Global network traffic report 2026

Understanding traffic trends is key to planning networks in the AI era

Advanced connectivity and the AI supercycle are inextricably linked. As AI changes the nature of global traffic, our networks must evolve to meet these demands. Nokia’s annual Global Network Traffic Report projects network traffic trends over a 10-year period to help communication service providers and AI and cloud providers plan their networks and infrastructure investments.

Highlights from our latest report

AI is transforming the networks that connect our world. What started with consumer chatbots and generative AI tools is expanding into a vast ecosystem of autonomous agents, robots, vehicles and machines — each generating and consuming data in new ways. Our new report projects global traffic will grow nearly fourfold by 2035, with almost half of it driven by AI. But AI's impact goes beyond sheer volume. It is changing how traffic behaves, where it flows and which direction it travels.   

Global network traffic report 2026

Download now

Key statistics/facts: 

44%

The portion of network traffic driven by AI in 2035

36%

Amount of AI-native traffic served by the edge and metro core in 2035  

160x

Agentic AI traffic growth over 10 years

105x

Physical AI traffic growth over 10 years

AI is reshaping global network traffic

AI is not just adding more traffic to the world's networks — it is fundamentally changing the nature of that traffic. A future network connecting autonomous software agents acting on our behalf and robots sensing and interacting with the physical world is also a network exhibiting very different traffic behavior. This expanding AI ecosystem will not only demand more capacity, but also low latency, uplink performance and distributed intelligence.

AI will be the key engine of traffic growth

By 2035, we project global traffic will quadruple to 3,789 EB/month, and nearly half of it will be driven by AI. Large-scale consumer and enterprise AI adoption, media-rich AI sessions, and the rise of agentic and physical AI will fuel this growth.

Agentic and physical AI will join GenAI as major traffic drivers

Today, agentic and physical AI contribute only a small share of total traffic. Over the next decade, however, they will become two of the fastest-growing application categories, growing to 160x and 105x of their 2026 levels, respectively.

AI processing will begin to shift from the central cloud to the edge

We see clear evidence of a distributed AI infrastructure, commonly referred to as the AI grid, emerging. More-efficient models and the need for lower latency and data sovereignty are pushing AI inference and intelligence closer to the network edge. By 2035, we project the edge and metro core will service 36% of all AI-native traffic.

Uplink performance will matter more than ever

AI is challenging the asymmetric design of traditional access networks. As applications increasingly sense their surroundings and stream raw data to the cloud, both consumer and enterprise AI are becoming more uplink intensive. In cells with the highest AI usage, uplink will account for 24% of total access traffic, and for physical AI, more than 90% of all traffic originates in the uplink.

"Cellular" and "mobile" will no longer mean the same thing

Fixed wireless access (FWA) will account for more than one third of consumer cellular traffic in 2035, turning cellular operators into home broadband providers. Higher data consumption per FWA connection and a growing global installed base are fueling this shift.

Explore our 2026 traffic projections in more detail

Download our Global network traffic report 2026

Frequently asked questions

How will AI impact global network traffic? 

AI is becoming the key engine of network traffic growth. By 2035, global traffic will quadruple to nearly 3789 EB/month, and about half of it will be driven by AI, according to the 2026 Nokia Global Network Traffic Report. Large-scale consumer and enterprise adoption, media-rich AI sessions and the rise of agentic and physical AI are fueling this growth. AI also behaves differently from traditional applications, producing bursty traffic, demanding continuous connectivity or generating bulk data uploads depending on the service.

What are agentic AI and physical AI, and why do they matter for networks?

Agentic AI refers to autonomous software agents that act on our behalf, while physical AI powers robots, vehicles and machines that sense and interact with the physical world. Their share of total traffic is small today, but the 2026 Nokia Global Network Traffic Report projects they will be two of the fastest-growing application categories over the next decade, growing to 160X and 105X of their 2026 levels, respectively. Their unique network requirements will push AI traffic further from the norms that shaped today's networks.

What is a distributed AI infrastructure?

A distributed AI infrastructure, or AI grid, will emerge as AI traffic processing migrates from the central cloud toward the network edge. More-efficient models and the need for lower latency and data sovereignty are moving AI inference and intelligence closer to users and devices. The 2026 Nokia Global Network Traffic Report projects that by 2035 the edge and metro core will service 36% of all AI-native traffic.

Why does AI require more uplink capacity from networks?

Traditional access networks were designed to carry far more downlink than uplink traffic. AI is challenging that model as consumer and enterprise applications increasingly sense their surroundings and stream raw data into the cloud for analysis. In cells where AI usage is highest, uplink will account for 24% of total access traffic, according to the 2026 Nokia Global Network Traffic Report. The trend is most acute for physical AI, where more than 90% of all traffic originates in the uplink.