Supply chains must be conceived with the end customer in mind

Interview with Jan-Peter Kleinhans from think tank Stiftung Neue Verantwortung

Since 2020, Jan-Peter Kleinhans has led the area of Technology and Geopolitics at Stiftung Neue Verantwortung, a non-partisan, independent and not-for-profit think tank in Berlin. His focus is on analysing semiconductors as a strategic product, how resilient the global semiconductor value chain is against external shocks, and how Europe’s competitiveness can be boosted. 

Mr Kleinhans, is it right to say that semiconductors now have the same significance within the economy as oil previously had?

Jan-Peter Kleinhans: Its relevance is actually much greater. But at the same time, the topic of semiconductors is also much more complicated than a raw material. Such comparisons show just how much our political thinking is still rooted in the 19th century, when geopolitical power was wielded over raw materials. The difference is that you either have oil or not, however. Semiconductors, on the other hand, are a man-made technology with a value chain which is organised across the globe and completely in private-sector hands. Governments don’t actually play a role at all here, which means they can only provide incentives for domestic semiconductor companies to become more competitive or to assume a stronger position in the supply chain. But in terms of the power structures themselves, they can only change very little. 

Does politics therefore need to intervene more in the value chains?

J.-P. K.: Fundamentally no! Under no circumstances should semiconductors be produced under state control as that would definitely be a waste of money. Even governments intervening in the value chains is a principle that Europe has not employed before. First and foremost, the governments in Europe and around the world should understand semiconductor supply chains better. This is because knowledge of the complexity, mutual dependencies, business relationships and market dynamics is very limited amongst the average political decision makers. They do not understand that it is not a case of a chip shortage, but rather that there are different shortages at different points along the supply chain for differing reasons. They would all need a wide range of regulation concepts in turn. Especially as, for many of these shortages, governments cannot do anything to mitigate them anyway. Because, at the end of the day, we are talking about private sector business relationships, and here you can only provide incentives from the outside – if at all. Currently, many of the concepts are not even half-baked, but have simply been taken from other eco-systems, such as vaccine production, and transferred to the semiconductor supply chain. But that will fundamentally not work here.

What can politics do then to make the ­supply chain more stable?

J.-P. K.: First of all, we need to diversify geographically. Currently, 90 per cent of state-of-the-art chip production is carried out in Taiwan and 10 per cent in South Korea. Geopolitically, it is not sustainable and most certainly not resilient to focus global production in two regions. Diversifying production using subsidies is thus an understandable step so that in ten years we are not necessarily independent of Asia but that there is at least a certain percentage in the USA or Europe, too.

The structure that we have today hasn’t come about for no reason. Can subsidies help here at all?

J.-P. K.: Subsidies do not make it possible to accelerate any kind of catch-up in terms of technology. I think that people have also realised that in Europe and the USA. They are extremely aware that state-of-the-art manufacturing – so currently five nanometres, in the future then four or two nanometres – is only possible with international partners. This means Intel or TSMC moving production to here in Europe. However, political discourse is focussing very much on this one production step and on a particular type of manufacturing, namely cutting-edge fabs. But the value-creation chain is of course much deeper and more complicated. And there are also areas within it where Europe already has a very strong position. But there is nothing to say European companies cannot catch up when it comes to chip design in the next ten years, for example, and be able to adopt a stronger position on the market. 

The question in the whole issue is where is the focus? Do I want a greater market share for European companies, do I want to be a technological pioneer or do I want security of supply? But what use is a greater market share for security of supply? The two have nothing to do with each other. Companies in the US were also hugely affected by the chip shortage, although they constitute over 50 per cent of the semiconductor industry. Take this example: say I need a hundred chips in a modern car. Even if a just handful of these chips cannot be delivered, then I am not able to make the vehicle. So what difference does it make if 20 per cent of these chips are produced in Europe but the other 80 per cent are still missing? Market share and security of supply have nothing to do with each other but are conflated politically.

So what do you think of the EU Chips Act?

J.-P. K.: The EU Chips Act is the geopolitical response that was necessary, but it was put together in haste. It is fundamentally divided into three pillars: research funding, investment and subsidies for production, as well as crisis management and monitoring. Research funding is very much standard, as is to be expected for Europe. There is nothing wrong with this, but Europe already has its strengths in research and we already have leading semiconductor research organisations such as imec in Belgium, Fraunhofer in Germany or CEA-Leti in France. With the second pillar, production subsidies, the focus has expanded a bit in the meantime. People have understood that this does not just involve two-nanometre fabs, but that other types of production can also be innovative and ­helpful. 

“We shouldn’t focus on averting the next shortage, but should prepare for the next shortage.”

