TheMarketForIdeas

“Semiconductors became as scarce as gold”. As weird this phrase it may look at first glance, it is actually a reality. And it may be a very grim reality if the demand rate will continue to be way higher than supply. These small electronic components are vital for manufacturing electronic devices and, by extension, for a very broad spectrum of common and industrial goods. Practically, modern humans depend on devices that run on electrical power, and almost all of these depend on semiconductors, chips and other tiny parts. In a way, even if it is not so evident, and even if we are so reluctant to admit it openly, contemporary human society has become deeply dependent on electronics. 

What are semiconductors? 

Very briefly, in common sense a semiconductor (SMC) is a special material that is neither conductor, like metal wires, nor insulator, like dry wood or rubber. Its most useful two characteristics in terms of electronics is that it can let pass the electric current in one direction, while blocking or reducing it in the opposite way and that its resistance can be variable regarding passing electric current. There are multiple types of semiconductors, made of different materials, but the most common ones are silica, germanium and different metallic alloys with special characteristics.

These materials are then used in manufacturing electronic components, like diodes, resistors and transistors, which in turn are base elements for chips and processors. The “chip” it can be regarded as the fundamental part of modern electronics. It is a small and very thin piece of material – hence the name –, usually made of silicon, but not exclusively, with microscopic integrated electrical circuits and electronic components attached to it. Transistors in the chip act as a very tiny switch, letting the current flow or not, diodes permit the flow only in one direction, resistors are passive elements that block, reduce or divide the electric flow, the small metal parts acts as wires while the silica act as the solid base and insulator from other chips. In turn, these microscopic chips are integrated into more complex electronic devices, such as chipsets, memory modules and processors.

The size of the transistors and integrated circuits varies greatly, from micro to nano meters, the smaller ones being harder to produce and therefore more expensive. For example, a modern CPU for computers has billions of transistors, and their size is measured in nanometres – one nanometre represents a billionth part of a meter. For electronics most of the time smaller is better, because a rather compact CPU (Central Processing Unit) of 150-200 square millimetres and 5-6 grams can accommodate more transistors, meaning more computational power, which in turn means a lower time to execute the commands. The same thing applies for a memory module, smaller chips usually mean bigger memory capacity and faster access. 

How does it work? 

Depending on the architecture of the build, it can trigger a specific response when electric current (DC – direct current) is applied. The simplest chips have one single type of action, for example turning a motor on or off. The more elaborate components can execute multiple instructions, according to the programmed instructions (aka the software component). A processor, a chipset or a memory module are the most common piece that is made out of a number (usually a great number) of chips, but this is not exclusive. Commonly speaking, the CPU does the calculations and logical operations, the chipset connects the CPU with the rest of the electronic components working like a dispatcher for electronic signals in or out of the CPU, while the memory (volatile or non-volatile) stores the information processed or to be processed. These sub-components in turn are used to create systems, and these systems in turn are part of the final products. One specific type of CPU is the graphical unit, called GPU (Graphics Processing Unit), which is a type of specialized processor designed to operate and alter graphic images. In a system that uses a display is usually complementary to a CPU, be it integrated physically in CPU or separate, and its main role is to relieve the load on the central processors by processing the video stream separately. 

What are they used for? 

In a word: almost everything electrical that has an electronic controller. Semiconductors in form of chips and other electronic parts are mainly used in IT hardware (PCs and laptops), mobile phones, common electrical consumer products, automotive sector and infrastructure, including military, communication satellites, railroads, utilities supply and so on...

Broadly and non-technically speaking, these chips store data, do computational or logical operations and or start, control, monitor and stop[1] the execution of each action made by electronic and electrical devices. When one pushes the button of a remote control or a coffee machine, electric current is let through and the chips execute what they were designed to – start the TV or the process for an espresso. Similarly, the chips are used for electric motors or actuators: when someone issues a command to the chip by pressing a button that puts it under electric current, the chip executes the command and, according to the design, starts or stops the rotary motor. Practically this turns electricity into kinetic movement, and a certain action is made. On a more complex level, the command to the chip is sent not by a person, like in the case of remote control or coffee machine, but by a more elaborate software system such as PCs or mainframe computers, creating what is called an automation. In reality, an automated action is made up of multiple “tasks” that are executed by different electronic, electric and/or mechanical parts, resulting in a succession of operations that define the entire process from start to finish. Sometimes these iterations are cyclical, as in the case of automatic production lines. 

