Building an AI-native Future of Industry
Beijing, China , 21 September 2026
Contact:
tsbevents@itu.int
The workshop programme is organized as four serial sessions.
09:30 - 10:00
| Welcome and Opening
|
10:00 - 10:30
| Session 1: Fundamental Technologies for AI-Native Architectures
|
| The transition to AI-native architectures demands a new set of foundational technologies—where intelligence is not overlaid onto existing systems but structurally embedded from the ground up. This session brings together the FG-AINN community to present progress on architectural specifications, use cases, and foundational requirements, and to discuss the key technical pillars: structural frameworks for AI-native networks, deep integration of compute and communication resources, and green computing models that reconcile growing AI demands with energy efficiency.
|
10:30 - 12:30
| Session 2: AI-Native Networks: Architecting the Future of Intelligent Telecommunications
|
| AI-native networks are at the heart of this workshop—the architectural foundation upon which the AI-native future is built. This session explores the next generation of telecommunications by addressing the design, standardization, and deployment requirements of AI-native approaches. Drawing upon the active work items of FG-AINN, it brings together telecom operators, AI infrastructure developers, vendors, and academic researchers to discuss structural frameworks, compute-communication resource allocation, and green computing models. The session aims to bridge international research with standardized implementation pathways, facilitating open dialogue on the governance and deployment of intelligent IMT-2030 network elements.
|
12:30 - 14:00
| Lunch break
|
14:00 - 15:30
| Session 3: AI Agents in Enterprise Applications
|
| AI agents are transitioning from conversational assistants to autonomous task-executing systems—digital workers capable of perceiving environments, decomposing goals, invoking tools, and delivering outcomes end-to-end. This session will examine the enterprise adoption of AI agents, focusing on the shift from single-agent to multi-agent collaboration, the role of emerging interoperability protocols, and the practical challenges of deploying agents at scale. The relationship between AI agents and AI-native networks is bidirectional: agents are increasingly deployed to operate and optimize network infrastructure itself, while the scaling of enterprise agent deployments places new demands on intelligent network provisioning.
|
15:30 - 17:00
| Session 4: Trends in the Token Economyn
|
| Tokens have become the fundamental unit of AI computation—the kilowatt-hour of the AI era. This session examines the token economy as the value circulation layer of the AI ecosystem through an industrial-chain lens: the compute layer, where token-based pricing is shifting infrastructure from capacity provisioning to on-demand token delivery; the model layer, where token efficiency drives competition, with unit costs collapsing even as aggregate demand surges; and the application layer, where autonomous agents amplify token consumption per task by orders of magnitude. The explosive growth of token demand is driving compute and network from loose coupling toward deep integration, requiring networks to evolve from traffic pipes into value hubs. AI-native networks ultimately serve the AI economy measured in tokens; understanding these dynamics is essential to the ecosystem this workshop addresses.
|
17:00 - 17:20
| Closing Remarks
|