How to solve the chipmaking sovereignty challenge

At the end of last year, the European Commission launched the Chips Joint Undertaking . The agreement, part of the European Chips Act, will establish a platform for public-private partnerships, and spur innovation in building a sovereign EU semiconductor ecosystem. ?

I’ve written about sovereignty elsewhere , and it’s the same story here: in a world increasingly wracked by political and commercial (not to mention ecological) tension, establishing the security and sovereignty of your supply chains is key for long-term prosperity.

For #semiconductors, the issue is particularly critical. Not only are they the pivotal technology in today’s digital economy, but manufacture is also highly geographically concentrated: around 60% of the world’s semiconductors are manufactured in Taiwan.

This means it’s a sector unusually vulnerable to disruption. A shortage of just one part can hold up a whole supply chain, with effects cascading downstream to the sectors that rely on these chips. No wonder governments around the world are spending big to bring capacity in-house!

The headline bill here is, of course, the US Chips Act, which has made $52 billion available to support the domestic US chip industry. But not far behind is the €43 billion European Chips Act, which also aims to double EU semiconductor market share (from 10% to 20%) by 2030.

However, it’s not just a case of throwing money at the problem: The journalist Chris Miller?–?author of the bestselling Chip War: The Fight for the World’s Most Critical Technology – –?said as much in conversation with Intel CEO Pat Gelsinger earlier this year:

When you look at a big spending program like this, a lot is going to depend on how the governments involved actually implement it. Do they put money towards the right projects? Do they have structures in place to make sure funding is being used in the right ways? Do they have a vision for how the chip industry is going to develop?

Breaking down the chipmaking industry

Naturally, a vision that will guide such a long-term, large-scale transition needs to be a strategic one. But it also needs to be tactical: able to assess the practicalities of microchip manufacture “on the ground”.

There are two key considerations here:

1.???? Chip design: Moore’s Law observes that transistor density (and by extension, chip processing power) doubles roughly every 18 months. It’s been good for 50+ years. But now the physical limits of the universe are getting in the way of further progress.?

For example, the wavelength of the light used to etch chip components is now thicker than the (sub 5 nanometer) components themselves. That means a laser beam is now too blunt a tool for today’s chips! Designers are finding ways around this: for instance, YV beams, with a narrower wavelength than visible light, can create even smaller designs. Quantum computing circumvents traditional binary computing models entirely with multi-dimension qubits.

2.???? Chip manufacture: However, the most powerful sophisticated chip design in the world is useless if it can’t be manufactured at scale. So, of equal importance are the supply chains that can manufacture and distribute huge numbers of effective chips, at scale, at speed, and without error.

Most of this is done in so-called chip “fabs”, and this is where Taiwanese industries are particularly prominent. TSMC (The Taiwan Semiconductor Manufacturing Corporation) doesn’t design chips; it just makes chips according to the designs of third parties – including Apple, AMD, ARM. This differs from a company like Nvidia (another TSMC client) who focus purely on design; or Intel, who have design and manufacturing operations.

?

Seen this way, the chipmaking sovereignty challenge is twofold: accelerating the design of ever-more sophisticated microchips, while also supporting the industrial capacity to make them at scale. It’s a question of end-to-end resilience: from lab to fab. And I believe virtual twins can be a key role in building that resilience.

Building resilience, from lab to fab

Netherlands-based chipmaker 恩智浦半导体 are exactly the kind of company that will support sovereign EU silicon ecosystem – but an ecosystem is only as strong as its constituent parts. For NXP, there are a lot of things that can go wrong.

Their semiconductors can have as many as 5,000 components, and one component might have 100,000 elements in it: from resistors to capacitors to transistors. Ensuring consistent quality means keeping track of all of this; small defects can quickly add up to big issues in the field.

IP challenges were also particularly worrisome for NXP, whose chips are used in payment technologies. If customers are going to use semiconductors for something as sensitive as payments, then the security of design – and of design processors –?must be put at an absolute premium.

For NXP, virtual twins – and in particularly those supplied by Dassault Systèmes – provided a game-changing solution . “Our demand managers tell me how happy they are with the flexible navigation provided by the DELMIA Quintiq solution.” Says Evelien Klein, Director of Demand Management at NXP.?“We can now rely on a single source of truth,”

And that’s what it all comes down to. Truth. The semiconductor game is not just a case of commercial rivalry. Semiconductors lie at the heart of our most critical systems, from hospitals to finance to defense networks. Those looking to bolster sovereignty cannot afford to compromise on that trust. And when trust is involved, I can’t think of anything more powerful than virtual twins.