Share of total worldwide chip spending per sector in 2019

Source: THINK Economic and Financial Analysis – Why semiconductors are as scarce as gold, ING Bank Study, 8 April 2021, https://think.ing.com/articles/why-semiconductors-have-become-scarce-as-gold. 

But, more importantly, beyond the normal usage for home appliances, cars and common electronics products, semiconductors are the backbone of manufacturing industry, because all automated or semi-automatic production lines, whether we are talking about fabrics, toys or cars, use chips to control the flow. This will be even more prevalent in the future, as a result of the so-called 4^th Industrial Revolution. The age of big scale industrial robots will need a huge supply of more and more sophisticated semiconductors in form of chips, multi-role processors and memory modules that will coordinate the production lines. Therefore, without exaggeration, we can say that the entities that will have these small but vital electronic components will rule the manufacturing businesses. In a way, semiconductors will act de facto as a “currency for development”. 

Where are semiconductors and semiconductor-based items produced? 

Since these parts are mass produced and unit value for each piece is rather low, most of big tech hardware companies opted not to invest themselves into manufacturing lines for each kind of items, but to purchase the small sub-components and parts from external producers. These suppliers are based almost exclusively in East Asia, mainly Taiwan and South Korea (approximatively 70% of world’s production is made by entities based in these two countries[2]), and to a lesser extent China, Europe, Middle East and United States. At this level, Asia heavily dominates the market in terms of bulk production. 

Global semiconductors design & manufacturing facilities

Source: Moorhead, P. – On semiconductor issues. A differentiated annual CSR report, Forbes.com, 23 July 2020, https://www.forbes.com/sites/moorinsights/2020/07/23/on-semiconductor-issues-a-differentiated-annual-csr-report. 

However, this is true for rather simpler types of semiconductors like cheap transistors, motherboards (integrated circuits) and low performance memory modules, but not for state-of-the-art processors that include the latest technology, like the ones used for video processors and CPUs. In terms of quality (design, high-end products, prototypes etc.), the US and Europe dominate the global flow.

The technology for making nanometre-chips is a well-guarded industrial secret and only a few players in the world can actually build these types of semiconductors. Even if the actual factories are scattered all over the world, the main producers of CPUs are mostly US or Japanese companies: AMD, Apple, Intel, Nvidia, Broadcom, Qualcomm, VIA, Samsung, Hitachi or Fujitsu.

Synthesizing, the generic logistic chain is as follows: chips and chip-based parts are made mainly in Taiwan or South Korea, high-performance CPUs are made by US or Japanese firms all over the world, from US, Europe and the Middle East to Eastern and South-Eastern Asia (including China), and the final product is usually assembled in China due to the lower cost of the workforce and existing concentration effects. At least this was the case until the recent years.

The geographical imbalance for bulk chip production and electronics manufacturing and is not something new, and the phenomenon is not even limited to this industrial branch. In very simple terms, as globalization deepened, mass production firms looked around the globe for the best set of competitive advantages, and since East and South-East Asia offered cheap workforce and investment facilities (like economic free-zones, permissive labour and environmental legislation etc.), this was the best alternative. Of course, this transition implied a cost for transport of goods from Asia to the US and European markets, but even so the transport charge was covered by the gains resulting from low wages paid to workers. Overall, it sounded like a good profit.

But pushing manufacturing more and more into Asia had an effect in the long term – the home market became dependent on foreign countries, and these countries were not always sound democracies or market economies. So, indirectly, the US and EU domestic markets were subject to influence from another country. This became a big economic security issue, and its real consequences were proved in the beginning of the Covid-19 pandemic. 

The need for more… 

The IT hardware area (including physical networking) has grown exponentially in the last two decades. More and more mobile phones, tablets, PCs, smart watches, smart TVs, smart-“anything-you-want” are needed for consumers. The entire global semiconductor supply chain is estimated to be around half a trillion dollars, while the production process can consist of up to 1000 separate steps, from design to final consumer[3]. Each new generation of devices is more developed than the previous and in turn needs more and better chips. So, economically speaking, we have a market that grows both in terms of quality and performance but also in quantity, in terms of numbers of components built. Moreover, the lifespan of a product of these types is much shorter nowadays, because of obsolescence, requiring new models to be built much faster.

