Overview
Artificial intelligence is entering a new strategic phase. While public debate continues to revolve around large language models, the next generation of economic and geopolitical competition will increasingly be shaped by world models, AI systems capable of learning, simulating, and predicting the dynamics of the physical world. These models are emerging as the foundational layer for robotics, autonomous systems, industrial digital twins, disaster response, and next-generation defense capabilities.
This strategic assessment argues that Türkiye has a stronger strategic fit with world models than with frontier-scale language models. Rather than attempting to match the pre-training race led by the United States and China, Türkiye can build on its existing strengths in defense technologies, autonomous systems, simulation, and advanced manufacturing to develop domain-specific world-model capabilities. Drawing on international developments, comparative experiences from South Korea, the United Arab Emirates, and Singapore, and an assessment of Türkiye’s evolving AI strategy, the report concludes that world models should be treated as sovereign infrastructure and outlines a phased policy roadmap for securing Türkiye’s place in the emerging physical-AI economy.
Key Takeaways
World models are a distinct AI category, not a variant of large language models. They are designed to model environmental dynamics and predict future states, enabling planning and control in ways text-only systems cannot. The field emerged as a competitive domain in 2025 with NVIDIA’s Cosmos, Google DeepMind’s Genie 3, Meta’s V-JEPA 2, and a growing body of academic research establishing world models as the next foundational AI layer.
World models are becoming dual-use strategic infrastructure. As compute inequality widens and the United States and China consolidate their leadership, middle powers including South Korea, the United Arab Emirates, and Saudi Arabia are increasingly pursuing sovereign AI strategies to reduce dependence on foreign AI stacks.
Türkiye’s comparative advantage lies in application-specific world models rather than generic foundation-model competition. Its strengths in defense, autonomous systems, disaster response, urban simulation, and an emerging civilian ecosystem position the country to compete where world models meet the physical economy.
Compute, advanced semiconductors, talent retention, and national data infrastructure remain the principal constraints. Rather than pursuing full-stack AI sovereignty, Türkiye’s most realistic strategy is concentrated investment in the middle layers of the AI stack, where domain-adapted world models are developed and integrated.
Türkiye’s 2026–2030 AI Action Plan substantially expands physical-AI capabilities but does not yet recognize world models as a distinct strategic policy category. The greatest risk is strategic drift, relying on foreign world-model platforms despite possessing the industrial base to develop domestic capabilities.
Türkiye should treat world models as sovereign infrastructure rather than as a research topic. Building targeted capabilities in domain-specific world models, sovereign compute, data infrastructure, procurement, and research centers, while improving coordination across existing institutions, would allow the country to convert its industrial strengths into a durable position within the emerging world-model layer.
