Opinion: Money’s not the problem for Europe’s semiconductor rebuild
But money alone is not enough. There needs to be cultural change.
For once a European project is perhaps mentioning figures that are in the right ‘ball park’ (see Europe will try to rebuild semiconductor capability using pandemic recovery funds), although even that is debatable. And I can agree that rebuilding European semiconductor capability, after decades of watching it decline under global competition and globalization, would be of strategic significance.
First off we should not underestimate the task at hand. So let us compare the €145 billion (about US$175 billion) the joint-initiative has its eyes on, with what Samsung is planning to spend simply to catch up with TSMC at 3nm.
According to reports, Samsung is planning to spend US$116 billion.
That’s one company budgeting almost as much as a continent and Samsung is starting from 7nm manufacturing with 5nm prepared but a little behind TSMC in terms of deployment. Meanwhile Europe has more or less let go the art of making digital circuits in advanced processes.
And then there are the 101 other strategic good causes that have their eyes on the same money, which is the “digital transformation” part of the European Recovery and Resilience fund. So it can be seen that while that money is an exceptionally large amount, even if it was all devoted simply to progressing semiconductor capability it would barely be enough. And what about EDA capability and packaging?
That said, leadership, or at least autonomy, in semiconductors is definitely desirable. The question is how best to achieve it.
Next: Now leaving California
Through the combination of world-class research and entrepreneurial spirit the United States created the semiconductor industry and that country’s companies, most notably Intel, have led it ever since, up until now.
Now the US is fighting hard to maintain its strategic position against the rising tide of technical and volume capability in China (see US government reinforces Huawei chip embargo). That is a problem of the United States’ own making. In the 1990s it relaxed the Cocom regulations it had imposed on the world because it’s semiconductor ecosystem companies wanted to be able to sell to a larger market.
Meanwhile raw technical leadership in semiconductors has moved to South Korea and Taiwan. These are, effectively, client states whose independence from China is guaranteed by the US. But the fact that the leadership has now passed from Intel to Samsung and TSMC and the most advanced manufacturing could become unavailable to the US, is clearly driving the US government to provide strategic support for on-shore manufacturing (see CHIPS for America Act promises $22.8 billion in aid and Second piece of US legislation earmarks $25 billion for domestic chip making).
But even if European politicians have also grasped the seriousness of the situation they may have limited understanding of the complexity of technical development and how long it takes to attract and nurture the talent and build up capability.
So we now come to the time-scale element.
Napoleon Bonaparte said of battlefield dynamics: “Space we can recover, lost time never.”
The €145 billion — if it is real money rather than a European Union accounting artefact — may be enough money to alter the semiconductor landscape in Europe and the situation is urgent. But if you spend that money too quickly you will almost certainly spend it unwisely and waste it. Do not underestimate the amount of time it takes to develop a semiconductor ecosystem. Perhaps the politicians do not realize how far behind the leading-edge Europe has fallen and that the leading-edge is now a very specialized place devoted to consumer electronics and high-performance computing.
The idea that Europe will design AND manufacture at 2nm within five years is not only almost impossible to imagine but may be ill-advised despite the strategic aims of self-sufficiency in key technologies.
Repair the disconnect
If we set aside the amount of money and the timescale there are other issues this initiative needs to address. Not least that at present it is solely a political instrument that lacks the commitment of the companies that would be required to act out its requirements and achieve its goals.
We have been here before. Back in 2013 the European Commission was talking about creating an Airbus of Chips (see European Commission repeats call for “Airbus of chips”). The semiconductor companies in Europe, then as now, are multinational enterprises responsible to their shareholders, not to the European Commission or even to their national governments, and they have to deal with the world as it is – including the cost of capital.
As a result, the CEOs of the day had little interest in European strategic thinking and continued to pursue fab-lite strategies albeit with some additional focus on specialized manufacturing. I am thinking of power electronics in the case of Infineon and sensors in the case of STMicroelectronics.
So how will these 17 countries’ politicians change European chip companies’ responsibilities and actions in the 2020s?
At present I don’t think those politicians know the answer to that question, and that is worrying.
Next: Don’t spread the butter too thin
And then there is the European Commission’s tendency to spread the butter too thin.
