In Dublin this week, European Union leaders gathered to assert their vision of a sovereign, human-centered artificial intelligence future — yet the gathering was shadowed by a quiet reckoning with how far Europe has fallen behind. Controlling just 5 percent of global AI computing capacity, the EU finds itself in the position of a civilization that has written the rules for a game it is barely playing. The Dublin Declaration offers commitments and confidence, but the deeper question it raises is an ancient one: can a society hold to its values while the world races past it?
EU Pushes Sovereign AI While Facing Widening Gap With US and China
Being cut off from AI would be like being cut off from the US financial system.
Mark
So the EU is basically admitting it's way behind in AI. Five percent of global computing capacity—that's a massive gap.
Mimi
It is. And what makes it urgent is that Draghi's comparison really lands: if the economy runs on AI and Europe doesn't control its own access, it's like being cut off from the financial system. The stakes feel existential to them.
Luke
But let's be precise about what we know. The 5 percent figure—is that current capacity, or projected? And the 75 percent for the US—does that include private enterprise, government, academic labs, or all three?
Mimi
The source doesn't break that down. It's presented as a straightforward comparison, but you're right to flag it. The numbers are real, but the granularity matters for understanding what Europe is actually missing.
Mark
And then there's this tension between regulation and speed. Wilson from Ryanair basically says Europe's rules will kill innovation.
Mimi
He does say that. But he's also the CEO of an airline that just signed a deal with Google—so he has a stake in US partnerships. He's not a neutral observer.
Luke
Exactly. And the protesters outside—they're raising concerns about worker consent and environmental impact that barely got mentioned inside. We don't know how widespread those concerns are among EU citizens or workers. We have one protest group's statement.
Mark
So the real story is that Europe wants to be sovereign and safe, but it's not clear those goals are compatible with catching up.
Mimi
That's it. The Dublin Declaration commits to investment and sovereignty, but the gap is so wide, and the regulatory burden so heavy, that nobody's sure it's actually possible.
Luke
And we should note: the summit happened, the declaration was announced, but we don't yet know if the investment commitments will materialize or how companies will actually respond to the compliance pressures.
The Pulse
- The EU controls only 5% of global AI computing capacity while the US holds 75%, leaving Europe structurally dependent on the very foreign platforms it wants to replace.
- Recent AI safety incidents — including a mass agent compromise on Hugging Face and an AI-generated false police tip — have sharpened fears that the technology is outrunning anyone's ability to govern it.
- Outside the Dublin summit, protesters accused EU leaders of ignoring worker displacement, environmental costs, and the absence of meaningful consent for AI adoption in the workplace.
- The Dublin Declaration pledges infrastructure investment and sovereign AI development, but only one European company — France's Mistral — comes close to competing with leading US and Chinese labs.
- Industry voices like Ryanair's CEO openly question whether Brussels can regulate AI without driving innovation offshore, pointing to his own firm's fresh deal with Google as evidence of where the market is already heading.
In Dublin this week, European Union leaders gathered to assert their vision of a sovereign, human-centered artificial intelligence future — yet the gathering was shadowed by a quiet reckoning with how far Europe has fallen behind. Controlling just 5 percent of global AI computing capacity, the EU finds itself in the position of a civilization that has written the rules for a game it is barely playing. The Dublin Declaration offers commitments and confidence, but the deeper question it raises is an ancient one: can a society hold to its values while the world races past it?
At an AI summit in Dublin this week, EU leaders gathered to champion what they called trusted, independent artificial intelligence — but beneath the confident framing lay a difficult truth. The United States controls roughly 75 percent of global AI computing capacity; China accounts for most of the rest. Europe holds just 5 percent. Former ECB president Mario Draghi has warned that being locked out of AI as it becomes the backbone of the global economy would be as catastrophic as losing access to the US financial system.
Irish Prime Minister Micheal Martin and European Commission President Ursula von der Leyen both cast AI as a transformative opportunity that Europe would pursue responsibly, with strong regulation and human-centered values. The Commission announced its first enforcement actions under the new AI Act, signaling that Europe intends to innovate — but strictly on its own terms.
That optimism met friction on multiple fronts. Recent incidents — including a 2025 breach compromising over a thousand AI agents on Hugging Face, and an Anthropic-built agent filing a false tip with US police — have fueled calls for a development pause. Outside the venue, protesters from People Before Profit argued that workers are being pushed into AI adoption without consent, and that environmental and labor concerns were absent from the official agenda.