Theoretically, it is a win-win situation for both producers and clients, except for the fact that demand exceeds supply by far, which is not good because a) prices are soaring for the final product and b) the time for delivery of these goods is largely growing. Along these lines, the issue can be framed thus: a “semiconductor crisis” started in 2020 and is ongoing. The global market needed more parts than the producers could supply in a short time.

Like in a perfect storm, multiple disruptive factors overlapped. The chip scarcity of the last years started mainly in the automotive sector, since modern cars resemble more a computer than a traditional pure-mechanical and electrical device. Each car contains hundreds of specialized chips[4] that monitor and control everything from fuel consumption to lights and the internal climate.

The shortage was deepened even more by the new crypto-mining fever. Cryptocurrency can be mined by connecting electronic hardware and pooling its resources via a network; in return for sharing computational power, the “miner” is awarded these crypto-coins. Since the entire domain witnessed not just a simple increase but a true boom, more and more computer hardware parts are needed, mainly video cards and secondarily CPUs. A good mining rig contains multiple video cards connected to one motherboard with a single CPU and RAM memory. More extended rigs could be constructed with up to 24 video cards and multiple motherboards, CPUs, memory modules and power supplies. In this light, it is not hard to understand why the market for these computer parts exploded and prices went up. In consequence, the producers of hardware increased the demand for electronic components, especially for chips – these being the main part of a video processor or CPU.

Last but not least the Covid-19 pandemic changed the way people work, putting a higher emphasis on flexibility concepts, in the form of work from home or work-anywhere modes of labour. This emergency situation pushed the pace of the digital transformation[5] and created an even greater demand for the IT hardware manufacturing sector since employees had to build their own workplace at home, one that resembled the one from the office. For a daily 8-hour job, a small portable laptop, tablet or smartphone represent insufficient equipment both in terms of computing power and health reasons, so the need for more PCs, big screens, webcams and other computer peripherals slightly increased. Again, more reasons for chip demand to go up.

Economic theorists would say: what is the problem? We need more semiconductors. Very well, this means that there are solid incentives for more and more businesses to produce the components, at least until the market reaches an equilibrium between supply and demand. This is exactly what is happening now, but what the economists do not say is that these chips require a special technology to produce… even if there are relatively simple parts, being so small poses a true challenge. It is not easy to build things that have dimensions of just a few nanometres. In fact, it requires a lot of time to develop a production line for microchips, it requires a lot of financial resources and also the deep know-how. No wonder that mostly the big tech companies can manufacture these chips in the quantity and quality demanded. A smaller company could produce specific chips but in a smaller number or with lower performances. And if this was not enough, to make the shortage even worse, there are the political factors that tamper with the economic sphere. 

Geopolitics jumps in… 

Since before the Trump Administration, a form of commercial war emerged between China and the US. It is not complete war since the two countries are so interconnected at the economic level that a complete separation may not be even possible anymore. For example, in 2020, the US imported goods from China worth more than 450 billion USD[6]; this means more than 1 billion USD per day. In fact, even if few are willing to admit it, the US and China are economic “conjoined twins”, each one depending on the other – China needs the money from the US and American consumers need the Chinese products. Still, in some specific areas, the US imposed a ban on Chinese trading, and one of these categories were the microchips. So, broadly speaking, US interdicted its main producers to sell high-end CPUs to China as a state, including some of the Chinese-based hardware manufacturing companies like Huawei. The main motive is that the Chinese authorities may use the technology to develop military programs. Maintaining proportions, it can be seen as an ongoing mini-Cold War[7] on semiconductors between the two main economic powers of the world, each part seeking to maintain or even obtain new advantages. The SMC battleground is in turn a part of the larger “Sino-American” geo-economic and geopolitical rivalry.

Even if the chip shortage was not intended, for whatever reason, this supply chain rupture imposed by the US also seriously affected the electronics manufacturers, especially the video cards and the car parts industrialists, and to a lesser extent the mobile terminals. Since most manufacturers have their facilities located in China, all of a sudden, the required components were out of reach or were available at a higher price and at a later date than was initially planned.

This manoeuvre, in conjunction with a general higher demand for electronics, sparked the beginning of the semiconductor crisis. The effect was not instant, and in fact it emerged gradually, since most of the manufacturers had some buffer stock of parts, but slowly even these stocks have run out.

