The Bloomberg Innovation Index is a worthwhile read for those who want to understand how countries fare in a multi-faceted score of parameters that, when put together, describe their innovative capability. The index has seven components, namely R&D intensity, manufacturing value-add, patent activity, productivity, high-tech density, research concentration and tertiary efficiency.
It is therefore not a surprise to us that the top three nations currently on that score are Germany, South Korea and Singapore in that order.
It is not the intention of the writer to expand on the details and scoring mechanisms of the index but what is to be noted here is how Malaysia has fared when compared with the rest of the world. For 2020, Malaysia has been ranked 27th overall. A middling score — nothing to shout about, but not too bad either, considering that a total of 60 countries were ranked.
However, if one looks deeper at how our ranking was achieved, we will begin to see some nuances behind our performance too. Across the seven components, Malaysia’s best score was ninth for manufacturing value-add, and the worst score was 46th for productivity.
Let us reflect on that for a moment. Our innovation index tells us that we are relatively good at adding value to items that are being manufactured, moving it further towards its end state, but at the same time, we do it rather inefficiently and labour-intensively too.
Another observation — and this is where I would like to make my point — is that the majority of countries that have been ranked high above us are countries that have significant activities in integrated circuit (IC) design, semiconductor fabrication, systems design and development, software development and engineering, systems integration and other science- and technology-based activities. These activities are the core competencies of their industrial ecosystems.
The structure of the industry and value-adding
The electrical and electronics (E&E) industry has been one of the pillars of Malaysia’s economy for nearly five decades and is a very significant export contributor, totalling more than RM330-plus billion per year over the past three years. This is more than the combined exports of our oil and gas and palm oil.
However, the challenge for the local E&E industry is that our imports are significant too. If one takes the difference between the exports and imports, and expresses it as a percentage of exports, then the picture emerges. This “local value-adding” has been effectively stagnating between 25% and 35%, and has gone on for decades too. Why is this so?
A look at the industry structure gives the answer. The E&E industry is nominally structured as front end and back end. At the front end of the industry, the ICs or chips are first conceptualised, designed, developed, prototyped and eventually brought to mass fabrication in very expensive fabrication facilities, or fabs. The final outcomes of the front-end stage of the industry are integrated circuits that have been fabricated onto wafers of silicon (or gallium nitride, silicon carbide, gallium arsenide and so on).
These ICs on the wafers are called “dies” and are then tested and brought to the next stage of production. This is when the back end begins.
At the back end of the industry, the wafers are diced into individual dies and are then readied to be connected to the outside world using various connection technologies. This step is called packaging. Following that, the dies are encapsulated in a suitable material for protection from the elements, eventually taking the familiar form that we refer to as “chips”.
Different chips are then assembled onto a circuit board or some other substrate, where they are eventually connected to other electronic and electrical components to become modules or systems. These are then put together to finally become the goods in the markets that we see.
Chips have become one of life’s constants, just like death and taxes.
There are clear distinctions between the front-end and the back-end industries. At the front end, the requirements are very stringent and exacting, putting a premium on understanding the fundamental sciences and technologies. The deep tech of manipulating atoms and molecules requires extreme control of the environment and process conditions. It follows that value-adding is very high at this stage of the industry.
Fabs are multi-billion ringgit facilities, but once built can last for 50 years or more very easily. However, due to the start-up costs involved and the human resources and ecosystems needed to maintain them, there are not many countries in the world that have bothered to invest in the efforts needed. This is a tough endeavour indeed.
A strong back with a weak front
Malaysia has done very well in establishing the back-end industry. Because of the enabling environment created by the MIDA Act 1965, the growth of the E&E industry back end is synonymous with foreign direct investment (FDI). Multinationals, when suitably incentivised, had set up camp in our so-called Free Trade Zones and created huge employment opportunities for our people. But the fundamental characteristic remains: value-adding is relatively low at the back end.
The supporting ecosystems also grew in tandem and have resulted in the creation of our own world-class companies in the respective supply chains. These companies, usually specialists in automation and manufacturing systems, have also spread their wings, becoming preferred vendors of their parent MNC and replicating their global footprints too. Some of these local vendors have also become the darlings of Bursa Malaysia.
In the late 1990s and early 2000s, Malaysia took steps to enter the extremely challenging and competitive area of IC fabrication as well as expand into IC design. South Korea, China, Taiwan and Singapore also followed this pathway, and to date have shown that their front-end industries are thriving.
What about the situation in Malaysia? Despite the unwavering visionary statements by our administration, the current landscape in the country can only be described as a work in progress. We cannot show much in terms of physical assets on the ground and other tangible developments in the industry ecosystem. There seems to be a mismatch between what was envisioned and what has been achieved thus far.
