.jpg)
A panel discussion hosted by Efinix discussed how Malaysia's electronics and semiconductor manufacturing strengths and capabilities can be leveraged to accelerate the country’s deep-tech ecosystem development.
Celebrating its 50th anniversary, Malaysia’s electrical and electronics (E&E) sector continues to be a key growth driver. For August 2022, the E&E segment registered exports valued at RM50.89 billion ($10.78 billion at $1:RM4.72), up by 48.5% year-on-year and accounting for 36% of the country’s total exports, according to the Malaysia External Trade Development Corp. (Matrade).
Matrade expects E&E exports to continue to perform well this year, underpinned by the strong growth in the semiconductor devices and integrated circuits (ICs) sector, despite the ongoing global pandemic challenges and the continuing chip supply shortage.
The AmBank Group, meanwhile, forecast the E&E sector to expand at around 15% this year as it benefits from a still-strong global semiconductor sales, which is expected to grow by 13.6%.
Overall, Malaysia’s E&E sector remains strong and continues to pave the way for the country’s economic growth.
However, the country cannot rest on its laurels. Over the past decade, the “deep-tech” landscape has evolved significantly, with technologies such as big data, Internet of Things (IoT), artificial intelligence (AI), blockchain, and quantum computing, growing at breakneck speed and changing the way people live, work, and interact.
“The past 10 years have seen the rapid emergence of the digital era. And this fast pace of digital transformation continues to gain momentum across the globe. This is all the more apparent amid the challenges of the pandemic; we saw technology and digital tools becoming more essential than ever,” said YB. Datuk Haji Ahmad Amzad Hashim, Deputy Minister of Science, Technology, and Innovation (MOSTI), during his opening remarks at the recent panel discussion hosted by Efinix Inc., in line with its celebration of its 10th anniversary.
With the theme “Emerging Deep-Tech Ecosystem in Malaysia: Now, Future & What’s Next,” the panel featured industry experts, VCs, members of the academe, and representatives from the government, discussing how Malaysia’s electronics and semiconductor manufacturing strengths and capabilities can be leveraged to accelerate the country’s deep-tech ecosystem development.
“This panel’s theme is indeed relevant and timely. Without a doubt, the pandemic has greatly accelerated digital adoption. In the last two years, ASEAN has registered over 60 million new users, with internet penetration increasing to 75 percent,” said Hashim. “Today, ASEAN is reported to be the fastest growing internet market in the world. Southeast Asia’s digital economy alone is projected to reach $1 trillion in gross merchandise value by 2030.”
With its prime position in the heart of the ASEAN, Malaysia is well poised to seize these ample opportunities. According to Hashim, the Malaysian digital economy is expected to contribute 25.5% to its GDP by 2025. But to spur this and other emerging industries such as fintech, edtech, healthtech, and AI technologies, he said it is vital to foster a strong entrepreneurial spirit and a highly conducive ecosystem for startups to further drive innovations in these areas and propel the development of Malaysia’s digital economy.
And one prime example of this is Efinix. Based in Silicon Valley in the United States, Efinix set up its R&D hub in Penang, Malaysia, with investments from the Malaysia Venture Capital Management (MAVCAP) and the National Innovation Agency (Agensi Inovasi Malaysia or AIM) under the Economic Transformation Program (ETP).
The investments in Efinix, the only IC startup company in the ETP, was envisioned as a strategic investment to grow Malaysia’s high-tech IC design center, according to Hashim.
“I must commend MAVCAP and AIM for having the expertise and foresight to recognize the potential in Efinix,” said Hashim. The investment in Efinix also paved the way for the transfer of technology and job opportunities, thereby creating a highly skilled Malaysian R&D team.
Today, Efinix is at the forefront of the global semiconductor industry with its innovative and highly disruptive FPGAs. And the company is actively developing an ecosystem in Malaysia for advanced FPGA technology by collaborating with various partners.
“This includes cooperating with Universiti Teknologi Malaysia (UTM) to adopt this curriculum, as well as with MIMOS to develop AI-based machine vision solutions targeting Industry 4.0 technologies,” said Hashim. “Collaborations such as these will support the initiatives put in place by the government, and mostly to nurture a resilient and skilled Malaysian workforce to propel the deep-tech industry and R&D.”
Not there yet
One of the questions asked during the panel discussion was where Malaysia is in relation to the deep-tech developments happening worldwide.