Right now, we are in a middle of a shortage that was un-imaginable a few years back. It reached the point that for a new car a client could wait up to 6 months, and no, it is not a Lamborghini or a Tesla, but a common car. The producers simply do not have enough electronic components because the suppliers cannot meet the demand. This is how a small microscopic piece of silica could stall an entire industry when multiple disruptive factors converge. 

The re-shoring of strategic industries will be an imperative 

The current situation can be briefly summarized as follows:

1. a) a big demand for semiconductors and electronics due to more complex automotive products (self-driving or assisted driving cars), more and more mobile communication terminals and IT hardware, including Virtual or Augmented Reality gear, combined with the crypto-mining frenzy;
2. b) the supply cannot keep the pace with the rising demand; some estimative figures state that actual demand is 50% higher than production capabilities[8];
3. c) logistics bottlenecks in supply chains for parts and materials, since some metals like wolfram or iridium are found only in some areas of the globe;
4. d) imposed regulations for exporting high-end technology and trade bans due to military or political reasons, and in this case strategic reasons surpassing economic interests;
5. e) cost of transport (land, sea or air) is rising due to higher fuel prices;
6. f) mass-production for semiconductors is concentrated in certain areas of the globe and this is viewed now as a state security problem.

In terms of future growth, a Roland Berger study[9] points out that, in 2022, supply would rise at a steady rate of 6%, a rate that is somehow similar to previous years, while demand would further increase by 17%. This actually translates into a deepening gap instead of a balancing in the SMC market. 

The world market for semiconductors – recent evolutions

Source: Roland Berger – Steering through the semiconductor crisis. A sustained structural disruption requires strategic responses by the automotive industry, December 2021, p. 3. 

That being said, the logical response is to increase production, look for alternative materials and technologies and, most important, to diversify manufacturing locations in order to avoid a future monopolistic crisis. And this is exactly what is starting to happen, but it will require some time until the market will balance.

Increasing production on existing plants is just a partial solution because, firstly, it cannot be doubled overnight, secondly, it needs more inputs and thirdly it does not solve the imbalances of production.

One option is to search for new materials or technologies, but this requires substantial financial resources, and even if will generate good technical results, it may not be economically feasible. Even if it will be, it will take a lot of time until it will be implemented in mass production of semiconductors. And the crisis is taking place now and needs to be addressed in short time, not on a timeline counted in decades. Moreover, increasing inputs also give rise to another challenge: more raw materials; these are not so easy to obtain and traditional mining is not exactly ecologically friendly either. Realistically speaking, there could be other ways, but not so great in terms of viability. For example, one alternative idea was to filter sea water to obtain rare metals for high-end semiconductor alloys[10]. Good in theory, but way too expensive in comparison with traditional mining, at least until the efficiency for large scale operations of this new alternative is achieved. It is true that mining minerals from sea water is not something new in essence, since for example mining salt out of the ocean was done in ancient times, but it is a totally new thing when talking about scarce metals that are found in very small quantities. Practically it will entail the filtering (by active filtration, reverse osmosis or sun-based evaporation) of millions of cubic meters of water in order to obtain small quantities of rare metals.

Then what shall we do? And we come to the re-internalization of production. This is contrary to the phenomenon of externalization of bulk manufacturing processes that occurred in the past four decades. At this moment, given that the labour costs in China, Taiwan, South Korea and the ASEAN area increased steadily, the cost of transport rose in the last few years; and, moreover, we now acknowledge that there is a security concern regarding the supply of items, suddenly the overall rentability calculus shows that it can be equally or even more profitable to relocate manufacturing plants back to European and American soil. In this way, the production will come close to the consumer market, reducing transport charges.

Major US tech companies like Apple, Amazon, AMD or Nvidia would desire to reduce their dependencies on Taiwan/South Korea production as well. Not because Taiwan or South Korea would be geopolitical rivals, but because already the supply cannot meet demand and also because the entire area is “hot” due to political problems between Beijing and Taiwan. In case of military escalations, all of the supply chain is threatened, so therefore for US manufacturers having SMC and chip production plants “at home” would be much safer.

Certainly, there will be serious costs involved, especially since the new factories will have to meet higher ecological standards and, also, a higher paid workforce, but in the (very) long run these investments will pay off. Also, it is imperative to point out that since new facilities will be built, usually in the form of greenfield investment, these can accommodate from the beginning the newest production technologies: industrial robots, IoT devices, fully automatic and flexible production lines. The advantage is that it will require fewer employees since the actual workflow will be made by machines, and by this way the cost of labour it is estimated to cease being a serious problem.

