Australia’s place in the semiconductor world: Some humble thoughts on developing a native semi industry

Today in or ongoing editorial series, Australia’s place in the semiconductor world, Dr Venkata Gutta considers why we need a domestic semiconductor industry, what other countries have done to develop theirs, what we can learn, and some proposals.

Millibeam is a mmWave fabless semiconductor company based in Sydney, Australia and we develop high-performance integrated circuits and systems that enable energy-efficient and cost-effective radio networks for mmWave 5G. 

For a company such as ours, a semiconductor ecosystem that includes access to capital, talent, foundries and links to domestic and export markets cannot be understated. 

Access to some of the above-mentioned resources are available in Australia. However, a fully functional semiconductor ecosystem is needed to build and grow a local industry that has scale. 

This idea has been gaining some traction in a landscape that is being continually shaped by geopolitics, supply chain shocks and the economic and strategic importance of technologies such as 5G, satcom, artificial intelligence (AI) and quantum computing. 

Building a thriving Australian semiconductor industry is within the bounds of reality. The factors working against it are complacency and the legacy of successful globalisation that have seemingly made some countries semiconductor superpowers and others dependent vassals. 

While semiconductor supply chains will continue to stay global, Australia need not remain highly reliant on supply chains that have been severely exposed in recent times. 

We can look to semiconductor superpowers and middle powers for lessons on how they managed to build domestic semiconductor industries that have gone global.

Semiconductors drive the global economy

Semiconductors are ubiquitous and are central to technologies such as the internet, medical diagnostics, wireless communications, and automobiles. 

It would be no exaggeration to describe semiconductors as engines of the global economy and catalysts for productivity growth. 

Notwithstanding their outsized influence, they occupy little to no attention in the minds of the public and policy makers. 

If not for the much-written about supply chain shocks and the resulting shortages during COVID, semiconductors would have at best occupied a perfunctory role in public imagination. Supply chain shocks coupled with the recent ‘chip wars’ have lent a geopolitical significance to semiconductor technologies and their supply chains. 

Semiconductors: Strategic bargaining chips in the geopolitical great game?

Semiconductor supply chains represent the best of global collaboration. 

Companies from different nations come together, pool resources, and advance the state of the art. 

No one country enjoys a technological monopoly in all aspects of the semiconductor supply chain, and co-operation is the name of the game. 

For instance, TSMC of Taiwan, which is dominant in chip fabrication, relies on highly accurate semiconductor fabrication machines from ASML of Netherlands or LAM Research of the USA. At the same time, industry  giants such as US-based Broadcom and Qualcomm or Dutch-based NXP might rely on TSMC to fabricate their chip designs. 

Finally, the assembly of chips itself is carried out in countries such as Korea, Taiwan, the Philippines, or Malaysia. 

All countries must work together and rely on each other to ensure a well-functioning global supply chain. 

However, against the run of the grain, there are ongoing attempts by some nations to decouple and create independent supply chains. 

This realignment due to strategic and geopolitical reasons is a source of uncertainty and unpredictability, and as such, semiconductors according to some commentators have become the ‘new oil’. 

While the global semiconductor supply chain has been a source of strength, its weaknesses have been exposed during the pandemic. 

While recent chip shortages have abated, there may be other potential shocks on the distant horizon, which have the power to sorely test the supply chain resilience of not just Australia, but the entire world. 

Picture credit: ANFF

One obvious potential flashpoint is the tensions across Taiwan Strait. Taiwan fabricates and exports somewhere between 60-70 per cent of the world’s semiconductors. Any disruption to the outbound semiconductor exports from Taiwan due to conflict portends a supply-chain shock and an economic disruption the likes of which the world has not seen. 

Any changes to terms of semiconductor trade either due to war or conquest have the power to re-alter or realign semiconductor supply chains, giving enormous control to countries in control of these supply chains. 

Keeping in mind what the on-going war in Europe has wrought in disruption to supply of commodities such as crude oil, natural gas, and grains, one can imagine how a black swan event disruptive to semiconductors might affect a country such as Australia, which is heavily dependent on imports of semiconductor products.

