European AI ambitions clash with a readiness to propose new technology taxes. France's gigabyte amendment and rejected AI-spending levy illustrate the tension, while German statements show competing priorities. Europe has real AI builders. Tax design should support shared prosperity without unnecessarily raising the cost of adopting technology or competing at scale.
On 30 October 2026, France's National Assembly adopted a budget amendment proposing €0.055 per gigabyte delivered by content providers to French users, with turnover caps. That vote did not enact a tax.
I think this captures a European reflex worth breaking: reaching for technology taxes before building businesses with the scale to lead.
A narrow vote, a wide net
Thirty-five deputies voted: 11 for, ten against, 14 abstaining. Éric Coquerel, recorded among the supporters, subsequently said he intended to vote against.
The government opposed the proposal, warning about the burden on smaller French providers. Its author, Laurent Lhardit, projected €10 billion annually; that was his estimate, not an independently verified forecast.
From gigabytes to AI spending
That evening, deputies rejected a separate amendment proposing a 1% levy on companies' qualifying AI expenditure above €100,000, including tokens. The gigabyte measure concerns data flows; this one would have taxed AI use.
Germany's digital minister Karsten Wildberger advocated taxing high AI revenues on 13 October, in a discussion about financing workforce adjustment. His ministry stressed growth first on 16 October. By 2 November, Chancellor Friedrich Merz was calling for technological leadership.
These are different proposals and political positions, not a shared enacted AI tax. But the eagerness to identify a new tax base deserves scrutiny.
In Terry Gilliam's Brazil (1985), heating engineer Harry Tuttle repairs Sam Lowry's malfunctioning system without official authorization. Central Services responds by taking revenge.
I see a useful warning in that absurdity: an institution can become so invested in administering a system that the person making it work becomes an inconvenience. Europe's tax proposals are hardly Gilliam's dystopia. But I would like policymakers to spend more time asking what would help the builders succeed before designing another obligation for them.
Give Europe's builders room to grow
France has Mistral; Germany has Aleph Alpha. These companies deserve a policy environment that helps them win customers and grow.
An AI-spending levy would raise taxable project costs regardless of whether the investment pays off. That risks discouraging customers European suppliers need. A charge based on bytes measures traffic rather than profitability; neither tax base automatically distinguishes an entrenched foreign giant from a local challenger.
I want AI's gains to fund public services and support people whose work changes. That requires careful tax design, not a presumption that every new digital activity deserves its own levy.
Europe's leadership ambition is worth taking seriously. I would start by making it easier for European suppliers to win substantial customers and finance their expansion. Politicians should show as much ingenuity there as they do in finding new ways to tax them.