“In a sense, we are not quite there yet as a country,” said Dr. Iskandar Samad, CEO of MIMOS Berhad, Malaysia’s national applied R&D under MOSTI. “If we use market capitalization as a proxy, and we dial back in time 10 years ago and compare ourselves to countries like South Korea, Taiwan, Israel, the United States, they all have grown leaps and bounds. As we know, market capitalization is a proxy for value. And most of their value is derived from technology companies. If we deconstruct the top 20 companies in the world, I think you will find that a good 40%, if not more, consist of technology and consumer-oriented technology companies. The Apples and the Googles of the world.”
But it is not all dreary, according to Dr. Samad, as Malaysia is well poised to leverage its composition in the E&E sector.
“We have a vast and very capable assembly, test and packaging (ATP), and electronics manufacturing; and we have several nascent IC design companies—I will qualify Efinix partly as an IC design company as well,” Dr. Samad said. “And there is an inflection point happening in the industry, with electrification, connectivity, sensory, and cognition right at the edge. That is happening. If you look at solar generation, how much that will supplant power generation plants, combined-cycle gas turbines. Vehicles will be electrified, and then there will be an entire world of autonomy and sensory that will be applied. So, I think Malaysia is well poised.”
Datuk Seri Lee Kah Choon, Special Investment Advisor to the Chief Minister of Penang, echoed the same sentiment. Building up its manufacturing excellence over the past 50 years, Malaysia has positioned itself as one of the major E&E manufacturers in the world, he said.
But what the country needs is to leverage its existing strengths to move up the value chain.
“I think the next stage for us to move forward is going to be how we are going to develop our deep-tech excellence to not just do manufacturing but move up the value chain; to see how we can use our brainpower with our understanding of the technology that we have nurtured so far,” said Lee.
Dr. Samad explained, “We need to orchestrate this future very well. MOSTI has undertaken the creation of five roadmaps at the beginning of this year, along the lines of E&E, AI, blockchain, robotics, and advanced materials. If we orchestrate that future very well, and MIMOS has a role to play in that—and we are committed to the cause—we have to align the dots between a clear and credible market opportunity, and a well-poised starting point in terms of where the country is. [We have to] orchestrate plans and execute well against those plans to position Malaysia for the future.”
According to Datuk Ooi Boon Chye, former SVP of Broadcom, no company is big enough to take care of everything, therefore the industry still need help from the government.
“I was also saying that we probably need a wafer bumping fab in Malaysia. We used to do wire bonding, but wire bonding is out. The technology is out. The technology is such that you cannot do wire bonding anymore. You cannot bond those wires to be that small. You cannot make wires that small. Whether it is gold wire or aluminum wire, you cannot make them that small,” Ooi explained. “So, today, it is all bumping technology. I would really encourage help to provide that kind of support.”
Over the next 10 to 20 years, Ooi expects manufacturers will no longer be building the biggest chips. “They are going to chiplets and 2.5D stacking and 3D,” he said. “That technology is where I think Malaysia can really excel in and continue to be a technology leader.”
Addressing the talent conundrum
While Penang is fortunate in that many good companies have come over to set up shop, the talent in the region is still mostly focused on the manufacturing side.
“Efinix is probably one of the first full R&D capable organizations in Malaysia,” Ooi said. “It is such a great startup, and I am really heartened as a Penang citizen who has been in the industry, and somebody who’s been involved with Malaysia all this time.”
Assoc. Prof. Ir. Dr. Muhammad Nadzir from the School of EE at UTM said Malaysia does produce talent, “but we need more.”
“Technologies are defined by innovations. In computing systems, we have more deep learning algorithms that come in with the availability of data; and we have massive connectivity of systems, devices, businesses, humans. All these require a different view at what computing is all about,” he explained. “From a traditionalist approach, computing involves hardware and software. From our perspective as academicians, trying to instill this to students, to our graduates, is not straightforward because we have seen the emergence of different disciplines. We have seen over the past, we have computer engineering—which then merges with EE and CS. At the same time, we have to look not only from the hardware and software perspectives, but from algorithms and architectures.”
Therefore, from an EE perspective, Malaysia needs a concerted effort for it to further its technology development.
“We have so many roadmaps and policies, but I am not sure whether it is enough at this point,” said Dr. Nadzir. “We have churned out talents; but the number of talents that we produced—through our local universities, be it public or private—is not sufficient to satisfy the needs of our industries.”
He agrees that Malaysia’s E&E talents are mostly focused on manufacturing. “Our industries started off with a very strong backend—manufacturing, assembly, and test,” said Dr. Nadzir. “Over the years, we have seen more frontend capabilities in our country. More and more R&D centers are being established. From the perspective of talents, we produce them, but we need more. And there is still a lot more that is needed to be done for universities to fulfill this requirement.”
