Chips Acts and Strategic Independence
May 28, 2024
After the Chips Act is before the Chips Acts
While governments around the world are shelling out gazillions of dollars in subsidies and tax breaks for the semiconductor industry to create “strategic independence” – China, USA, EU, South Korea, India, just to name the most prominent – the question may be allowed:
Will these subsidies become a recurring issue for the decades to come for European taxpayers? And what may be missing?
The various Chips Acts around the world – mainly subsidies for the semiconductor industry to support local manufacturing – are in full swing. Funds, which are partially dedicated to R&D and ecosystem investments are in the process of being granted, so are massive subsidies and tax breaks to support the building and equipping of new fabs at the leading edge – in Europe, the USA, Japan, South Korea plus a few other regions.
In the US, the total amount going in reshoring the core semiconductor production is, according to the White House communications, a little short of 300 billion US Dollars until 2030. In Europe, the sum may be “a tad lower” – there is the Chips Act with 43 billion Euro (of which some billions are repurposed from older programs) and the subsidies from several governments to support distinctive fab projects – Germany alone has reserved a minimum of 10 billion Euro for Intel in Magdeburg and 5 billion Euro for TSMC in Dresden. A few billion Euro more for other projects in other countries round out the seriousness with which the “strategic independence in chip manufacturing” is pursued.
Now, everyone has an opinion about the sense of these billions of taxpayer Euros spent for a profitable industry (with admittedly high risks as a nature of the technology). I personally believe that strategic independence in manufacturing leading edge chips only makes sense for the EU if we would have the necessary IP (or companies with this IP) to design such chips and a market with enough critical mass to justify the horrendous investment. But that should not be my point here.
I would rather discuss two different questions here (admittedly related to the opinion above):
1. What happens after the money of the first Chips Act (plus local subsidies) runs out?
2. Where is the strategy of the European Union regarding technology innovation in general?
What happens after the money of the first Chips Act (plus local subsidies) runs out?
This will potentially happen sooner rather than later. In the US, the question has already been raised as the US Inflation Reduction Act is being implemented with rapid speed, and despite some delays in fab projects, by 2026 or 2027 the money will be gone. If Europe’s Chips Act does not drown in bureaucracy, the same may happen here around the same time or a little later. Will everything be rectified by then and independence achieved? I know, a rhetoric question.
I tried a bit of simple math to highlight the challenge.
- Let’s say the semiconductor industry will grow from today’s 550 billion US Dollars to 1 trillion US Dollars in 2030.
- The accumulated revenue in those six years will amount to 5.5 trillion US Dollars, globally.
- The capital needed for chip manufacturing to support the growth in the same timeframe will amount to 1.3+ trillion US Dollars.
- If Europe wants to achieve the goal of 20 % of global chip market manufactured in Europe, as communicated by the Commission, the total capital that needs to be invested in Europe by companies (subsidies included) within that timeframe is somewhere between 250 and 300 billion US Dollars, just for manufacturing.
- Companies expect subsidies between 20 and 50 %, depending on the negotiation talent and the criticality of the project. Let’s assume an average of 30 % of the CAPEX sported by governments, so between 75 and 90 billion US Dollars until 2030.
- The subsidies currently assigned are less than half of that.
- In other words, another 50 billion US Dollars of subsidies might be needed before end of the decade. Potentially more.
I heard opinions of semiconductor CEOs saying that achieving strategic independence for Europe in semiconductors may need more support: 500 billion instead of 50 billion US Dollars (or Euros); not just for manufacturing, but the entire package or ecosystem.
As EU countries are running either giant debts already (Italy and France) or want to save their way out of the crisis (Germany) by letting their infrastructure fall to rubbles (let’s assume not intentionally), I wonder where this investment support should come from? We will certainly have some interesting discussions not too far in the future about massive trade-offs. Nevertheless, I think that serious financial support for staying in the leading-edge game is absolutely mission-critical. The big question is for which endgame?
