Smart factories (e.g., the industrial Internet of Things (IoT), automation and artificial intelligence (AI)) are radically changing manufacturing. The smart factory revolution is currently on hold, however, due to the shortage of one of its key components, semiconductor chips. Globalization, which has been an important driver of growth and the creation of wealth for decades, has also made our world far more complex. Have we reached a point where this interconnectivity is turning into a liability?
To understand why this question is being asked, we must first look at the reasons for the current semiconductor chip crisis and its critical role in the modern-day world. Measures that are being introduced to deal with the current bottleneck will also be considered.
How did we get here?
As is so often the case, the current semiconductor crisis is not a one-off problem that appeared out of the blue, it resulted from multiple overlapping crises that have characterized the early 2020s. They include the industry structure itself; limited capacity available to meet the global surge in demand for semiconductor chips, and rising protectionism and geopolitical instability.
What is not likely to change anytime soon is the fundamental characteristic of the semiconductor industry itself. The production of semiconductor chips is not only extremely capital-intensive, it is also a very time-consuming process that consists of hundreds of individual steps. First, 99.99 per cent pure mono-crystalline silicon ingots (cylindrical rods) have to be produced, diced into slices and polished. They are then coated with photoresist, exposed to extreme ultraviolet light (in so-called lithography, ultraviolet light hits the photoresist, causing chemical changes).
In a next step, the wafers are cleaned, doped and etched multiple times to create the intended pattern. The complexity of this production process becomes all the more apparent when we consider positive or negative photoresists, wet or dry etching, the wide range of available silicon-dopants and the fact that those processes need to be repeated countless times. Finally, the wafers are ionized. Once all of these steps have been completed (which can take up to three months), the wafers are diced into individual chips, inserted into cover-media and bonded (the so-called “packaging”). This final production step is extremely labour intensive, that is, the COVID-19 lockdowns had a severe impact on this stage of the production process. Such complex, multi-stage production processes are challenging in the best of times, but are nearly impossible to maintain in the face of lockdowns and supply chain disruptions.
Moreover, the semiconductor industry is highly capital-intensive (depending on the size of the wafers. Building a semiconductor fab can cost well in excess of USD 10 billion). Hence, in economically uncertain times, manufacturers are hesitant to expand capacity to more than the bare minimum, not least given the long-term implications such a decision has – building a fab is a long-term investment and takes around 3–4 years to be online. Manufacturers’ reluctance was further buoyed when the automotive industry cancelled all orders at the onset of the COVID-19 crisis to optimize its own stock levels. Fabs do not make a profit below an 85 per cent capacity load, i.e. a near optimal capacity load is necessary for a fab to be economically viable.
When in 2020 car sales resumed quicker than expected, and remote working and home schooling increased demand for laptops, the demand for chips suddenly skyrocketed. The perfect semiconductor storm was created when China began stockpiling chips1 in response to ever-increasing trade tensions and the industry’s limited capacity failed to meet the explosive demand.
The implications of the semiconductor shortage
The semiconductor industry reflects the state of the global economy. Optimized to maintain the lowest possible level of inventory and maximum capacity utilization, it is hardly surprising—at least in hindsight—that global supply chains are in disarray. This may be easy to say now, but warning signs were already visible prior to 2019. Globalization started crumbling amidst the often abstract discussions on protectionism materializing in the real world, export stops and stockpiling. The impact of the semiconductor shortage on the automotive industry can only be described as dramatic. Around 11.3 million cars could not be produced in 2021 as a result of the semiconductor shortage2, and a further 7 million cars are assumed to not have been produced in 20223. Forgone sales in this industry alone have run into the hundreds of billions. Given the fragility of global value chains and the expectation that the underlying conditions will not improve any time soon, questions about how this will all play out are growing.
In particular smart factories critically depend on semiconductor chips. The Industrial Internet of Things (IIOT), which connects machines and devices and uses sensors and actuators to continuously monitor all stages of the production process, runs on semiconductor chips. Edge computing combined with 5G networks transmit data in real time, feeding into artificial intelligence (AI) control systems, also heavily depend on semiconductor chips. With the huge volumes of data being generated, exchanged, analysed and used in real time, this demand for chips and programmable logic controllers (PLCs) is only bound to rise in the future. Especially as it is not only smart factories and the Industrial Internet of Things (IIOT) which depend on semiconductor chips, but also smartphones, personal computers and cars.