MAVCAP CEO Shahril Anas Hasan Aziz said that right from the start, when Efinix was starting, one of the key challenges was getting the right talent. “It is tough to get people from established companies to join startups. Part of it could be the package, but part of it is looking to the future.”
Efinix CEO Sammy Cheung, who was moderating the panel, shares the same sentiment. “As Efinix continues to grow, we want to double, triple, even quadruple our resources on different fronts. But being able to find and build up more talents is going to be a daunting task in Malaysia,” he said.
“For sure, we are developing talents within the company to continue to grow their capabilities,” said Tony Ngai, CTO at Efinix. “But definitely, we also need to get it from the industry, from the universities. So, it is important to talk about what we are looking for, and what skillsets we want.”
On top of that, there is also a depth versus breadth discussion, according to Ngai. It is no longer necessary for people in this area to have the same expertise as there are a lot of technology developments happening all at once.
“I would encourage everybody to focus a bit more on being curious to understand more, to understand the wider scope of technology—what the dependencies are, the surrounding groups, the surrounding industries,” explained Ngai. “Rather than just hardware, how about the software guys? How about the software guys either sitting in Hong Kong or Sydney or other locations—what are they good at? And then how to leverage them, get the technology, and skillsets you have in this area. Building the breadth will help us be able to foster much wider innovations and industrial applications.”
Ooi said there are fewer and fewer requirements for people who are very technically capable to be graduates. “The reason is that there are a lot of things that can be learned on the internet by the young people themselves. But what these young people require is guidance from senior people inside the company,” he said. “I was very impressed that Efinix has more than 20 interns. That is a substantial number of interns that you are actually committing to.”
According to Ooi, practical engineering skills and guidance will save many years for Malaysia.
“Cross pollination” approach to nurturing talent
According to Cheung, Efinix’s development team in Malaysia can do anything from transistor level design, RTL, analog design, full-chip mixed signal integration, to tape-out.
“I am definitely proud of our world-class development team,” he said. “Majority of the development is done here. On top of just hardware, we also do full stacks: self-developed software, all the way from frontend GUI, place and route, the whole FPGA software program. We also do high-level system-level defined from soft IPs, RISC-V stack, to provide all the flexible system level solutions to our customers, we also have AI teams working on the inference.” On top of all those, Efinix also does production test development.
“I think having a full ecosystem in a region—if it is Penang already, that’s great—if it is experimenting from Hong Kong, from other regions, it will be good,” said Cheung. “We are doing lots of ‘cross pollinations’ in terms of technology. How to bring engineering in Penang, the software team, the application team; we have people from other regions to bring up the custom interfacing actions. All those are building up the engineering capability in this region, from Efinix’s perspective, and will benefit the wider industry.”
And that is what sets Efinix apart from other MNCs: cross pollination.
“It is functional,” said Cheung. “Personally, I would like them to understand the full business, why they are doing this and that. We really like to get people—they will change functions within the company because we are a growing worldwide business. It is not all day we are sending to this plant here, but also, more importantly, how the employees have a chance to go to other places worldwide, to learn about the business, and how to support it. Those could be ways that we try to attract more people coming into the field. As long as they join us with good fundamentals in the technology side, they all have a chance to grow into the next phase, which usually takes a few years—but we are willing to invest and hire as many people as possible to grow at various stages.”
Experience is indeed hard to get, but Cheung believes that when other companies hear about Efinix’s success in Malaysia, more people will come. “It is better to get more people to invest here, so that we’ll get more talents. We play the long game—we want to be here in the next five to 10 years and have as many talents as possible.”
From an industry perspective, Dr. Lee said what Efinix is doing is a very important aspect of nurturing talent. “We have always been complaining about talent and its quality. But of course, when we look at the industry, it is very heartening to see companies like Efinix, whereby you take up this mentorship program to nurture talents,” said Dr. Lee. “We do have many talents. You just need a mentor to really transfer some of the technology and the knowledge to these youngsters. This mentorship program is something that is very commendable, and I think that more companies should pick it up as well.”
Replicating E&E success to boost deep tech sectors
Dr. Samad of MIMOS said Malaysia cannot always be dependent on commodities or agriculture. “They are finite resource. When we study other countries, for example, Made in China 2025, everybody knows about that; and there is also the Fabs and Chips Act that the US has come up with to strengthen upstream fabrication activities,” he said. “Historically, some of you might know about the history of Japan’s VLSI Project back in the 1970s; the world of ITRI, TSMC, and Morris Chang, and the repatriation of Taiwanese talent back from the US; and the role of chaebols in building up South Korea’s high-tech industries. The reason I mentioned these is because of concerted effort, holistic effort—what this country needs to state as an aspiration in the semiconductor space to build real critical mass.”
