Europe can fabricate a die it still cannot finish.

Public discussion of semiconductor sovereignty in Europe is almost entirely a discussion about fabrication. Which node, which site, which subsidy. The step that turns a finished wafer into something a robot builder can solder onto a board rarely appears in the same sentence.

That step is assembly, packaging and test. It is where the die is cut, bonded, encapsulated, wired to the outside world and checked. For most of the components inside an industrial robot, it is also where a large share of cost and nearly all of the geographic risk sits.

Why packaging stopped being the boring part

For three decades packaging was treated as commodity work. Performance came from shrinking transistors, so value moved upstream and the back end moved to where labour and land were cheapest, overwhelmingly in East and Southeast Asia.

Scaling slowed and the industry started buying performance a different way: stacking dies, putting several chiplets on one substrate, moving memory closer to logic. Those techniques are packaging techniques. The step everyone offshored became the step that decides how good the product is.

What this means for a robot builder

A motion controller does not need a leading edge node. It needs a mature part that will still be manufactured in seven years, in a package that survives vibration and heat in a factory cell, delivered on a schedule the builder can plan a production line around.

All three of those properties are set at the back end, not the front. So when a European robotics firm talks about supply exposure, it is usually describing lead times on packaged parts rather than a shortage of wafers.

The European position, stated plainly

Europe holds genuinely strong positions in lithography equipment, in research institutes with advanced packaging programmes, and in substrate and materials specialists in Austria and Germany. It does not hold meaningful high volume assembly and test capacity.

The research capability is real and it is not the same thing as industrial capacity. A pilot line that can demonstrate a stacked package is not a facility that can deliver a million qualified units a quarter.

What to watch over the next two years

Three indicators. Whether any announced European fab investment comes with committed volume back end capacity on the same continent. Whether substrate suppliers add lines in Europe rather than Asia. And whether robotics and automotive buyers start writing packaging location into supplier qualification, the way they already write in second sourcing.

If those three do not move, Europe will keep producing wafers that leave the continent to become products somewhere else, and the dependency the policy was written to fix will survive the policy.