Where is the strategy of the European Union regarding technology innovation in general?
By comparison, the semiconductor industry is a small business. Of roughly 100 trillion US Dollar of world production of something real (except the hot air the finance industry is generating), semiconductors represent about 0.5%. However, it triggers an electronics industry worth five times that.
And this electronics industry in turn – especially computing and communications - incubates innovation across all industries and life areas that could easily end up at 50 % of the global wealth.
Computers and communications technologies, which roughly make 2/3 of the world’s components consumption, aren’t happening in Europe anymore. While we used to have stakes in telecom and mobile until ten years ago, computers were always dominated by US companies. And it seems this is not going to change, with Artificial Intelligence becoming a booster for computing and communications power.
According to McKinsey and other researchers like Next Move Strategy Consulting, the opportunities of artificial intelligence will dwarf past innovation cycles. By 2030, the AI market in general is supposed to weigh about 5 trillion US Dollars. While software and application revenues (marketing, customer experience) will make the lion’s share of this gigantic market, hardware opportunities – chips, communication devices, servers etc. – probably will end north of 20% of the market, with AI-supporting chips like GPUs and memories easily amounting to 400 to 500 billion US Dollars or 40% of the total semiconductor market.
Apart from the originally British ARM Ltd., the inventor of the ARM architecture (most successful microprocessor/microcontroller architecture ever) and owned by shareholders from around the world, Europe has little to nothing to offer in terms of AI hardware, the juice of any accelerated AI software and application strategy.
Europe’ fate hinges on the strength of its industrial and automotive industry, who today represent 20% of the global components consumption. While automotive is purportedly “driving” a new dynamic with double digit growth in semiconductors, the dynamic is dependent on a successful BEV (Battery Electric Vehicle) strategy, combined with a major effort to turn cars into autonomous vehicles. At the moment, the problem seems to be that this dynamic, or rather hope, is shattered by the sobering development of the BEV market and the investments in the needed infrastructure. The requirements to drive this market to hit the climate goals by 2035 are there, but the reality defies the projections. Unless, of course, everyone falls for the cheap Chinese solution that are beginning to be offloaded at seaports. Meaning, the components market won’t happen here, but in China.
Industrial electronics encompasses a wide field of applications, from automation to energy, from IoT to medical. Many European companies are leading here, globally, but face increasingly aggressive competition from Chinese companies. Ironically, the German “Energiewende” or energy transition to CO² neutral energy depends on Chinese solar panels, wind power or heat pumps.
While industrial electronics mostly does not represent leading-edge for semiconductor suppliers, it could well grow faster on average than the last few decades, helping especially to create this all-electric society that many European associations and companies are talking about (unless most of it will be realized with Chinese technology).
The point I am trying to make here is: the EU in particular and Europe in general have no real plan how to maintain their global competitiveness and no culture of risk or venture to achieve or maintain leadership in fields that will count in the future. We are focusing on things we know and try to improve it incrementally while the world moves on. This is a controversial and black and white statement, I understand, and many people may contradict. But look at how we spend subsidies in the EU: five times more for agriculture (60 billion Euro) than for high-tech (12 Billion Euro). Microsoft spends more money on AI research than all of Europe together.
If we are speaking about future support for the semiconductor industry – once the funds of the current Chips Act are spent – we must understand first, not where we can continue to increment and regulate ourselves into insignificance, but where the music will play globally and what it would take to play a dominant instrument in that orchestra. The opus that will be played is certainly AI Computing. And each Euro given to the components industry in direct subsidies or tax break would need to be topped with another 5 Euro for AI and adjacent technologies. At a minimum!
Why? Imagine our world in 50 years – provided it’s still there and we haven’t ruined it: What do you think will be the predominant technology to run all systems, ecosystems, networks, controls, sensors, data and make sense of it all?
© Teaser Image, Shutterstock