data-centric approach, machine-to-machine communications, autonomous machines, and robotics—all these are increasing the challenges in test and making current approaches like BIST [built-in self-test] and DFT {design for test] less effective in meeting the needs of the industry.

In the past ten years, the focus was primarily on technology improvements in the front end. However, what we see now is a tremendous investment in the backend, involving packaging and test. Companies are moving to heterogenous packaging to assemble multiple IC blocks into a single chip package for application-specific needs. So, the various blocks, even on different technology nodes – for example, you can have a logic block on 7nm, or the power management block on an older technology node, for example 14nm. The heterogeneous packages result in devices being much more difficult to test using the traditional means.

In monolithic chips, you can scan through and test a high percentage of transistors. But as technology nodes advance, the number of transistors being missed using traditional ATEs is growing. The problem is compounded—not only the number of untested or decreased coverage using traditional ATEs, but it is also further complicated by the inter-die connect between the different IP nodes.

This means that the only way to get the maximum coverage and ensure the product’s reliability and safety is by putting your SOC or SIP into a System Level Test mode. Simply put, it means the system-level approach is where the engineers have started testing the chips in mission mode. For example, they would take a CPU or a GPU, put it in a reference mother board, download the appropriate software, and excite the device as if it were in a real use-case, to ensure that it works in the way it’s supposed to.

More and more companies have been building AI chips, complex chips, multimode IP chips—all in a single package. They realize that the only way to ensure or assure safety, reliance, and test coverage is to use System Level Test. And this is a great opportunity for AEM as the world moves more and more into SLT for higher-end devices because this is where we have a tremendous lead in deployment experience and technology.

EE Times Asia: What is the role of data analytics and machine learning to reduce the cost of test, while at the same time increase coverage and improve time to market?

Nair: We have been looking at big data for a long time. Data analytics and machine learning are an integral part of our future test and handling methodologies, spanning its use for both intelligent predictive maintenance on the tools and to be able to better slice and dice data to understand trends, patterns and risks, so that the customers can make decisions faster and in real-time.

At AEM, we are working with our customers to continuously evaluate the best use of data analytics and machine learning to make sure that our customers can get insights into test processes to reduce their cost of test and increase their yield.

This is a work-in-progress because this requires continuous engagement both from the customers and our R&D team to ensure that we are using the data appropriately to be useful to our customers and us.

AEM CEO Chandran Nair

EE Times Asia: How is 5G empowering IoT and AI, and what’s the role of AEM and testing solutions in this ecosystem?

Nair: 5G is such a buzzword. The applications in 5G essentially bring together high-performance computing at the edge; low-power sensors, and high-speed connectivity to enable powerful and transformative use cases spanning from agriculture, smart power grids to autonomous transport.

As the applications get more complex, so do the semiconductor devices that power these applications. This means that the need for AEM System Level Test, and what we call the Test Cell approach, providing the ability to do high-volume testing for complex devices, increases. Thereby, they will increase the opportunities for AEM to do more.

EE Times Asia: What end markets do you think will drive the semiconductor and test industries over the next year?

Nair: When you look at the industry and the proliferation of semiconductors, you can see where some of the customers’ needs are going—the growth in the automotive, especially for clean energy and electrification of that space, is one enabler that will drive the semiconductor industry. Mobility devices will continue to grow.

In addition to those two obvious ones, more and more companies are coming out with application-specific AI-powered chips, and that will also spur the growth of the industry in addition to what we already see, which is the data-centric approach.

The difficulties this global pandemic has thrown at us is tremendous, but it has also enabled the faster adoption of technology that can help us enhance remote working capability and remote learning ability.

In the new normal, the way we live, learn and play are different. There are opportunities for us to look at how we can best use technology while minimizing the impact on the environment and moving forward with our ability to become more productive.

Leave a comment