A unique opportunity to strengthen the Spanish chip ecosystem

The global semiconductor supply crisis has highlighted the weaknesses of the current globalized production model, especially when an explosive cyclogenesis is generated in its calm waters by the concurrence of over-demand, rising energy prices, long-distance transport bottlenecks, trade wars... and wars of war.

In the specific case of semiconductors, not only have the consequences of the relocation of microchip manufacturing to Asia become evident, but also its extreme concentration in potentially unstable areas (66% of chips on order are produced in Taiwan). This problem has caused concern in the Western world, which has reacted with plans to support the relocation of this industrial sector.

In Europe, this plan is the Chips Act, which is reflected in Spain in the PERTE Chip (Strategic Project for Economic Recovery and Transformation), endowed with 12,250 million euros. The situation in Europe is weaker than in the USA because here there are no major cell phone and computer manufacturers, which are the recipients of the chips manufactured with the most advanced technologies (currently only available in plants in Asia: Samsung in Korea and TSMC in Taiwan, with Intel's permission), nor do we have any of the major fabless prescribers who design these advanced chips (but do not manufacture them, hence the name fabless).

The European ambition with the Chips Act is to regain a market share of at least double digits. The problem is that there is no European technological capacity to enter advanced nodes. These nodes are known by the name of the prevailing manufacturing technology, CMOS, and are measured by the minimum dimension that can be defined inside an integrated circuit, which in turn determines the computing power (or performance in general) of a chip of a given size. No European manufacturer, or foreign manufacturer on European soil, manufactures in our continent below 20 nanometers (nm), when the current frontier of CMOS technology is in a few nm: 5 nm evolving towards 3 nm (little joke, 3 nm is the diameter of the DNA double helix).

The possibility of attracting the establishment of such advanced factories in Europe was initially seen as a way of recovering technological sovereignty, but this is debatable because the know-how in these nodes must necessarily come from outside, and advanced devices would be manufactured in foreign ownership for a global market, eminently outside the EU.

Even so, producing advanced chips in Europe has intrinsic benefits: it is a technology on which much of the digital muscle of modern and future society rests, and important supply chains of high added value are generated around it; producing chips is like having oil: you make a good business out of it, even if you do not consume what you produce yourself.

Spain is like Europe, but worse. If in Europe there are few semiconductor factories of any kind, in Spain there are none to compare (there were from the mid 80's until the turn of the century). However, associating a PERTE to the world of chips identifies them as a key sector for the future of our economy. Its 12,250 million are not all thought of as direct investment in the Spanish ecosystem. In fact, a good part of them, 9,000 million, are destined to attract advanced foreign factories in line with the Chips Act. This is the minimum bet to put Spain on the European (and global) chip board.

In the Spanish case it is a risky bet because other European countries have more developed microelectronics ecosystems. Here the authorities will have to present convincingly the geostrategic advantages and the potential for generating scientific and engineering talent.

In any case, the rest of the investments contemplated in the PERTE are a unique opportunity to strengthen the Spanish ecosystem. In fact, three of its four pillars are aimed at this, seeking to strengthen scientific capacity in lines of future interest, including quantum chips, personnel training, the establishment of pilot lines that help to transition certain technologies in the field of semiconductors from the laboratory to the market, and the promotion of the generation of fabless companies.

Three more points. The first is that in the semiconductor world there is a dichotomy between design and manufacturing, but we must try to combine both activities. Having separated them geographically is what has led us to the current situation. In the last few months we have been successful in attracting powerful semiconductor design houses to Spain, such as Intel or Cisco. This is a good sign, it will enrich the ecosystem and will lead to the generation of highly qualified employment. Designing can be done from Spain, and it is, in fact. Moreover, Europe is seeking technological sovereignty in the design of microprocessors based on open access architectures, and part of this effort is currently being led from Spain.

The second point is that the factories of highly advanced CMOS nodes move a large part of the world, but not the whole world. Semiconductor devices are not only used to store and process information, but also to exchange this information, and also energy, with their environment. These other chips are essential for the efficient management of the power grid, electric vehicles, the electrification of industry, communications and the Internet of Things. The implementation of such manufacturers would be equally enabling for the industry and would produce devices with abundant use in the Spanish and European market. Advanced nodes may be the spearhead of micro and nanoelectronics, but the long handle behind it is what makes a spear a functional throwing weapon.

The third point is that there is a need for environments that allow the manufacture of small and medium series. This is what start-ups and SMEs, which make up a large part of the Spanish ecosystem, need to develop the demonstrators of their products, or their final products when they are aimed at small niches with high added value. This aspect is part of the fourth axis of PERTE, but its economic allocation could well be greater given its importance.

Along this line of strengthening domestic production, all studies on chip manufacturing point to a persistent demand for CMOS technology nodes in the micron or slightly sub-micron dimensions. These are the so-called resilient nodes, and unlike the advanced ones, they could be developed in Spain. Due to their lower cost and complexity, they are the right traveling companions to provide intelligence to those other devices listed above. This is an option to be explored. It would also be an interesting sovereign technology for sensitive applications, where resorting to foreign technologies is a security risk. In this way, PERTE would not only allow us to have production capacity in Spain, but also to have Spanish production capacity in this field.

Luis Fonseca, director of the Institute of Microelectronics of Barcelona (IMB-CNM-CSIC)