We have spent years talking about semiconductors as though they were territory. Countries must take the lead, defend their position and win the chip war. It is stirring language. It is also a terrible description of an industry whose most advanced product may cross borders dozens of times before it reaches your hand.
On Monday, Intel announced a new €5 billion investment in its Leixlip campus in Ireland. The money will expand advanced manufacturing, increase Intel 3 capacity and support current and future Xeon server processors. Most of the spending is expected by the end of 2027. 1
Resilience is not the same as isolation. A stronger supply chain has more trusted connections, not fewer.
Thirty-five years is not a photo opportunity
Leixlip matters because this is not a factory dropped onto an empty map. Intel has operated there for more than three decades. The site sits inside a web of engineers, suppliers, universities, transport links, energy systems and institutional memory. Irish reporting puts Intel’s historic investment in the country at roughly €30 billion and its workforce at about 4,900 people before this new commitment. 2
That long relationship is what industrial strategy often misses. Governments announce a subsidy and photograph a cleanroom, but a fabrication plant is only the visible centre of a much larger capability. The scarce asset is not just equipment. It is the accumulated judgement required to run that equipment reliably, at scale, every day.
For Europe, the investment is a vote for continuity at a moment when AI is pulling more demand toward data-centre processors. It helps preserve skills that would be extraordinarily expensive to rebuild after they disappear. It gives customers another source of advanced manufacturing in a politically stable region. Those are real gains even if they do not produce a European smartphone made exclusively from European components.
Sovereignty needs a better definition
Technological sovereignty is too often presented as national self-sufficiency: mine the materials, design the chip, make the tools, fabricate the wafer, package it and assemble the final product within one border. No country can do all of that at the leading edge without becoming poorer, slower and less inventive.
A useful sovereignty is the ability to make choices under pressure. It comes from knowing which dependencies are dangerous, which are mutually beneficial and where a single point of failure could shut down an economy. That means domestic capability in selected areas, diversified suppliers in others and durable partnerships almost everywhere.
The unglamorous questions
The Leixlip announcement is not automatically an industrial-policy success. Capital spending must turn into competitive production. A fab needs enormous quantities of reliable electricity and water. Europe must train and retain people who can work across materials science, precision engineering and software. Public support should come with transparent commitments on jobs, environmental performance and the length of the company’s presence.
There is also the uncomfortable fact that Intel itself is trying to recover manufacturing momentum. A new investment is evidence of intent, not proof of the final outcome. Serious support for European capacity should allow optimism without demanding credulity.
Retire the battlefield
The chip-war metaphor makes every partnership sound like weakness and every foreign supplier sound like a threat. It invites expensive duplication while concealing the deeper strength of the semiconductor industry: extreme specialisation coordinated across borders.
Europe should want more fabrication, packaging, research and equipment expertise. It should also want healthy Asian and American partners. The point is not to replace one brittle dependency with a fantasy of autarky. It is to build enough capability and enough relationships that no single disruption gets to decide Europe’s future.
Intel’s €5 billion does not win a war. It strengthens a network. That sounds less dramatic, which may be why politicians rarely say it. It is also much closer to how durable technological power is actually built.