Insights from the World Artificial Intelligence Cannes Festival 2026

Agentic AI remains a major topic, while Open Source AI and World Models are gaining significant traction.

The maturity of Agentic AI is growing, and discussions covered all associated aspects: from platforms and software to security to the new roles humans will play in organizations leveraging AI Agents. One challenge is the growing complexity of systems and architectures. Another is maintaining agent orchestration under strict human supervision, which will become increasingly challenging as the number of AI agents grows. Managing autonomy at scale will be a critical challenge.

Regarding the impact of Agentic AI on organizations, it is valuable to reflect on the future enterprise operating model - especially the role of leaders in a future where complex AI agents collaborate with humans to orchestrate highly specialized AI agents.

Cognitive World Models (based on Yann LeCun's work on VL-JEPA - Joint Embedding Predictive Architecture for Vision-Language) are now emerging and go beyond LLMs (this marks the end of "just add GPUs," according to Pascale Fung). Cognitive world modeling will enable AI to better plan, explain, adapt, and collaborate.

At the end of the day, core foundational AI disciplines (AI Governance, AI Literacy, AI Supervision, Trustworthy AI, Data Management) remain essential—even more so for organizations deploying Agentic AI.

Many thanks for the inspiring presentations and bilateral discussions, especially with Dr. Luc JULIA, Eric Salobir and Milind . See you all there next year!