European Tech Sovereignty Requires European Funding

In June 2026, Europe got a live demonstration of what dependence costs.

The US government ordered Anthropic to disable its two most capable AI models, Fable 5 and Mythos 5, for every foreign national on earth, citing national security (Al Jazeera, 2026). Not a tariff, not a review, not a phased restriction. One letter, and within hours the models went dark for everyone outside the United States, including foreigners working inside the US, including Anthropic's own non-American staff. Europe was not targeted. Europe was simply on the wrong side of a border when a switch was flipped in Washington.

The European Commission's response was measured and, read closely, an admission. Contingency measures "should not be discriminatory against partners," its spokesperson said, while conceding the episode is "another illustration that Europe needs to strengthen its technological sovereignty" (Euronews, 2026). Translation: we did not see this coming, we could not stop it, and we are now reading about our own dependence in the news.

Sovereignty is not about having the best model. It is about who can turn it off.

This is the part the debate keeps getting wrong. The question was never whether American AI is good. It is excellent. The question is who holds the switch. And in June 2026 the answer became impossible to ignore: a capability that European companies, researchers, and public bodies had quietly built into their workflows could be revoked, overnight, by a government that does not answer to any European voter.

A dependency you cannot influence is not a partnership. It is an exposure. And this one is not unique to AI. Swap the noun and the same sentence holds for cloud infrastructure, advanced chips, payment rails, and satellite capacity. The Anthropic shutoff is not the story. It is the preview. Who knows what is switched off next, and on whose schedule.

None of this is an argument to wall Europe off or to stop using the best tools in the world. Autarky is a fantasy and a bad one. The argument is narrower and harder to dismiss: Europe needs the *optionality* to keep functioning when someone else flips a switch. Optionality means alternatives. Alternatives mean European companies actually building the critical layers. And companies do not appear because a Commission communique wills them into being. They appear because someone funds them.

This is no longer a fringe position

Two things are true at once now. The dependence is real, and Europe has finally started treating it as an emergency.

Weeks before the Anthropic shutoff, on 3 June 2026, the Commission had already put one of the largest industrial-policy bets in its history on the table. The Technological Sovereignty Package, spearheaded by Commissioner Henna Virkkunen, targets an estimated 422 billion euros of investment across chips, cloud, AI, and data centres over the next decade (Fortune, 2026). Its pillars are concrete: a Cloud and AI Development Act to triple Europe's data-centre capacity, a Chips Act 2.0 that deliberately aligns chip designers with European industry demand, and Data Center Acceleration Zones to fast-track the physical build-out (European Commission, 2026). Virkkunen has been explicit that this is not protectionism. The aim is to stop depending on a single dominant supplier who does not share Europe's interests, not to close Europe off to anyone.

Then the switch flipped, and a policy debate became a fire alarm. The chorus stopped sounding technocratic. France's Édouard Philippe called AI "critical infrastructure, as essential as electricity or the internet." Bruno Retailleau put it more bluntly: "a nation that depends on others for its technology is a nation that can be unplugged overnight." A UK security minister observed that national security is now "more about code than cannons," and MEPs in Finland and Germany called for pooling European compute (Fortune, 2026). Ursula von der Leyen had already set the frame: "We cannot afford to depend on others for the technologies that keep our hospitals running, our energy grids stable and our services secure." This is no longer a worry voiced by a few hawks. It is the Commission, heads of government, and parliamentarians across several countries, saying the same thing in the same month.

And the most credible voice said it first. Mario Draghi's competitiveness report put the price of staying in the game at roughly 800 billion euros a year and told Europe, in plain terms, to stop behaving like a regulator and start behaving like an investor (European Parliament, 2025).

Which makes sovereignty a funding problem

Now notice what every one of these announcements quietly shares: a very large number, and no confirmed way to pay for it.

