Gen. John Allen

A candid conversation with General John R. Allen and Lieutenant General Ben Hodges on the future of transatlantic security, military readiness and the strategic lessons shaping NATO and European defence.

Watch as Czechia's President, Petr Pavel welcomes guests to the GLOBSEC Forum 2026. 

Followed by a 1 on 1 Chat:
Europe is entering a decisive decade in which technological capability will shape sovereignty, security, and its place in an increasingly contested global order. The choices made in the next five years will determine whether Europe acts as a producer of strategic technologies — or remains dependent on the ecosystems of others. The risks and responsibilities of leadership in the intelligent age are becoming more urgent, as Europe navigates rivalry, partnership, and values in a world where technology has become inseparable from geopolitics.

Executive Summary

In a chat with Gen. John Allen, the Czech President argued that AI mastery is now a decisive geopolitical variable., citing the stated ambitions of both Russia and China as evidence that Europe cannot afford complacency, and called for the removal of barriers between Europe, the US, and with technologically capable partners worldwide. On AI governance, he rejected the framing of a US-EU competition between speed and regulation, arguing instead that the two approaches are complementary and that internal democratic rivalry over models risks benefiting China. He pointed to Ukraine's defence innovation cycle as a concrete benchmark: drones designed, produced, tested at the front and iterated within days represent a tempo that European procurement and legislative systems are currently incapable of matching, and which must change. He closed by framing that the defining geopolitical variable of the coming decade will not be territorial size but speed of innovation.

Livestream Link:  https://youtube.com/live/Ihmif6sAbGQ?feature=share

Europe has no shortage of digital strategies, regulatory frameworks, or political ambition. The harder question is how quickly those ambitions are translating into real adoption, diffusion, and scale—fast enough to sustain competitiveness in a world accelerating toward frontier AI, advanced compute, cloud infrastructure, and data driven growth.

As global powers move decisively to industrialize AI and embed it across their economies, Europe faces a pivotal moment: the question is which levers will accelerate the conversion of governance leadership and policy coherence into technology uptake that drives productivity, resilience, and growth—and how to reduce friction from deployment gaps, market fragmentation, and slow diffusion so Europe can scale delivery.

Progress against the priorities set out in the Draghi Report — and the evidence measuring real world adoption of critical technologies across Europe — are at the centre of the agenda, alongside an evaluation of the effectiveness of policies designed to accelerate scale.

• Where does Europe need to accelerate tech uptake and diffusion in pursuit of the Draghi roadmap?
• What does tech adoption evidence reveal about the most urgent gaps to fix?

Executive Summary

This session brought together senior figures from the European Commission, national government, defence industry and global technology to confront the specific mechanisms behind Europe's AI scaling problem. The central diagnosis was unambiguous: Europe is not losing because it lacks researchers, ideas or ambition, but because fragmented capital markets, slow procurement cycles, regulatory sequencing that precedes rather than follows adoption, and inconsistent public-sector leadership combine to prevent promising innovations from reaching scale. Henna Virkkunen outlined the Commission's forthcoming tech sovereignty package, distinguishing genuine strategic independence from protectionism and identifying AI, quantum, chips and cybersecurity as priority sectors for European capacity-building. Lisa Monaco's Microsoft AI diffusion data reinforced the message that adoption is the decisive variable, and that the countries leading globally (UAE, Singapore, South Korea) are achieving results through public-sector adoption, investment in skills and local language model development rather than frontier research. Micael Johansson offered the defence industry's perspective: AI is being embedded in systems at pace, but the acquisition model must shift from periodic procurement to continuous capability development in genuine partnership with military end users, and the AI Act's burden on industrial applications must be revisited. The session closed on the question of trust, where panellists converged on the view that trust is inseparable from resilience.

Livestream Link:  https://youtube.com/live/zKJDFe24cac?feature=share

Anthropic's Claude Mythos Preview and the launch of Project Glasswing have, in weeks, opened a new chapter in cybersecurity. Mythos is a frontier AI system with unprecedented capability to find and exploit software vulnerabilities — capability that emerged not from cybersecurity training, but as a byproduct of general advances in AI reasoning. Glasswing is the response: structured deployment to twelve trusted partners, including Microsoft and Palo Alto Networks, to give defenders a head start before adversaries catch up. For Europe, this raises questions that cannot wait for the next regulatory cycle. How do European governments, militaries, and critical infrastructure operators participate in a coalition built on American capabilities, governed by an unsettled American policy framework, and arriving faster than European institutions can absorb? What does digital sovereignty mean when the most consequential cyber-defensive tools sit inside private companies and a transatlantic relationship in flux?

• Mythos-class capabilities emerged from general AI advances, not cybersecurity engineering — which means similar surprises are coming faster than European regulatory cycles can absorb. How should European governments and enterprises plan for that?
• Glasswing is, by design, exclusive. European finance ministers have called for broader access. From inside the coalition, what does responsible European participation look like — and what must Europe bring to earn it?
• The Anthropic-Pentagon dispute and the federal court ruling make clear that the governance framework around frontier AI is unsettled in Washington itself. How should European policy principals navigate that uncertainty?
• If a European critical infrastructure operator — a grid, a bank, a hospital network — asked you privately what they should be doing in the next six months that they are not already doing, what would you tell them?

Executive Summary

This session combined a classified-grade briefing on Anthropic's Mythos frontier AI model with practical cybersecurity guidance for European operators, all within a 25-minute GeoTech Centre conversation. Helmut Reisinger's central message was one of controlled urgency: the technology is currently in a controlled testing environment with 12 companies, European institutions and enterprises depending on critical infrastructure have a finite window to prepare, and the preparation required is operational rather than primarily regulatory. His framing of the sovereignty paradox, that good cybersecurity requires global data correlation which sits in tension with the data localisation impulse behind many sovereignty frameworks, was the session's most structurally challenging insight, particularly given the UAE example of physical infrastructure destruction that rendered data location irrelevant. General Allen framed Project Glasswing as the first chapter of a new era rather than a contained experiment: frontier AI emerging from general advances rather than cybersecurity engineering means capability surprises will continue to arrive faster than regulatory cycles can absorb them, and the geopolitical consequences of access asymmetries between the US, Europe and China are already measurable.

Livestream Link:  https://youtube.com/live/S3N5NzUNOpU?feature=share