Infrastructure vs Integration: What Central Asia's AI Race Tells European Policymakers

Kazakhstan's Play: Infrastructure and Visibility

Kazakhstan has made a full-throated commitment to AI ambition. Alem.ai, the International AI Centre operating from Astana, is the nation's flagship. The facility houses a supercomputing cluster built on NVIDIA H200 processors, powerful enough to rank 86th globally on the TOP500 list. The government has committed to training 1,000 AI specialists annually, with a broader target of 100 domestic AI startups. A domestically developed multilingual model, KazLLM, now powers local applications. All of this sits within Kazakhstan's "2026 Year of Digitalisation", a sustained push to drive digital change across every sector of the economy.

By 2025, the Kazakhstani state had already deployed 54 million eGovernment services. IT exports reached approximately USD 1 billion annually, supported by a workforce of 200,000 digital workers, including 20,000 AI specialists. These are not trivial numbers for a country of 19 million people.

The centrepiece of Kazakhstan's international ambitions is GITEX Central Asia, arriving in Astana on 02/06/2026 for its inaugural edition. Organisers expect up to 1,000 visitors, modest by the standards of Hannover Messe or VivaTech, but a clear signal that Astana intends to position itself as a regional tech showcase. Meanwhile, Astana's own Smart City project is deploying AI analytics systems through local suppliers, with 60 per cent of contracts awarded to domestic firms. Presight Kazakhstan is channelling technology transfer into the local ecosystem.

Uzbekistan's Approach: Quiet Depth Over Flashy Scale

Uzbekistan is demonstrating that operational depth sometimes matters far more than hardware headlines. The nation's digital ambitions are anchored by MyID, a biometric identity platform that has already onboarded 14.5 million users and processed 130 million authorisations. MyID connects 28 banks and 17 payment systems, making it load-bearing infrastructure for the country's digital economy. In February 2026, Uzbekistan introduced OneID auto-registration to reduce friction for new users. More significantly, mandatory biometric SIM verification came into force in January 2026, a sweeping move that makes MyID not optional but foundational.

This is not flashy; it is penetrating. While Kazakhstan invests in AI centres and supercomputers, Uzbekistan is making digital governance so seamless that citizens barely register the technology underneath it. The IT Park in Tashkent remains a crucial innovation hub, though recent expansion data has been sparse.

The contrast resonates strongly with debates happening right now in Brussels and Westminster. Anu Bradford, professor of law at Columbia and author of Digital Empires, has argued consistently that the EU's regulatory approach risks prioritising governance frameworks over actual deployment, leaving member states with rules but not results. The Uzbekistan model, whatever its civil-liberties complications, is the inverse: deployment first, governance catch-up second.

• 130 million authorisations processed through Uzbekistan's MyID platform
• USD 1 billion in annual IT exports from Kazakhstan (2025)
• 54 million eGovernment services deployed by Kazakhstan in 2025
• 14.5 million MyID users across Uzbekistan
• 28 banks integrated into Uzbekistan's MyID payment infrastructure

Two Models, One Region, and What Europe Recognises

The strategic question is genuine: does it matter more to host prestigious conferences and build world-class AI research facilities, or to embed governance systems so deeply into daily life that the technology becomes invisible infrastructure? European policymakers will recognise this tension immediately because they are living it.

Margrethe Vestager, during her tenure as European Commission Executive Vice President for Digital, repeatedly argued that Europe's strength lies in trustworthy, rights-respecting AI systems rather than raw compute. That framing maps neatly onto Uzbekistan's integration-first model. But critics, including those at Mistral AI, the French frontier model company, have argued just as forcefully that without serious infrastructure investment, Europe's trustworthy AI will simply run on American or Chinese hardware. That maps onto Kazakhstan's infrastructure-first model.

Both nations are acutely conscious of digital sovereignty. KazLLM and MyID are homegrown alternatives to global AI platforms and identity systems. Neither country wants to be merely a consumer of Western technology; both want to be producers. The European parallel is obvious: the EU AI Act, the European Health Data Space, and national large language model programmes from France, Germany, and the Netherlands all reflect the same instinct.

The Comparison Table That Should Interest European Planners

• Flagship AI initiative: Kazakhstan, Alem.ai centre plus NVIDIA supercomputer cluster; Uzbekistan, MyID biometric platform
• Primary strategy: Kazakhstan, infrastructure, R&D, and international visibility; Uzbekistan, governance integration and digital identity
• Key 2026 milestone: Kazakhstan, GITEX Central Asia (June); Uzbekistan, mandatory biometric SIM verification (January)
• Digital workforce: Kazakhstan, 200,000 workers including 20,000 AI specialists; Uzbekistan, not publicly disclosed
• IT exports: Kazakhstan, approximately USD 1 billion (2025); Uzbekistan, not publicly disclosed
• Citizens on digital identity systems: Kazakhstan, 54 million eGovernment service interactions; Uzbekistan, 14.5 million MyID users with 130 million authorisations

What This Means Beyond the Region

If Kazakhstan's infrastructure model prevails, Central Asia becomes an AI R&D hub attracting investment from firms seeking regional offices. If Uzbekistan's integration model prevails, the region becomes a reference point for digital-first governance in emerging economies. But the more interesting question for European observers is what happens when both coexist, because that is exactly what is happening across EU member states right now.

Germany's sovereign cloud ambitions and France's investment in Mistral and national compute capacity resemble Kazakhstan's approach. Estonia's long-standing e-government architecture, which handles everything from tax filing to medical records, resembles Uzbekistan's MyID model. The difference is that Estonia built its system with strong data-protection foundations from the outset; Uzbekistan's mandatory biometric SIM rollout raises questions that neither Tashkent nor international observers have fully answered.

There is one more variable worth watching: execution. Kazakhstan has the infrastructure and the international partnerships; Uzbekistan has the scale and the integration speed. Kazakhstan's supercomputers and AI centres are impressive on paper. But Uzbekistan's ability to deploy systems to 14.5 million users, integrate with the banking sector, and enforce mandatory biometric verification at scale speaks to operational maturity of a different kind. Neither nation has yet hit the kind of stumble that would definitively tilt the race, which means both will continue pushing harder. For European public-sector AI teams, that competitive dynamic, two models forcing each other to deliver faster, may be the most instructive lesson of all.