For example, those involving new materials such as silicon carbide or gallium nitride. But particularly when it comes to these materials, European companies are already in a strong position. The third pillar contains many mechanisms – especially for the subject of crisis management – which are guaranteed not to work in their current form. For example, that Europe can use joint purchasing power like with the vaccine procurement or, when chips are in short supply, can stipulate which orders have priority. This will not work for such a diverse product as semiconductors. 

In an interview with Deutschlandfunk, you said that we had better concentrate on the design of the chips. But how does that contribute to a resilient supply chain?

J.-P. K.: Not at all. That’s exactly what I meant earlier on – I have completely different political aims. If I am concerned about security of supply, then I don’t need to increase my market share. Instead, I have to see to it that I improve my security of supply through both international partnerships and increased domestic production. But here, it doesn’t matter what role semiconductor companies play in the global market. 

But if I am concerned with this 20 per cent global market share stated in the EU Chips Act, then I have to rely on production steps with high added value, and invest in new European semiconductor companies. The highest share of added value in all process steps is in the chip design, however. Approximately 50 per cent of the share of added value is in the chip design, around 20, 25 per cent in front-end manufacturing, and then approximately 5 to 10 per cent in back-end manufacturing. 

If I am concerned about market share and competitiveness, and about helping to shape technology in Europe again, then the question of security of supply does not arise. First of all, I need companies that develop these chips. And then it makes no difference if this AI accelerator developed in Europe is manufactured in Taiwan or South Korea – because it has been developed in Europe and so the largest share of the added value remains with the European company. But it is precisely this – the distinction between supply security and technological competitiveness – that has been lost as a clear aim by the world of politics. 

So is the EU Chip Act much too complicated for what we want?

J.-P. K.: I wouldn’t say that it is too complicated. But there is no clarity about the aim. You can already see it in the specifics. Sometimes there is mention of 20 per cent production shares, i.e. wafer capacity, and sometimes a 20 per cent market share. But these are two completely different things. For example, the USA has a market share of approximately 50 per cent, but only a 12 per cent production share. Taiwan is precisely the opposite. Taiwanese companies have around seven per cent of the semiconductor market globally, but around 22 per cent of the production capacity. The second pitfall is the issue of supply security. Even if additional production capacities are developed in Europe, operations in these fabs will still be transnational. That means the future fab in Dresden or Magdeburg will still be reliant on Japanese chemicals, US equipment and wafers from Taiwan. So the fab in Dresden or Magdeburg would also be affected if there were supply issues in Japan, the USA or Taiwan. This means we shouldn’t focus on averting the next shortage, but on preparing for the next shortage in all dependent supply chains. 

Here, the end-user supply chain must be adapted, i.e. the automotive supply chain or the medical device supply chain. Strategic warehousing thus becomes a point of focus, for example. But here, even without action from the government, you can see a response in the different user industries. In the automotive industry there are initiatives such as Catena-X, for example, which are aiming to improve visibility and transparency in the supply chain. 

“Market share and security of supply have nothing to do with each other but are conflated politically.”

So should a resilient supply chain be designed with the end customer in mind?

J.-P. K.: Absolutely. Especially as the capacities currently being developed everywhere around the globe are above all cutting-edge capacities for producing high-end chips such as the five-nanometre chip. Today, there is mainly a lack of chips with older technologies, however – 40 nanometres and more. The foundries have hardly any financial incentive to invest in older production technologies though because they have already been completely written off. If you now try to compete on price with a newly built 40-nanometre fab (that is not written off), it is not possible. It is a different state of affairs when the semiconductor customers pay the operators of the fab in advance. This is precisely what UMC is doing, for example. They had a complete 28-nanometre fab pre-financed by a customer. And this means the customer has guaranteed capacities – no matter what happens on the global market, they get the wafers they have paid for in advance, come what may. This means the risk is shifted. Previously, it rested with the semiconductor manufacturer. If there were fluctuations in demand, orders were cancelled and the contract manufacturer was stuck with unused capacity. Now, the risk is shifting and rests on the client. And no government can bring this about through regulation. 

Will the industry manage to make the supply chain more resilient? Or should we keep expecting there to be shortages?

J.-P. K.: We have to keep expecting it. Also because behaviour is irrational in the semiconductor industry – just like everywhere else. There was panic buying here too between the first quarter of 2020 and the third quarter of 2021. Regulation will not prevent that either. Irrational and selfish behaviour on the market will always lead to fluctuations in demand in the supply chain, and therefore shortages in production. This is because it becomes extremely difficult to calculate future demand and plan the necessary investment in production capacities for the future. It is for precisely this reason that it is important to expand our viewpoint. Of course we can do a lot within the semiconductor supply chain; it must be strengthened and diversified. But we also have to look at the end-user industries and get them to also make their supply chains more resilient. Looking at the end-user industries and their responsibilities is essential for overcoming chip shortages in the future.