Taiwan: When Silicon Stops Being Neutral
Jan 16, 2026
5 min read
Author
Maxime Pasquier, Principal @ Blackwood
Painting: Philippe Jacques de Loutherbourg’s “Coalbrookdale by Night” (1801)
Watching international law get stress-tested again, I find it hard not to think about Taiwan’s fate. More specifically, what would actually be at stake if its sovereignty were ever seriously challenged. With the exponential growth in demand for microchips, Taiwan is no longer just a geopolitical flashpoint, it has become one of the most important industrial choke points in the world.
A few numbers explain why better than any rhetoric ever could. Taiwan accounts for roughly 60% of global foundry output, contract chip manufacturing. When it comes to the most advanced logic chips, the ones that actually matter for AI, high-performance computing and modern defence systems, Taiwan’s share rises above 90%. One company, TSMC, sits at the centre of that ecosystem.
Now looking at the other side of the equation. China is the world’s largest semiconductor buyer. Last year, it imported around $385bn worth of chips, more than it spent importing crude oil… a slight dependency bottleneck, and Beijing knows it.
For years, this arrangement made perfect economic sense. Chipmaking concentrated where yields were highest and costs lowest, supply chains were optimised for efficiency. The problem, of course, is that efficiency assumes stability, and stability no longer feels like a given, nor something some world leaders seem especially invested in preserving. It turns out “just-in-time” works best when nothing ever goes wrong.
You can see the shift in behaviour almost daily.
In the Netherlands, a court is now hearing a fight over Nexperia, a European chipmaker owned by China’s Wingtech. Officially, it’s about governance and control but in practice, I believe it to be about whether Europe is comfortable with strategic chip capacity being controlled from Beijing. The Dutch government briefly stepped in last year, then stepped back to avoid diplomatic fallout, which tells you roughly how straightforward “strategic autonomy” looks in practice..
This happens while Chinese customs officials have reportedly been instructed that Nvidia’s H200 AI chips are no longer permitted for general import. Which raises the question I keep coming back to: why would China restrict the very chips it most wants?
China knows it can’t replicate Taiwan’s advanced manufacturing overnight. But by tightening access to cutting-edge foreign silicon, it nudges domestic demand toward local substitutes, buys time for its own ecosystem, and reduces strategic exposure. Short-term pain in exchange for long-term optionality, at least, that seems to be the plan. It helps that last year China accounted for roughly 42% of global semiconductor equipment spending. Not just buying chips, but buying the tools to make them!
What I find interesting is that the same logic is now showing up at the corporate level. Google’s decision to design its own chips, most recently Willow, is officially framed around performance and efficiency. But owning the silicon roadmap also means fewer supplier surprises, fewer export-control headaches, and a little more control over destiny. It’s the same calculation governments are making, just executed inside a balance sheet rather than a parliament. Sovereignty, it turns out, scales remarkably well inside a hyperscaler, at least for now.
This is the part investors should probably pay closer attention to.
When both states and hyperscalers start investing in chip independence, something structural has changed. Chips are no longer just a cost line, they have become a risk variable, which explains why everyone suddenly wants two of everything.
So what does all this mean for Taiwan?
Taiwan’s dominance is often described as a “silicon shield”, the idea that the economic shock of disrupting Taiwan would be so severe that it deters conflict. That may well be true today. But here’s the uncomfortable implication I can’t quite shake: the more successfully major powers and large corporations reduce their dependence on Taiwanese production, the thinner that shield becomes.
If China can meet more of its own needs, if the U.S. and its allies rebuild meaningful capacity elsewhere, the global cost of disruption falls. Deterrence becomes less economic and more purely political or military.
I don’t think that makes conflict inevitable. But it does suggest the incentives are shifting, slowly, deliberately, and mostly out of sight.
Sources