Think big picture: we need long-term crisis management strategies
The three inter-locking crises, namely the processes of the industry, the semiconductor industry’s limited capacity, and the rise in protectionism and instability will need to be tackled to solve the semiconductor shortage. Long-term and big picture thinking about the global economy points to a need for more structured approaches.
High fab utilization and continuously rising demand (smart factories, hybrid work, etc.) call for an increase in capacity on the part of semiconductor manufacturers. Global supply chains have always been complex; supply chain disruptions and protectionist measures have only magnified their complexity. Global supply chains will increasingly hinge on stronger, more durable supplier relationships; the widespread short-term transactional approach with the aim of optimizing own stock levels, often at the cost of downstream suppliers, is no longer sustainable.
To achieve long-term stability, we must also take the changed geopolitical situation into consideration. The semiconductor industry is concentrated in a handful of countries, including Taiwan, Province of China, the Republic of Korea, Japan, the United States and to a lesser extent, smaller players in Europe. While no country exclusively controls all stages of the complex semiconductor production process (that is, no country has achieved what the European Union refers to as “technological sovereignty”), the level of concentration of semiconductor production stages in a handful of countries is nonetheless alarming.
The design of semiconductor chips is firmly in the hands of the United States, with established players (the so-called fabless companies such as Qualcomm and Nvidia) and new players budding to enter the fray (Tesla or Alphabet, to name just two). Production capacities, on the other hand, are heavily concentrated in Taiwan, Province of China, and the Republic of Korea. With so much depending on the availability of semiconductor chips, the concentration of much of the global foundry production in two potentially highly unstable geopolitical regions gives cause for concern.
Large multinational corporations (MNCs) have already taken action. A flurry of greenfield projects indicates their intention to shorten supply chains and to de-risk their operations. Samsung, for instance, is rumoured to invest USD 17 billion in the United States, while Intel has announced plans to invest up to USD 80 billion in Europe. Micron plans to invest USD 7 billion in Japan.
FDI in semiconductor manufacturing
Within this context, the European Commission introduced the European Chips Act on 8 February 2022, with the intention of kick-starting major developments in technological capacity building, the construction of a 2 nm chip-fab to ensure the supply of super-fast chips, and to increase Europe’s market share in global chip production to 20 per cent. While some may criticize this “EU fab” or “EU foundry” (at present, Europe primarily uses 7 nm chips), Europe would benefit from a more visionary approach rather than settling on the smallest common political denominator, especially given that the automotive industry, medical device manufacturers, as well as the very existence of cutting-edge smart factory concepts, depend on access to the latest chips generation. Hence, refraining from taking decisive action to address the current supply chain crisis and concentrating chip production in potentially unstable regions would be rash. Similar policies are also being pursued by China and the United States (see, for instance, the U.S. Chip Act and the ‘Executive Order’ signed by President Joe Biden in February 2022, earmarking USD 52 billion in incentives to support domestic semiconductor production).
Ensuring supply chain resilience may imply a retreat from globalization
There are no quick or simple solutions. Highly complex production processes heavily rely on specialized machinery, which first need to be built; what may pose an even bigger problem is that key consumables in semiconductor production, such as neon, krypton, or argon, depend on key suppliers, in particular Ukraine and the Russian Federation. These key consumables are currently also experiencing severe shortages, and have exacerbated the semiconductor crisis even further.
In the semiconductor industry—as in the global economy—profit margins might become slimmer in exchange for ensuring continuity of supply; hyper-optimized, JIT-like models might give way to goods in stock, and transactional zero-sum supplier-customer relationships might be replaced by more strategically collaborative networks. Hyper-globalization is on the retreat, which is not necessarily a new trend, but global circumstances are increasingly forcing countries to adopt strategies that have been sensible all along, namely a stronger emphasis on regionalized, as opposed to globalized, value chains.
Disclaimer: The views expressed in this article are those of the authors based on their experience and on prior research and do not necessarily reflect the views of UNIDO (read more).