At the 10th edition of the India Mobile Congress (IMC) 2026, held at Yashobhoomi, Qualcomm showcased demonstrations focused on the convergence of connectivity and computing to support agentic artificial intelligence. The centerpiece to the company’s presentation is its “Connectivity Plus Compute” framework, designed to distribute AI workloads across user devices, edge infrastructure, and cloud data centers via high-performance networks.
Historically designed primarily to route data, telecommunications infrastructure is shifting as autonomous, multi-step AI agents increase overall computational demand. Under this model, telecom operators are expected to function alongside cloud providers and hyperscalers as active nodes in a shared computing ecosystem, balancing latency, cost, and device constraints.
Device-Level AI and Inference Disaggregation
At the consumer interface layer, Qualcomm highlighted how form factors including smartphones, smart glasses, earbuds, and wearable pendants are serving as front-end interaction points for AI agents. Central to the device-level demonstrations is the Snapdragon 8 Elite Extreme Gen 6 platform, which the company stated can execute Mixture-of-Experts (MoE) models of up to 30 billion parameters directly on a smartphone without continuous cloud offloading.
Complementing on-device processing, Qualcomm demonstrated what it terms the “disaggregation of inference.” Under this workload-partitioning strategy, sensitive or latency-critical reasoning tasks execute locally on hardware, while heavier reasoning steps route dynamically to edge servers or centralized data centers. Alongside this, the company presented high-efficiency silicon solutions tailored specifically for data center inference and agentic pipelines.
AI Integration in Radio Access Networks and Sensing
Looking toward future network infrastructure, the showcase outlined the role of AI in 6G architectures, which are being developed to be inherently AI-native:
- Autonomous Network Management: Qualcomm demonstrated Agentic RAN Management within its Dragonwing RAN Automation Suite. The system interprets operator-level intent, evaluates real-time traffic and network variables, and executes adjustments while retaining human oversight mechanisms.
- Integrated Sensing and Communication (ISAC): Developed in collaboration with Tejas Networks, a live demonstration showed a commercial 5G base station performing dual functions: standard cellular communications alongside radar-like object detection and spatial tracking. The technology is aimed at municipal traffic monitoring, logistics tracking, and drone oversight, serving as a functional precursor to native sensing frameworks planned for 6G standards.
Industrial Applications and Automotive Deployments
In enterprise and industrial settings, on-premise compute was exhibited through an autonomous mobile robot making real-time navigation decisions. In partnership with digital engineering firm Nagarro, Qualcomm displayed an edge-based predictive maintenance solution running on the Qualcomm Dragonwing IQ9 Series. The system ingests and evaluates live telemetry from industrial machinery to detect operational anomalies and forecast equipment failures prior to downtime. A commercial smart mirror deployment was also shown to demonstrate contextual edge computing in consumer retail environments.
In the automotive sector, the showcase featured the Mahindra BE 6 SPORTEQ electric vehicle, integrated with the Snapdragon Cockpit Platform. The system powers Mahindra’s Artificial Intelligence Architecture (MAIA), handling natural language processing and contextual cabin adjustments tailored for the domestic market.
Qualcomm’s demonstrations are scheduled to run through October 10 at Booth A10, Hall 2, Yashobhoomi.
Commenting on the showcase, Rajen Vagadia, President, Qualcomm India, said:
6G is being designed to be AI-native, with networks that integrate connectivity and compute. The adoption of agentic AI will significantly increase demand for both. Trust, security, and affordability will be fundamental to scaling agentic AI and meeting them will require workloads to be distributed intelligently across endpoints, devices, the network, and the cloud. Our showcase at IMC offers a glimpse of this transformation, including AI agents running across devices.