We can frame Intel Corporation’s new initiative in this way, in concordance with the strategic shift of the US Administration, to establish a mega-factory for semiconductors in Ohio. Initially, there will be an investment of 20 billion USD[11], but it can be expanded in time to 100 billion USD[12], making it virtually the biggest industrial investment into a single facility in US history.

On the other hand, we must take into consideration the opposite arguments too. Based on current figures and future estimations, it may not be the best solution in financial terms. And here is why – building and equipping new factories plus training personnel will need solid investment that must be recovered in time. This means that a small part of the final worth of each unit has to contain this passive cost, which will make prices at least a bit higher in comparison to Asian-manufactured ones, considering other parameters to be constant. In the end, strategic reasons may demand it, but profitability calculus, at least on short and medium terms, may be well up against it. 

We did not start the fire 

War. War in Eastern Europe. Any military conflict is generally bad for business, with a few exceptions, like weapon dealers. But in this case, since it involves the Russian Federation and Ukraine, it is especially bad, since these countries are suppliers of base materials and complementary elements for the semiconductor industry. As if the situation was not difficult enough as it was, now the flow of raw materials is severely hindered by the war and the sanctions. It was already underlined that the technology for making nanometre-size chips is not something trivial. In fact, the manufacturing process involves high precision, very thin and very potent laser beams to cut the pieces of material at such small dimensions. There is no other option to cut things at nanometre level. Unfortunately for the whole chips industry, these lasers are not so easy to build and operate either. They require controlled environments with no humidity or dust, and some rare gases, like neon, to be able to operate. The main problem is that Ukraine was a top supplier of neon, accounting for nearly half of world’s production (between 45% and 54%)[13]. Due to the war, the neon supply nearly halved overnight. And even if the producers still have some neon in storage, in the medium term, if no alternatives are found, it will be another major disruption in the entire production process, sharpening the shortage even more and, of course, rising the prices sky high.

On the other hand, the Russian invasion in Ukraine could be the triggering moment for establishing a new geopolitical order and shaping a new form of a Cold War, again antagonizing West and East for a very long time. If this will be the case, then Western economies will no longer be able to rely so much on manufactured products from China and any other country that will be part of the “opposing side”, especially if we are talking about high-tech products. Like every event of this magnitude, it implies a series of problems concurrent with a range of opportunities, depending on the point of view. There will be serious drawbacks in terms of R&D since the information and know-how will no longer be shared so easily and there will be more and more shortages in various sectors, at least until most of the industries will be re-internalized and up and running. The good side is that there will be new job openings and, after the transition period, the domestic markets will no longer depend on foreign and distant countries. 

Some conclusions 

Firstly, we must admit that this supply crisis may last for a while, particularly because of the war in Ukraine and the cooling relationship between China and the United States. These factors represent just two of the major barriers that hinder the market for the moment, but in the long run there are others as well: higher prices for materials, environmental regulations to be fulfilled, rising costs for labour and transport, greater cost for research and development of new products etc. If the demand will have a greater rate of growth than the supply, this crisis may become chronic (arguably it is already, since it has been going on for two years at least). We can expect an even greater scarcity of components, and this means that prices will stay quite high.

Secondly, even if it is not quite in the headlines for news, this issue is actually very important for businesses around the globe because these tiny bits of silica and metal are basic ingredients for every electronic, home appliances, automobile or industrial equipment. Practically any shortage in the semiconductor sector expands to other sectors as well, resulting in higher prices or longer waiting periods for final products. As the strong laws of economics state, higher prices and longer delivery times mean fewer buyers, which in turn mean lower revenues for companies and, indirectly, for states. In other words, and without being melodramatic, in a bigger picture this means the slowing down of economic development. It is true that the impact is not catastrophic, but we must admit that it does some damage to a world economy already shaken due Covid-19. Or at least it reduces the pace of post-pandemic recovery.

What will happen after the Covid period? Probably, as studies from McKinsey & Co. pointed out in 2020[14], there will be an increase in R&D and a focus on developing the next generation of semiconductors, ones that ideally will be cheaper, faster to produce and with better performances. However, the road form prototype to mass production is long and difficult.

