One of the consequences of the ability of artificial intelligence (AI) systems to reason, plan, and execute tasks autonomously is that mobile traffic patterns have changed significantly, increasing the importance of uplinks.
InterDigital research shows how the emergence of agent AI is redefining the demands placed on devices, networks, and cloud infrastructure. Among the survey results: Migrating a decentralized network to enable on-device AI According to a report conducted by ABI Research for InterDigital, the rapid adoption of agent systems – expected to increase across enterprise and consumer markets over the next three years – is increasing uplink traffic from AI devices and changing the way modern networks operate. As a result, network design will be reconsidered.
The study notes that modern mobile networks have historically been optimized for downlink throughput and video delivery. However, unlike traditional mobile applications that primarily consume data via downlinks, agent AI systems continuously generate and exchange contextual information, enabling real-time reasoning and decision-making. Therefore, as the amount of upstream data generated by AI devices increases, there is a risk that networks will become overloaded, leading to increased delays and costs.
The study found that four key devices are driving uplink traffic: smart glasses, wearables, smartphones, and IoT sensors and devices. Smart glasses continuously capture video, images, and environmental context and send the data upstream for real-time AI inference and assistance. ABI Research predicts that by 2030, smart glasses shipments will reach 70 million units, with cellular-enabled devices accounting for more than 12% of shipments.
In contrast, wearable devices (including next-generation technologies that collect voice, biometrics, and contextual signals) support persistent agent-based AI interactions. Smartphones increasingly send multimodal inputs such as voice, photos, videos, and sensor data to cloud and edge AI systems. During operation, IoT sensors and devices continuously stream operational and environmental data to AI models for analysis, automation, and decision-making.
The study also found that uplink pressure is already noticeable in video-intensive applications such as live streaming and real-time video collaboration, and that many users uploading simultaneously can cause localized mobile cell congestion. Unlike these temporary spikes, agent AI systems generate continuous upstream data exchanges from connected devices, which can put sustained pressure on uplink capacity, it added.
The report suggests that the industry needs to move to a distributed intelligence architecture to meet the AI demands of modern devices. In a distributed intelligence architecture, AI workloads are coordinated across on-device processors and cloud platforms based on their complexity. The company says that by embedding intelligence deep into the network infrastructure, AI-enabled applications can operate efficiently without sacrificing performance.
The study found that as the entire mobile ecosystem rapidly innovates and integrates the latest AI technologies, ensuring consistent and complementary directions of movement is essential to enabling future AI applications and related experiences.
This is especially true for 6G networks. 6G networks are designed to improve mobile broadband (MBB) access for smartphones by increasing network speeds, reducing latency, and improving device battery life.
However, InterDigital cautioned that this is only the foundation on which additional services can be built. By integrating AI into the network, smartphones can offload demanding applications to the edge of the network or to a centrally managed location to ensure optimal resource utilization and enable a distributed intelligence fabric.
“Agentic AI introduces a new set of requirements for both networks and devices,” said Larbi Belkhit and Paul Schell, senior analysts at ABI Research and co-authors of the report. “Supporting autonomous AI systems will require far more distributed computing architectures and significantly more intelligent networks. Operators will need to manage increasingly symmetrical traffic patterns while enabling real-time AI workloads across devices, the edge, and the cloud.”
“Agentic AI represents the next stage in the evolution of intelligent connectivity,” said Rajesh Pankaj, Chief Technology Officer, InterDigital. “Intelligence must be distributed across devices, networks, and clouds, and efficiently delivering these AI-enhanced services will require new computing architectures that balance performance, latency, and energy efficiency.”