Thinking beyond supply chain shocks: Why Australia needs a semiconductor industry

Much has been made about why Australia does not need a semiconductor industry. 

The usual argument goes something along these lines: Australia is too small and cannot support industries that must compete in global markets. 

Some intellectual credence to this argument may be found in ‘New trade theory’, which posits that access to a large domestic market can encourage exports, which in turn may help these companies achieve economies of scale in production, thereby giving them the ability to differentiate their products without cost. 

It is possible to rebut and provide counterexamples to the above argument. 

There are countries comparable to Australia (in terms of GDP and population) which have managed to build thriving local semiconductor industries and ecosystems, despite the lack of a supporting domestic market. For example, Taiwan, Korea, Singapore, and the Netherlands are all smaller or comparable to Australia but are home to highly successful semiconductor companies. 

1. Taiwan is home to more than 90 per cent of the world’s advanced semiconductor manufacturing. 
2. Korea has 18 per cent share of the world semiconductor markets and >56 per cent share of the global semiconductor memory market. 
3. Singapore accounts for 11 per cent of global semiconductor market share. 
4. The Netherlands has semiconductor champions such as ASML, NXP Semiconductors and others, which together have combined annual revenues ~ USD 30 Billion. 

None of these countries have inherent advantages that would make them an obvious choice for a semiconductor industry. 

What they had were committed policy planners or entrepreneurs, who recognised early the importance of economic complexity and diverse income sources in an economy. 

This was especially the case for Taiwan and Singapore. The Netherlands and Korea piggybacked on the successes of their local champions, such as Phillips in the case of the Netherlands and Samsung and SK Hynix in the case of Korea. 

The success of these local semiconductor champions in turn created a sustainable semiconductor ecosystem. Furthermore, all the above-mentioned countries also have educational systems that can build a workforce capable of supporting a growing semiconductor industry. If there is anything thing we can learn from the success of the above countries, it is this:

1. Committed policy planning is central to the emergence of a semiconductor industry.
2. The presence of a local global semiconductor champion can lead to the emergence of a supporting supply chain, which in turn can support emerging semiconductor players.
3. In the absence of a global semiconductor champion, it is even more imperative to have a Govt policy to encourage entrepreneurial endeavours and the development of a semiconductor ecosystem.
4. Educational systems and institutions capable of building a knowledge-based workforce for a growing semiconductor industry are needed.
5. The timescale needed to build an industry spans decades rather than years.

Based on these examples, we can strongly argue that it is feasible to build a thriving semiconductor industry in Australia, provided there is the will to do so.

But why would we ever need one? 

There are some strong arguments for developing a local semiconductor industry in Australia:

1. Increased economic complexity and diversified income sources,
2. Improving resilience in the face of supply-chain shocks, and
3. Sovereign capability in strategic technologies such as 5G & satcom and emerging technologies such as AI and quantum computing.

Curious cases of semiconductor startups in Australia

Picture credit: Getty Images

On the surface, Australia would appear to be a curious choice of place for a semiconductor startup such as Millibeam or even a startup that has managed to achieve scale such as Morse Micro. 

Firstly, there have been very few semiconductor companies of note that have achieved significant success on a global scale. 

Of course, there was Radiata, which achieved some measure of success, but there has not been a global semiconductor champion that has emerged from Australia or one that calls Australia home. As such, there has not been time and space for the emergence of a thriving ecosystem. Secondly, the human capital needed to build a semiconductor industry and ecosystem in Australia can be described as limited at best. 

In spite of the seemingly insurmountable roadblocks, Australia has a number of strengths, such as: 

1. Educational institutions capable of building a knowledge-based workforce,
2. Australia is a destination country for skilled talent,
3. A growing venture industry, and
4. Growing recognition of the importance of 5G and satcom, AI and quantum computing.

Building an Australian semiconductor industry: A few proposals

The strength of Australia in emerging technologies such as artificial intelligence and quantum computing (silicon-based quantum computing), and the desire to gain a foothold in the global semiconductor industry that is expected to hit $1 trillion by 2030, appear to have been the driving factors behind the establishment of the Semiconductor Service Sector Bureau (S3B) by the NSW government. 