Thirdly, the current semiconductor crisis may be exactly the trigger and incentive for industrialists to re-shore the manufacturing of strategic products, as chips and semiconductors are at the top of the list. Yes, it will cost a lot to create new factories, to equip them with state-of the art machinery, to select the workforce, to train it and, most importantly, to pay it at US or EU wage levels, but it is mandatory in order to keep the future economy safe and sound, especially since the geopolitical perspective is not very bright. If there will be two political and economic blocks like it was in the Cold War era, the globalization phenomenon in terms of trade will take a serious blow, since domestic economies will no longer be able to rely solely on imports for certain parts, especially if these items are made “on the other side of the Iron Curtain 2.0”.

Last but not least we can expect that the future manufacturing industry dominated by IoT and VR/AR devices will need an even greater supply of chips, semiconductors, integrated circuits and all other electronic components needed “to do the magic” and create fully automated production facilities. In this light, it is not an exaggeration to say that semiconductors will be the future currency for development. So, tell me how solid is the supply chain for these parts of prime importance so I can tell you if your economy will be a competitive one or not. 

Notes:

[1] Kamasa, J. – Microchips: Small and Demanded, CSS Analyses in Security Policy, ETH Zürich, No. 295, December 2021, p. 1.

[2] Choudhury, S. R. – Tough road ahead for U.S. firms trying to cut reliance on Taiwan chipmakers, CNBC.com, 13 April 2021, https://www.cnbc.com/2021/04/13/semiconductor-shortage-us-tech-companies-and-their-reliance-on-taiwan.html.

[3] Khan, S.M., Mann, A., Peterson, D. – The semiconductor supply chain: assessing national competitiveness, Center for Security and Emerging Technology, January 2021, p. 5, apud. Alam, S. et al. – Globality and complexity of the semiconductor ecosystem, Accenture and Global Semiconductor Alliance, 2020.

[4] Wu, X., Zhang, C., Du, W. – An analysis on the crisis of “chips shortage” in automobile industry – based on the double influence of COVID-19 and trade friction, Journal of Physics: Conference Series, No. 1971, 2021, https://iopscience.iop.org/article/10.1088/1742-6596/1971/1/012100.

[5] Pitkin, W.W. Jr. – Chip shortages: created by demand, geopolitics, pandemic and mother nature, State Streets Global Advisors, Fundamental Growth & Core Equity, June 2021, p. 4, https://www.ssga.com/international/en/institutional/ic/insights/chip-shortages-created-by-demand-geopolitics-pandemic-and-mother-nature.

[6] Office of the United States Trade Representative – The People’s Republic of China, U.S.-China Trade Facts, https://ustr.gov/countries-regions/china-mongolia-taiwan/peoples-republic-china.

[7] Eurasia Group – The Geopolitics of Semiconductors, September 2020, pp. 3-4.

[8] The Times of India – How chip shortage has changed the industry, created new new giants, 9 November 2021, https://timesofindia.indiatimes.com/gadgets-news/how-chip-shortage-has-changed-the-industry-created-new-new-giants/articleshow/87601039.cms.

[9] Roland Berger – Steering through the semiconductor crisis A sustained structural disruption requires strategic responses by the automotive industry, December 2021, p. 3, https://content.rolandberger.com/hubfs/07_presse/20211214_RB_ART_21_037_WPs_CES_Part2_Semiconductor_Shortage.pdf.

[10] Loganathan, P., Naidu, G., Vigneswaran, S. – Mining valuable minerals from seawater: a critical review, Environmental Science Water Research & Technology, No. 3, 2017, pp. 39-41.

[11] Semuels, A. – Intel reveals plans for massive new Ohio factory, fighting the chip shortage stateside, Time.com, 20 January 2022, https://time.com/6140476/intel-building-factory-ohio.

[12] The Real Deal – Intel investment in new Ohio chip plant could reach $100B, 23 January 2022, https://therealdeal.com/2022/01/23/intel-investment-in-new-ohio-chip-plant-could-reach-100m.

[13] CNBC.com – Russia’s attack on Ukraine halts half of world’s neon output for chips, 11 March 2022, https://www.cnbc.com/2022/03/12/russias-attack-on-ukraine-halts-half-of-worlds-neon-output-for-chips.html.

[14] Bauer, H. et al. – How the semiconductor industry can emerge stronger after the COVID-19 crisis, McKinsey & Company, Advanced Electronics Practice, June 2020, p. 6.