The emergence of agentic artificial intelligence (AI) will redefine demands of devices, networks, and cloud infrastructure, according to an analysis from InterDigital and ABI Research. These changes, the companies said, could create upstream bottlenecks in mobile networks.

The analysis, entitled “The Distributed Network Shift Enabling AI on Device,” said that, while the dynamic to date has focused on delivery of information via downlink, the agentic dynamic will feature generation and exchange of contextual information to enable real-time reasoning and decision making.

In other words, there will be more data shared and it will be done so in an increasingly symmetrical manner.

“Agentic AI marks the next phase in the evolution of intelligent connectivity,” Rajesh Pankaj, InterDigital’s Chief Technology Officer, said in a press release.

“As AI systems become capable of reasoning, planning, and executing tasks autonomously, we are beginning to reimagine our wireless networks for 6G. Intelligence must be distributed across devices, networks, and the cloud, and delivering these AI-enhanced services efficiently will require a new computing architecture that balances performance, latency, and energy efficiency.”

The analysis on agentic AI and mobile traffic identifies four main drivers of the increasing and more symmetrical traffic:

• Smart glasses that send data upstream for real-time AI inference and assistance. ABI Research predicts 70 million smart glasses shipments by 2030 and that cellular-enabled devices will represent more than 12% of shipments.
• Wearables that include next-generation devices that collect voice, biometric and contextual signals to support persistent agentic AI interactions.
• Smartphones that transmit multimodal inputs such as voice, photos, video, and sensor data to cloud and edge AI systems.
• IoT sensors and devices that continuously stream operational or environmental data to AI models for analysis, automation, and decision-making.

The agentic AI analysis says that the stresses are already apparent in such applications as livestreaming and real-time video applications. The answer is to evolve to distributed intelligence architectures including cloud platforms and scenarios in which AI workloads use multiple on-device processors.