We’ve heard of Samsung’s astronomical plans and have seen TSMC prepared to spend $30 billion for each giga-wafer fab, albeit over several years. But that €145 billion is intended for digital transformation across the whole continent, excluding the Brexiting United Kingdom of course. So maybe 80 percent of that money will go supporting software and Internet-mediated services and the remainder will have to be pushed towards a broad set of hardware claimants including electronic equipment, PCB makers and the whole electronics ecosystem. Then there is the multitude of flavors of semiconductor manufacturing, assembly, test and packaging that have to be more or less created from scratch. And just as China is finding very few chips get designed without the use of Cadence and Synopsys software, so Europe needs an entire EDA industry to be created for true strategic independence.
All too quickly the fund is doling out paltry amounts for highly specified projects and too much of the money is spent on administration and public relations.
And then there are the geographic interests. Malta and Cyprus have signed up to this joint-declaration not because they have a lot to contribute but because there might be European tax-payers’ largesse for them. I doubt they will want to see the money spent exclusively in Grenoble, Munich, Dresden, Eindhoven or Leuven. From Estonia to Portugal and Finland to Greece there will be expectations of “involvement.”
Which will lead to the classic European Commission problem of taking a large amount of money from the tax payers and then proceeding to spread it too thin to achieve anything.
More of what Europe’s good at
The joint declaration makes the point that Europe is already good at a number of things in electronics including power electronics, RF technologies, smart sensors for embedded AI, microcontrollers, low-power technologies, secure components and semiconductor manufacturing equipment. We have the world-class ASML, a monopoly supplier of extreme ultra-violet lithography equipment. These are good starting points.
We also strong in industrial and automotive sector so we can capitalize on that. But it should also be remembered that not so long ago we were leaders in mobile telephony with Nokia and Ericsson handsets ruling the roost. It is easier to lose a market than gain one.
Which is why this initiative needs focus not on technology but on changing the cost of capital and the commercial landscape. If that can be done established companies will change their actions and startups will be formed. There is then some hope a semiconductor ecosystem has a chance to follow – but not in two or three years; more like two or three decades.
I give credit to the politicians for recognizing that as the Covid-19 pandemic has lowered barriers and created an opportunity. I give credit to the politicians for having ambition.
But they have so far done nothing to change the commercial dynamic that blighted previous initiatives and killed off multi-company fab initiatives, such as those that helped establish Crolles.
Next: Legislate and let the money flow
As I have argued before, legislation is far cheaper and far more effective at achieving change than funded R&D and manufacturing subsidy.
For example if you pass a law that says all lighting on sale in the Europe Union must be diffused in the Europe Union, you would quickly have wafer fabs sited in the Europe Union churning out LEDs. And around those wafer fabs a semiconductor ecosystem would develop.
If you passed a law that mandated AI datasets must be developed in Europe and AI chip training done in Europe for all European machine learning applications, you could quickly develop European excellence in AI. And if there were no inward investors from off-continent you would at least have ensured some level of local competence or self-sufficiency in a strategically important form of computation.
Whether such legislation is in breach of World Trade Organization agreements, and whether European politicians care in the post-Covid era, is for others to decide. But this illustrates how setting the climate is more significant and cost-efficient than trying to fund the detail.
And legislation can be crafted in various ways. It can be more nuanced and yet have tremendous impact. The creation of the Groupe Special Mobile (GSM) standard for cellular communications did far more to generate European wealth than any amount of technical funding could have done. Although I am sure grants and incentives were also handed out to help create that achievement. GSM and the arrival of the cell phone paid hundreds of thousands of engineers’ and executives’ salaries for a couple of decades.
In conclusion, European politicians have identified a strategic issue and there is an enduring political will to do something about it. But politicians are mainly skilled in shuffling papers and spending tax payers’ money. If there is no change in the economic and/or legal landscape there will be no reason for the primary and smaller European chip companies to change what they are doing.
There needs to be bolder political thinking to reverse decades of decline. There needs to be entrepreneurial involvement from the chip companies. There needs to be decades of steady effort.
Related links and articles:
US government reinforces Huawei chip embargo
IPCEI fails to address European dithering
CHIPS for America Act promises $22.8 billion in aid
Second piece of US legislation earmarks $25 billion for domestic chip making
European Commission repeats call for “Airbus of chips”
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
