Driven by acute shortages and growing geopolitical competition, semiconductors have shot up the EU policy agenda over the past year. Caught in the crossfire of a global trade war and exposed to the vagaries of an undiversified supply chain, Europe has understood how its strategic autonomy is dependent on these chips. Looking further ahead, Europe aims to diversify its economic structure and be at the forefront of the next generation of digital technologies. Chips will be integral in this effort, both as an enabler for a range of technologies, from AI to supercomputing, and as an extremely valuable component of global value chains in their own right.
For now, Europe accounts for only 10% of the industry’s market share. European policymakers have committed to doubling Europe’s share of semiconductor industry production to 20% by 2030, including developing capacity for next-generation manufacturing at 2nm nodes. As part of the Industrial Strategy Update, an Alliance on Processors and Semiconductor Technologies was announced and a new Important Project of Common European Interest (IPCEI) is expected.
Much of the focus has been on manufacturing and on building up Europe’s existing strengths in automotive and industrial chips. However, the expense and level of technological sophistication required mean it will take years before Europe can develop cutting-edge manufacturing capabilities. Promoting Europe’s chip design ecosystem could be a cost-efficient way to climb the semiconductor value chain, diversify our economy, and carve out a strong position at the technological frontier without abandoning the benefits of interdependence.
The Global Semiconductor Industry
The semiconductor industry is global, segmented, concentrated, and highly sophisticated. Powerful economic and technological forces have resulted in most sections of the value chain being dominated by a handful of companies spread across America, Asia, and Europe. Only a few companies, such as Intel, both design and manufacture their chips. Instead, most semiconductor designers contract out their manufacturing to dedicated manufacturing companies like the Taiwanese TSMC, accounting for over 50% of market share. The technological sophistication required means that only TSMC and Samsung can manufacture the latest generation of cutting-edge chips.
Though it represents only 10% of the global market, Europe has some strengths, such as equipment manufacturers like ASML. Moreover, Europe has a strong presence in the chips necessary for cars and industrial processes, although these are less sophisticated than the digital chips designed by US companies. The various European companies operating in this space have some manufacturing capacity but still outsource much of their production. Hence, they have felt the brunt of the semiconductor shortage, as higher value costumers ranging from Qualcomm to Apple are prioritised.
It is important to distinguish efforts to develop high-end semiconductor manufacturing and the mature manufacturing capacity that is still critical for sectors such as the automotive industry. The options for state support in the latter case are limited, since state aid instruments like IPCEIs require technological innovation and cannot be used to subsidise mature industrial capacity. Whether firms develop greater European capacity will depend on their market-driven, cost-benefit assessments. The Alliance can serve as a mechanism for facilitating and coordinating private investment, as has been the case with the Battery Alliance, perhaps with loans from the EIB and national policy banks; however, it is still driven by the private sector’s decision-making. As semiconductor firms across the world increase their capital investments in response to the current shortage, European players may decide that the global supply chain will meet their needs. However, even if firms choose to develop more European capacity, it would do little to develop Europe’s position beyond exiting niches.
The challenges posed by high-end semiconductor manufacturing are of a higher order of magnitude. The latest generation of factories cost between $15 and $20 billion and carry considerable technological risk. Even Intel, the dominant semiconductor company, has struggled to keep up with the latest generation of manufacturing processes, while China has failed to make significant headway despite large investments. Only TSMC and Samsung have succeeded. Furthermore, the applications the European industry specialises in are unlikely to require advanced manufacturing processes at the necessary scale. American tech companies would need to place orders when they already have stable and fruitful relationships with established firms.
As such, policymakers appear to be focusing on convincing existing market leaders to set up facilities in Europe. Doing so will still require significant public support, with reports of $10bn in subsidies requested. Assuming that such public support could be given in a state aid conforming manner, whether the cost is justified will depend on three factors. First, an assessment of the benefits to strategic autonomy of having high-end manufacturing in Europe – the experience of ramping up vaccine production has shown this can be substantial. Secondly, the direct economic value created. Thirdly, the spill-overs to the wider industrial ecosystem. This third factor is perhaps the most important, since the ultimate goal is to create a rich innovation and industrial ecosystem that will foster high-value European companies. A careful assessment of the industry’s structure and the various costs and benefits will be required to make this decision, but it should not be viewed as the only way to develop Europe’s semiconductor industry.
If Europe wants to grow high-value digital sectors and enhance its strategic autonomy, chip design deserves as much, if not a greater, focus as manufacturing. Designers not only capture the lion’s share of the industry value-add. From a strategic autonomy perspective, they underpin the modern economy just as much as manufacturers do. Focusing on design over the medium term need not be seen as an alternative to manufacturing, since strengthening our design industry could improve the business case for more advanced manufacturing capacity.
Current indications are that European efforts beyond manufacturing will be concentrated on existing strengths in 5G, industrial, and automotive applications. This market is expected to grow considerably and maintaining Europe’s market share is also clearly important. However, the entire semiconductor industry is in considerable flux with the growth of cloud computing and AI, creating opportunities for new actors. Their associated geopolitical power and the implications for European strategic autonomy will be significant.
China’s experience demonstrates that it is significantly easier to catch up in design than in manufacturing. Europe’s strategy could focus on distributing R&D funding to a broad range of design projects and measures to direct private finance towards innovative young companies. After all, the current giants in the field, including Intel and Nvidia, began as start-ups with venture capital funds. Promoting new players in design could be particularly cost-effective and, unlike manufacturing, would not require a significant amount of expensive investments with a high degree of concentrated technological risk.
Europe has traditionally found this kind of industrial policy – nurturing a rich R&D ecosystem and financing young innovative companies – challenging. However, examples like UiPath and Northvolt show how European industrial policy can direct financing towards such companies. For such an approach to succeed, European policy will need to enable more financing for companies to scale up beyond their early stages and to support them to remain centred in the EU, particularly by strengthening the Capital Markets Union. Otherwise, we might see successful companies leaving; after all, UiPath may have begun in Bucharest, but its headquarters and initial public offering are now in New York.