S3B is tasked with the development of a market-aware semiconductor ecosystem in NSW. While the establishment of S3B is a positive development that should be welcomed, its limited budget, focus on NSW and its governance dominated by universities in Sydney (despite the participation of CSIRO and ANNF) give it far less national presence. 

Its greater focus on emerging technologies such as quantum computing may end up leading to a more specialised ecosystem that may not necessarily cater for all semiconductor markets and applications. 

A federally funded national body for semiconductors and advanced manufacturing

Building a native semiconductor industry is a nation-building exercise. In this context, what is needed is a federally-funded national body which can develop and implement a national strategy for growing a sustainable semiconductor industry, as well as an advanced manufacturing sector that can be an end user of semiconductors.

A national body with authority to formulate policy, bring together semiconductor startups, academia, established players and end users such as telcos (in the context of Open RAN), satellite communication companies and the defence sector can help grow the domestic and export market.

While some programs and initiatives do exist, they are at best loosely connected. 

Furthermore, building a semiconductor ecosystem should be a multi-decade initiative and without a national body to champion it, changing political and economic circumstances could all too easily derail it. Even a national body may find itself hamstrung under strong economic and political headwinds. 

However, a national body with a charter and structure stands to have a better chance than loosely connected initiatives.

Development of human capital

Ultimately, development of a semiconductor industry requires significant human capital over a long period of time. 

It is no accident that Silicon Valley emerged around Stanford University and in close geographical proximity to UC Berkeley. The reason for this has much to do with availability of talent, which is just as important as access to capital. 

In this regard, Australia faces a chicken-and-egg problem with the talent needed for a semiconductor industry. Firstly, the Australian semiconductor industry is a minnow in terms of job creation and secondly, there is not enough talent available for existing semiconductor startups. 

Picture credit: Intel

Without talent, new companies cannot emerge and without any successful semiconductor companies there won’t be the job growth to justify students flocking to science, technology, engineering, and mathematics (STEM) fields. Only increased government funding in universities targeted towards STEM degrees and research can short circuit this challenge.

Venture capital

Undoubtedly, private capital would have to play a key role in the development of a native semiconductor industry. 

While the Australian venture capital industry has seen rapid growth, it is subject to uncertainty in the macro-economic outlook. Secondly, the investment in deep-tech (a category under which semiconductors fall) lags far behind other investment areas such as SaaS or Fintech. 

The number of committed deep-tech VCs in Australia can be counted on one hand, with firms such as Main Sequence being a rarity rather  than the norm. Greater VC participation may require increased tax incentives specific to the semiconductor industry, than is currently possible under NISA or ESVCLP. 

The incentives provided to early-stage investors in semiconductor or advanced manufacturing startups should be high enough to even out the imbalance in funding between deep-tech and other sectors, an imbalance that is universal and not just limited to Australia.

In the absence of political will for such incentives, the Australian government should consider an investment vehicle like the Clean Energy Finance Corporation, considering the development of a semiconductor industry is akin to a significant nation-building project.

Attracting foreign investment

This is one of the most popular ways in which governments around the world have tried to build native industries. However, it can come with significant costs as foreign entities may demand tax concessions, subsidies, and other support. 

Such ventures have had a mixed record in many countries and should not be the sole approach to building a native semiconductor industry, no matter how attractive or surefire it may seem. 

Often, multinational corporations focus on low wage settings and a low-tax regime as first and foremost reasons, and occasionally access to talent for embarking on such ventures. 

Australia does not have natural advantages in any of these areas and governments may have to contribute heavily to attract such investments. Foreign investment could be a part of the solution, but not the primary one.

Conclusion

No country has the divine right or birthright to host a local semiconductor industry. 

A powerhouse such as America took several decades to become the semiconductor leviathan we know it to be. And an economic superpower like China is still trying to come to grips with the challenges of building a native semiconductor industry. 

All the countries that have managed to build native semiconductor industries of note got there through strategic intent, forward planning, and succouring innovation.

It is time we expand our vision in Australia to include semiconductors and advanced manufacturing.

Dr Venkata Gutta is CEO and founder of milliBeam, a fabless semiconductor company, which develops innovative chipsets for millimeter-wave 5G radios.