The semiconductor supply chain

Highly specialised expertise and a huge amount of investment is required to produce state-of-the-art chips. An extremely complex and global supply chain has emerged in the semiconductor industry over the last few decades in order to cope with this situation. 

Processors, logic modules and communication components are now in everyday use in many devices. Barely anyone stops to think about how complex the chip production process is: hundreds of companies across dozens of countries are involved in development, design and production. On average, a chip travels several times around the world before the microelectronic component can be delivered. 

Rapid development in chip ­technology

These complex supply chains have developed over the last few decades as a reaction to the increasing sophistication of the technologies. Since the integrated circuit was invented in 1958, the number of transistors per wafer in a logic chip has increased by a factor of 10 million – resulting in processor speeds 100,000 times greater than previously possible. Combined with technical innovations – for example in the field of packaging and material technologies – devices can therefore be developed with exponentially greater computing power, while having increasingly small form factors. To illustrate: today’s smartphones have more computing power than the mainframe computer NASA used to send Apollo 11 to the moon in 1969. They have more storage than a server in a data centre in 2010. 

Different regions, different tasks

Extensive expertise and considerable investment is and was responsible for the remarkable development in semiconductor technology: according to ZVEI (German Electrical and Electronic Manufacturers‘ Association), the industry is investing around 22 per cent of its annual revenue into research and development (R&D) – more than any other industry. The industry’s response is to rely on the division of tasks. As a result, a supply chain structure has emerged in which different regions are particularly strong in certain tasks: for example, the USA is the leader when it comes to R&D-intensive activities such as electronic design automation (EDA), chip design and production technology. On the other hand, East Asia is the leader when it comes to wafer production. China is the leader in the fields of assembly, packaging and testing – aspects which require fewer highly qualified specialists as well as less capital. All countries depend upon one another in this integrated global supply chain and rely on free trade to transport materials, equipment, intellectual property and products around the world to the ideal location to carry out the relevant task. This global division of labour creates enormous value: according to a study from the Boston Consulting Group, developing a complete value chain in an individual region would require an investment of around 1 trillion euros and result in micro-electronics becoming significantly more expensive – by 35 to 65 per cent.

Specialisation is risky

Geographical specialisation has resulted in increased efficiency, but it also comes with risks. According to a study carried out by SIA and the Boston Consulting Group, there are more than 50 tasks in the whole supply chain where one region has more than 65 per cent global market share. Around 75 per cent of semiconductor production capacity and many suppliers of key materials – such as silicon wafers, photoresist and other special chemicals – are concentrated in China and East Asia. Furthermore, South Korea (8 per cent) and Taiwan (92 per cent) are currently home to the most state-of-the-art production capacity in the world – with chip sizes under 10 nanometres. If there were to be disturbances in one of these countries – for example due to natural disasters or political conflicts – this would result in severe disruption to the global supply of chips.

Takeovers are changing the chip landscape

The increasingly stringent requirements when it comes to chip technology and the associated enormous investment over the last few years have resulted in significant consolidation for companies in the semiconductor industry. Every segment along the supply chain – from chip development and the manufacturers of production technology through to chip manufacturers themselves – is now dominated by a few companies. At the start of the 2000s, there were 25 companies who were able to produce state-of-the-art chips. There are now only three companies that have the relevant production skills and technologies. One Dutch, one Japanese and three American companies make up more than two thirds of revenue when it comes to the machines and systems required for the production of semiconductors. 

IoT as a new driving force

IoT now has the potential to significantly transform the semiconductor industry once more. This market ­requires an extremely wide range of chips – including for cloud integration and connectivity, for data processing or for sensors. Moore’s law (which predicts the increase in performance speed and reduction in chip costs) is ­increasingly less relevant here. More emphasis is placed on factors such as electricity consumption, size, software and configurability when it comes to IoT chips. The fact that IoT chips have to be developed for a wide variety of different applications has meant that there are now a number of new companies in the semiconductor industry. The advantage of this is that they benefit from the structures developed in the semiconductor industry over the last few decades. They can concentrate on developing new chips, but can rely on the existing specialist production capacity in order to produce them – the investment costs are therefore limited. This creates scope for lots of innovative ideas and once again completely alters the competitive dynamic in the industry.