Paul van Gerven

14 April

Germany sponsoring that Intel fab isn’t such a bad idea, except for the fact that it comes at the expense of bigger priorities.

At parties, I often get asked how I feel about the Intel fab that Germany is subsidizing. Okay, you caught me: I’m lying. A disappointing number of people actually care about Europe’s ambitions to reduce its dependence on the US and Asia for semiconductors. That’s a shame because I think it’s an absolutely fascinating topic of conversation.

I must admit, though, after thinking and writing about the subject for, well, years if I include Neelie Kroes’ efforts almost a decade ago, my answer would still be a hopelessly nuanced and boring “on the one hand … but then again” kind of argument. I just can’t seem to make up my mind. Just in case we happen to bump into each other at a party or somewhere else, I’d like to share my deliberations here, so we have a head start on our conversation.

Shot in the arm

While the European ambition to increase semiconductor self-sufficiency is generally supported, the fab doesn’t seem to have many fans among Bits&Chips’ contributors. Maarten Buijs, deeply suspicious of the French approach with which the Chips Act is infused, wrote that “it’s extremely important that the European Commission sticks to its guns of forbidding state aid to anything that’s not pre-competitive.”

Bram Nauta pointed out that there’s no market for leading-edge silicon in Europe. We don’t have industries that need it, nor do we have fabless companies operating in that market segment. So why pay a foreign company to make those chips here? We’d be much better off investing those billions in building our own fabless chip ecosystem and own CMOS development and manufacturing capacity, Nauta argued.


8-bit Microcontrollers Still Anchor the Majority of Embedded Designs Today

They are tiny, but vitally important. The market for 8-bit microcontrollers continues to grow strongly as a key part of the drive to digitalisation, highlighted by the current chip shortages. Read more about Microchip’s 8-bit devices.

I, too, have advocated such a bottom-up approach. Interestingly, it looks like this is actually going to happen, with several European companies joining forces to breathe new life into FD-SOI, a very ‘European’ alternative for all but the most bleeding-edge ‘regular’ CMOS. If successful, the alliance spearheaded by Globalfoundries and STMicroelectronics could very well give Intel a run for its money, according to veteran tech journalist Bolaji Ojo.

On the other hand – I told you it would be this kind of argument – what business is it of ours if the Germans want to sponsor an Intel fab? It’s their money – the EU made this kind of state aid possible but isn’t chipping in. Having a fab makes sense for the world’s fourth-largest economy, especially considering the fact that cars will become computers on wheels. And if ASML isn’t worried that government involvement around the globe will create devastating overcapacity, why should anyone else?

From an outsider’s perspective, it’s not a bad thing either if Intel gets a leg up in its efforts to become a strong competitor of TSMC. The Taiwanese foundry is notoriously averse to building leading-edge manufacturing facilities outside Taiwan, which understandably is making the West nervous in light of China ogling the company’s home base.

It’s a bit of a stretch, but there’s an argument to be made that a big, honkin’ fab in Germany will shine a giant spotlight on semiconductor technology, setting all sorts of good things in motion for the European technology sector. It could capture the imagination of students, researchers and entrepreneurs alike, be it in semiconductor technology or in domains that heavily lean on semiconductor innovation, such as AI or communication technology. Building that fab could give the European deep-tech scene a shot in the arm.

Keep everyone happy

I just hope that the Germans know what they’re getting themselves into, though. The next generation of IC technology, typically requiring a new fab to manufacture, is always right around the corner. EU member states supposedly have set aside 30 billion euros for fab funding and semiconductor collaborations (so-called IPCEIs). These funds will leverage additional private investments, but a 30-billion-euro budget stretching to 2030 falls woefully short to move the needle on Europe’s market share in semiconductor manufacturing. I suspect the 20 percent ambition requires a handful of new fabs at least.

Gun to my head, I suppose I do have an opinion. I still think the EU and its member states should prioritize building on Europe’s existing strengths as well as catching the next technology train, such as quantum technology and integrated photonics. I could live with that fab, if I was sure that already meager resources weren’t going to be spread too thinly in an unfocused, have-to-keep-everyone-happy EU fashion.

Main image credit: Intel