Take the case of Silterra. It was first conceptualised in 1995, and eventually launched in 2001 after six years of planning, fundraising, construction, commissioning and technology acquisition. The campus in Kulim was to house three cleanrooms and was supposed to confidently drive Malaysia forward in this global business.
It was meant to push our exports up towards the higher end of the value chain.
But the actual situation now is quite the contrary. After more than 20 years, the fab is still sub-scale in its output and is a miniscule player in this competitive global landscape. It is hard to recover the high costs of semiconductor process development and engineering when you cannot achieve adequate volume production. Did we not fully appreciate the challenges when we embarked on this path?
Is a reassessment warranted? To answer the question, we first need to understand where we are at today. Is the presence of a locally owned semiconductor fabrication facility a critical resource needed by the country? Why would we need a local fab when there are similar facilities within a few hours of travel in the region? Why indeed.
The effort of bringing a design of an IC into mass production is an iterative one. This is a process that consumes several tens to hundreds of millions of ringgit, with the efforts stretching months and even years. Time, money, expertise and good support are resources that are critical. Tangible and non-tangible factors play their roles in ensuring success.
Taking the design of an IC into the physical world depends greatly on the abilities and willing support of the fabrication facility. Even if a fab is capable of fabricating a design, it may or may not choose to proceed to fabrication. The decision is purely business-driven.
When engaging a potential customer, the fab shall work on developing and engineering the semiconductor processes that are needed to bring the design to life. Fabs constantly invest in process design kits (PDKs) as supporting inputs to chip designers. By utilising the PDKs of a particular fab, the designer is assured that their designs shall have a high chance of success and be manufacturable.
The healthy and virtuous relationships between the IC design market and the fab become an enabler for innovation in an industrial ecosystem. In the past, this was what Silicon Valley was all about and what Hsinchu (Taiwan) is about today.
It is clear that having an accessible and supportive fab is critical. Creating the prototypes, testing and developing the concepts are the early stages of innovation and require as much support as possible. Only a locally controlled fab can extend that quality support and create the synergies beyond the hard-nosed, commercial realities. Strategically aligned fabs are resources that are much valued in this ever-growing industry.
Divestment talk leading to the disinvestment effect
It is very unfortunate for Malaysia’s stillborn front-end semiconductor industry that the constant talk of divestment has dominated the airwaves for a very long time, creating a pall of disinvestment over our industry.
Silterra has become synonymous with divestment, not growth or possibilities. Whether justified or not, our local investors, bankers, media, business people, educators and the general public have all grown wary of this industry.
In spite of the explosive growth of electronics and the hyping up of the Internet of Things, IR4.0, AI, machine learning, green technologies and so on, no one here has stepped forward to actually invest in the creation of the chips, local systems and solutions that are needed. We seem to be content as importers of technologies only.
Is divesting the only option? Can we restructure and reorganise?
We do not need rocket science to tell us that investments of this nature are not where financial profits shall be immediately plentiful, especially when we are at the stage of trying to establish the semiconductor industry. With its long gestation, economic profits, however, have to be brought into the equation too. This globally competitive industry is not for the faint-hearted.
We should also ask whether for the long term and with inroads already made, is divestment the right option for Malaysia?
Our current investment has never been subjected to the full range of restructuring and reorganising possibilities. It may be time for the enlightened corporate wizards to do something that should result in Malaysia not selling out its technology resources to foreigners or Trojan horses.
It is a well-known fact that Taiwan has done a fantastic job of making its semiconductor industry a global success. While marvelling at its stellar performance, we should also appreciate that its success was due to its unwavering focus, stringent corporate governance and long-term perspective.
The same story is currently playing out in China. Nearly all of the country’s fabrication facilities are funded by government resources, in full cognisance of the long-term nature of these investments. Companies that run the fabs are unburdened by their capex debts and are then empowered to also grow their design industry.
They turn their abstract vision of technological independence into real business actions.
Malaysia too has had its consecutive industrial master plans spelling out visionary statements such as “move up the value-chain”, “knowledge-based industry” and “get out of the middle-income trap”. It now looks like we are at a watershed moment.
Our place in the global technology landscape is unique. Our costs are competitive and we already have an established back-end industry. We are seen as respectful of intellectual property laws and are considered neutral by the competing forces of the world today.
The opportunities are large enough for us to chart our own unique offerings in this continuously growing market and we need not compete head-on with the established players. If we lose touch with this industry and its foundational technologies, getting back in is almost impossible. In fact, we could and should pursue the development of the industry with improved vigour and focus.
Kamarulzaman Mohamed Zin is the immediate past CEO of Silterra Malaysia Sdn Bhd and was founding chairperson of the Semiconductor Fabrication Association of Malaysia