You cannot have sovereign technology you did not pay to build. The Commission's 422 billion euros is a target, not a cheque, and it has not said where the money comes from. Draghi's 800 billion a year is a gap, not a budget. Public instruments like the proposed European Competitiveness Fund can de-risk and seed deep-tech, but no public purse funds an entire transition. The bulk has to be private capital, deployed across exactly the deep-tech categories where the cheques are biggest and the patience required is longest.

And this is where the sovereignty ambition runs straight into a wall we have written about before. The money exists: Europe out-saves the United States, then sends roughly 300 billion euros a year out of the EU and funds barely one fifth the venture capital the US does (CEPR, 2025). The private capital that should be building European alternatives is sitting idle in deposits or quietly funding the very ecosystem Europe now says it wants to be less dependent on.

Virkkunen's instinct on chips, to align the builders with the demand, is exactly right. It simply has to extend to capital. You cannot fund a sovereign technology base with money that refuses to cross its own internal borders, and you cannot align European capital with European builders if the two can barely see each other. Sovereignty and capital fragmentation are the same problem wearing two different hats.

Funding the transition is a matching problem

So how does Europe actually pay for this? Not, mostly, with new money. As we argued in the last post, the capital already exists; what is missing is the ability of that capital to find the companies worth backing. The builders of sovereign European tech are scattered across 27 markets. The capital that could fund them is scattered across the same 27. Right now they largely cannot see each other, because the networks that move money still stop at national borders.

That is a discovery problem, and discovery is the cheapest lever Europe has. It does not require a treaty, a new fund, or a decade. It requires that a French investor with a deep-tech thesis can find the Estonian or Finnish or Portuguese company building exactly what Europe is about to need, and reach out, before that founder gives up or relists in Delaware.

Be honest about the limits. Money alone does not manufacture sovereignty. Talent, compute, energy, and sane regulation all have to show up too, and a single discovery layer fixes none of those. But of all the inputs, capital is the one Europe demonstrably already has and demonstrably wastes. Mobilising it is the fastest move available, and it is a prerequisite for everything else. Nothing gets built unfunded.

Floancer wants to be part of the build-up

We are not going to pretend a discovery platform builds sovereign chips. It does not. But every gigafactory, model lab, and sovereign-cloud startup on Europe's wish list shares the same unglamorous prerequisite: it has to be funded, and the investor who would fund it has to find it first.

That is the layer Floancer is building. Not the fund, not the factory. The rail that lets European capital find European builders across borders, investor-led so the channel carries signal instead of spam. The public packages are working to create the demand and de-risk the bets. The private capital already exists. What is missing is the connective tissue between them, and that is the part we want to help build.

We did not start Floancer because of one export-control letter. But that letter made the argument better than we ever could. Europe's ability to fund its own future cannot depend on capital that stops at the border, and we intend to be part of fixing that.

The bill is coming either way

Europe just watched its access to frontier capability disappear in an afternoon, decided by people it did not elect. The lesson is not that America is hostile. The lesson is that dependence is a position, and positions can change without your consent.

The transition to a more self-reliant European technology base is going to be expensive, and it will be funded by Europeans or it will not happen. The money is here. It needs to find the builders, and it needs to stop pretending the other 26 markets are foreign.

Floancer is building the cross-border discovery layer for European startups and investors, so the capital Europe already has can find the companies it needs to back. See how it works for investors or for startups.

Sources

• Al Jazeera, US orders Anthropic to disable AI models for all foreign nationals, June 2026. aljazeera.com
• Euronews, US export controls on Anthropic "should not be discriminatory," EU Commission warns, June 2026. euronews.com
• European Commission, Strengthening Europe's tech sovereignty, 3 June 2026. commission.europa.eu
• Fortune, Anthropic shutdown sparks a global scramble for sovereign AI, June 2026. fortune.com
• CEPR, The Venture Capital Challenge for Europe. cepr.org
• European Parliament, Report on the Draghi Report / Capital Markets Union, 2025. europarl.europa.eu