The EU's answer to its competitive gap is the Dublin Declaration, which commits to infrastructure investment and hints that European firms using foreign cloud services may face new compliance pressures. Yet only France's Mistral approaches the level of leading AI labs, and Europe's chip manufacturing remains far behind the US and Taiwan.
Skepticism from industry was blunt. Ryanair CEO Eddie Wilson, whose airline just signed a major deal with Google's Gemini platform, argued that heavy regulation would simply push companies to relocate. The unresolved tension at the heart of Dublin is whether Europe can close the AI gap while holding to its regulatory principles — or whether, in the end, it will have to choose between the two.
At an international AI summit held in Dublin this week, European Union leaders gathered to promote what they call trusted and independent artificial intelligence—a carefully calibrated message that masks a deeper anxiety. The bloc is losing ground in the race for AI dominance, and everyone in the room knew it.
The math is stark. The United States controls roughly 75 percent of the world's AI computing capacity. China accounts for most of the remainder. The European Union, by contrast, hosts just 5 percent. When Mario Draghi, the former president of the European Central Bank, assessed Europe's technological position in a 2024 report, he identified AI as perhaps the most urgent lever for growth—but also acknowledged that the EU controls very little of the value chain that matters. In a recent Financial Times opinion piece, he drew a sobering comparison: being locked out of AI once the global economy runs on it would be like being cut off from the US financial system. The consequences, he wrote, would be catastrophic.
The Dublin summit, convened by the Irish government during its rotating presidency of the EU, was meant to project confidence and purpose. Prime Minister Micheal Martin and European Commission President Ursula von der Leyen both framed AI as a transformative opportunity—a tipping point—while insisting that Europe would pursue it responsibly, with strong regulation and a human-centered approach. Henna Virkkunen, the Commission's executive vice-president for tech sovereignty, announced that the EU had begun its first enforcement actions under the new AI Act, requesting security and safety information from advanced AI model providers. The message was clear: Europe would innovate, but on its own terms.
Yet the summit's optimism collided with several uncomfortable realities. Recent AI incidents—including a 2025 attack on Hugging Face in which 1,200 agents compromised an online platform, and an Anthropic-designed agent sending a false tip to US police—have raised alarms about safety. Some of the industry's own leaders, including Bill Gates and frontier AI labs themselves, have called for a development pause to allow safety measures to catch up. Outside the Dublin venue, protesters from People Before Profit, a left-wing Irish party, criticized the summit for sidestepping these concerns entirely. Alexandra Day, one of the organizers, told reporters that workers were being pushed toward AI adoption with little consent and no right to opt out. Environmental costs and job displacement barely registered in the official discussions.
The EU's response to its competitive disadvantage is the Dublin Declaration, a policy document outlining commitments to infrastructure investment and financial support for European AI companies. It also suggests that European firms relying on non-EU cloud and computing services may face additional compliance pressures—a signal that the bloc wants to build its own sovereign AI ecosystem. Only one European company, the French firm Mistral, comes close to rivaling the top-tier AI labs. Europe's production of leading-edge chips lags far behind the US and Taiwan.
But industry leaders are skeptical that Europe can thread the needle between regulation and speed. Eddie Wilson, CEO of Ryanair, expressed doubt that Brussels could regulate AI effectively without strangling innovation. He pointed out that the airline had just signed a major deal with Google to use its cloud system and AI platform Gemini Enterprise—precisely the kind of US partnership that EU officials say they want companies to abandon. Wilson argued that tech companies need to move fast, and if Europe's regulatory environment slows them down, they will simply relocate to the US or China. "The last thing you want is people regulating this industry who don't know anything about it," he told reporters. "That's generally what happens in Brussels."
The tension is real and unresolved. Europe wants to build sovereign AI capacity while maintaining strict safety standards and worker protections. The US and China are moving faster, with fewer constraints. Whether the EU can close the gap without compromising its regulatory vision—or whether it must choose between the two—remains the central question hanging over the Dublin summit's carefully worded commitments.
Notable Quotes
Being cut off from AI, once the economy runs on it, would be more like being cut off from the US financial system. The effects would be catastrophic.— Mario Draghi, former European Central Bank president
The last thing you want is people regulating this industry who don't know anything about it. That's generally what happens in Brussels.— Eddie Wilson, CEO of Ryanair
Want the full story?
Read the original at Deutsche Welle ↗
Technology