He said if the critical mass is there, a lot of these enabling factors will fall into place. “At the end of the day, it is the matching of supply and demand. That reaches an optimum point; you do need strong demand drivers that will then help drive the supply—that is a virtual cycle,” Dr. Samad noted.
One of the things that MIMOS is doing on behalf of MOSTI is to develop a stronger aspiration for where the country needs to be in semiconductors and E&E. “But we cannot do this alone. To the industry, please come and talk to us, and provide us opinions on where you think this country should head to,” Dr. Samad said. “There are inflection points in the industry that, if we played our cards right, we could use as demand drivers to build in the shared success for local companies to partake. Regulatory policy drivers that favor local content and nurturing of local industry, is part of the equation.”
So how can the success of E&E be replicated? Dr. Samad said E&E is a necessary part of all those industries.
“We should not see AI, blockchain, IoT, as separate from the E&E sector. The E&E sector drives and enables those use cases and new technologies,” he said.
Dr. Lee agrees. “The foundation is the semiconductor industry. If you look at all these use cases—be it AI, solar, IoT, and deep-tech—all these have to build on the foundation of semiconductors,” he said. “From this foundation, we have built many diverse types of industries. The LED industry is one of it. And then, of course, we have the solar power, and others. But for us to move beyond that, Malaysia has to move away from pure manufacturing.”
He said the industry should use the “brains” more, rather than just the hands and legs to do manufacturing.
“It is very heartening to see that we have started to move the backend or frontend to the side. For example, the IC design side is very promising because it is something that we have the understanding, and the talent that understands the semiconductor industry,” Dr. Lee said.
Indeed, Malaysia is a semiconductor hub, contributing around 5% to 6% of the semiconductor production worldwide, according to Dr. Lee. Which is why he believes Malaysia will have a certain critical mass.
“But at the same time, I totally agree that we have to be very focused, we cannot do everything—we just have to be focused,” he pointed out. “If we can expand the IC design industry, I would say that is where Malaysia should move towards.”
UTM’s Dr. Nadzir said more inputs are needed from the industry—from setting up labs to sponsoring students and providing them experiential learning, beyond just doing internship.
“We want students to be able to work in the real industry, so that their experiential learning can actually improve them further, as well as to support universities,” Dr. Nadzir explained. “From the university perspective, sometimes we feel we are like a one-shop center. ‘Dr. Nadzir, we want to hire 50 engineers—can you provide us with 50 engineers?’ I would say it is rather impossible because our intake for electronics engineering, for example, is only around 120 to 140 per year.”
And then there is the challenge of talent exodus.
“Because of that, we encourage companies to invest in talent development, to be engaged with us,” Dr. Nadzir said.
He explained that it is not easy to provide a wholistic perspective because students, being engineers, need to solve problems. And that problem-solving skill requires access to tools.
“Sometimes, companies advise us to teach our students to use industry tools—which is actually very hard for us to get. So, I said, why don’t we just provide them with just open-source tools just so they can learn the trade, learn the design flows, before they upgrade themselves,” Dr. Nadzir said. He added that in as much as the academe needs real industry use-cases to be linked back to the students, the industry should also invest in academia by providing the latest technologies, breakthroughs, and know-how, which can be shared with the students better.
“Training the trainer is one of the approaches that the industry should take,” Dr. Nadzir said.
Cheung said it these are the right things to ask the industry. “At Efinix, we absolutely like to see more activities start early at the university level. So, we are waiting to see what we can do more to be able to bring more talents to the industry.”
“My hope is that you call Malaysia home,” said Dr. Samad. “There are plenty of local opportunities here, for sure. MIMOS, for one, is willing to assist and aid in matching you up. I mean, there is a lot of machine vision players here in Penang itself. So, in developing those kinds of local use cases, MIMOS can help facilitate that. Never leave the country; don’t pull out your design center in Malaysia.”
Being an IC design company, Efinix tends to be more conservative, according to Cheung, “because we have to be very careful of how things are being done.”
“As a company, we can hire hundreds of people, even 2,000 people—but that’s only 2,000. What we want is more people outside the ecosystem being able to leverage what we built. We also want to make connections globally. I believe, where we are, this is just the best place for Efinix, here in Malaysia. We can definitely leverage the environment, and we believe we have the right setup and platform for the next 10 years to do that,” Cheung concluded.
Source: www.eetasia.com