Last week’s ECIPE study, commissioned by the US-hyperscaler-dominated CCIA, has done something useful: it has put a price on lock-in. The study estimates that moving European workloads away from the incumbent providers would cost €115 billion over the next ten years. The CCIA presents this as a reason for customers to stay put. It is in fact the strongest argument yet for leaving. The CCIA has established the cost of lock-in, and that cost is small next to what Europe stands to gain by breaking free.
Set the €115 billion against the other side of the ledger. Last year, the Asterès study commissioned by Cigref, France’s forum of digital business leaders, found that Europe spends €330 billion every year on US cloud and software. This year, in a follow-up study, it found that above-inflation price rises by the hyperscalers are pushing that bill up by an average of €140 billion a year between now and 2030. Over the next five years alone, that adds up more than €2 trillion flowing out of the European economy. That is the benefit on the table: €2.2 trillion that could be retained, redirected and reinvested at home, against a one-off transition cost of €115 billion spread over a decade. The benefits to the European economy outstrip the cost of lock-in by more than twenty to one.
And the €115 billion is not money lost. Unlike the €330 billion that leaves Europe each year to create jobs and wealth in the United States, the cost of transition would be spent largely with European consultancies, integrators and cloud providers doing the work of freeing customers from lock-in. It would support firms and workers who pay tax into European exchequers. The CCIA’s figure is therefore less a cost than an investment in European sovereignty, and a very reasonable one.
The study is also a striking admission. Despite the Data Act, the Digital Markets Act and the Digital Services Act, customers of dominant players remain locked in, and the hurdles to moving data and workloads between clouds remain real. By quantifying them, the CCIA has strengthened the case for European cloud providers who abide by the letter and spirit of Europe’s rules, laws and values.
The lesson for those framing and debating the Cloud and AI Development Act (CADA) is clear. The playbook of the 1990s – raise the spectre of switching costs so that customers keep buying what they have always bought – should not set Europe’s policy. In the 90s it was software; today it is cloud and AI. If CADA is to deliver a real increase in European cloud and AI capacity and capability, it must not legitimise business as usual or offer a fig-leaf for non-European services to pose as sovereign solutions. There will be costs to moving to sovereign architecture. The CCIA has told us what they are. They are dwarfed by the €2.2 trillion cost – in jobs, growth, tax revenues and autonomy – of not moving.