Omdia view
Summary
In the past five years (2020–25), global cellular traffic growth has moderated. In the lead-up to 2020 there was a rush to acquire 5G spectrum, build up infrastructure, and launch commercial services. Then the COVID-19 pandemic happened. Commuting and travel stopped and populations became isolated. Home broadband use spiked as mobile cellular growth slowed. Innovative wireless B2B applications were part of the original 5G plan. These applications hold great promise but still need to gain momentum.
Looking back, Omdia expects the mobile industry will remember 2020–24 as a temporary slowdown for cellular. In 2025, people are back on the road and in the air. There are widespread return-to-office initiatives, and travel has passed pre-pandemic levels. The artificial intelligence (AI)/machine learning (ML) revolution has started, creating demand for remote sensors and devices to feed its applications. Mass market wearables applicable for consumer entertainment and business are nearing a tipping point. The end result: in the run-up to 2030, traffic is ramping up. In the coming years, Omdia expects 5G networks in Europe and globally to shift from underutilized to overloaded.
2025–33: Europe’s forecast for cellular traffic growth
For 2025–33, Omdia expects Europe to experience a CAGR of 27% for cellular traffic (see Figure 1). At a broad level, this aligns with global growth trends. However, Europe stands out due to its wealth, advanced technological infrastructure, and strong wireless footprint, meaning that it will embrace cellular AI applications ahead of other regions.
Figure 1: Total cellular traffic growth, Europe and global, 2023–33
Source: Omdia
Omdia projects a healthy 55% CAGR for cellular AI traffic in Europe over the next eight years, as the region transitions away from conventional applications (see Figure 2 – definitions and examples of net new AI, AI enhanced, and conventional traffic types are in the Appendix). The result is that Europe will reach two key traffic thresholds within the next eight years:
- Around 2031–32, Europe will reach a tipping point where cellular networks will carry more AI application traffic than non-AI traffic.
- In 2032, Europe’s net new AI traffic will exceed 10% of total cellular traffic. Net new AI traffic represents applications enabled by the rise of AI. There will be more net new AI traffic on Europe’s cellular networks in 2033 than all the cellular traffic these operators carried in 2025.
Europe’s transition to AI-enabled applications – and the region’s demographic stagnation – has put the brakes on conventional cellular traffic growth. Omdia projects an anemic 14% CAGR for conventional cellular traffic in Europe over the next eight years, compared to 19% CAGR for the rest of the world.
Figure 2: Cellular traffic by category (AI vs. non-AI), Europe, 2023–33
Source: Omdia
Cellular traffic growth driven by analytics and experiences
Video content is cellular’s foundation. Well over half of cellular traffic today is media streaming. Growth for this segment is slowing, as there is only so much time in the day and only so many pairs of eyeballs. However, video quality is improving and that takes more bandwidth – streams are inflating from SD (1–3Mbps) to HD (3–8Mbps).
Upstream video is now complementing downstream content for analytics, with cameras serving as universal sensors. A real-time video stream acts as a “single source of truth,” pairing with AI-powered recognition.
From counting people, reading license plates, and detecting hazards (such as leaks, smoke, and fires), to safety, theft surveillance, and asset tracking, AI-powered video can be deployed cost-effectively and unlock vast efficiency gains (see Figure 3).
Upstream video over cellular serves two major demand segments. The first includes verticals with remote needs, such as transportation, agriculture, and energy, as well as cellular fixed wireless access (FWA). The second includes mobile workforces across industries, such as field sales and technicians, retail delivery services, and public safety.
As upstream video grows, future media and extended reality (XR) devices and formats will ramp up in the longer term. These emerging devices offer many new types of experiences. They are more online and interactive, consume more bandwidth, and have tighter performance requirements than SD/HD video.
Figure 3: Cellular net new AI traffic growth by category, Europe, 2023–33
Source: Omdia
Global crunch: Europe’s cellular growth in context
Projected cellular traffic increases are not unique to Europe. Asia & Oceania remains by far the largest market, fed by growth across greater China, Japan, Korea, Southeast Asia, and Australia/Oceania. The Americas and the Middle East & Africa are projected to be the fastest-growing regions for cellular traffic over the next eight years. Europe falls somewhere in between, holding a share of about 18% of the world’s cellular traffic over the next eight years (see Figure 4).
Figure 4: Cellular traffic growth, Europe and global regions, 2023–33
Source: Omdia
What happens if infrastructure fails to keep pace?
Omdia’s methodology makes many assumptions about industries’ ability to keep pace with demand. We generally assume there is enough infrastructure (fiber for backbones and distribution, cell density, and spectrum) and competitively priced services to support business and consumer traffic growth. However, Omdia predicts a temporary slowing of cellular growth during 2029–30. This accounts for a gap when future media and XR devices become widely available, while 5G remains insufficient to support the traffic demands of fully immersive experiences. Omdia expects a future commercial 6G rollout to resolve any barriers in 2031 and beyond.
Technically, access management tools such as network slicing can deal with any contention by prioritizing access. As cellular networks approach overload, critical communications and premium-tier subscribers get enhanced service quality, and others have reduced performance. Best-effort cellular users will need to resort to offload options such as site Wi-Fi when the cellular network does not measure up. Analysis by The World Bank has shown that investment in digital infrastructure helps economic growth. An industry can always find ways to route around peak capacity shortages, but there is an economic cost when the supply of an underlying utility is not fully reliable.
Appendix
Methodology
Omdia’s AI Network Traffic Forecast, first published in 2023, is the product of cross-industry research from Omdia’s analyst groups, including hardware and component shipments, IoT device shipments and connectivity services, the evolution of consumer technology and mobile devices, AI software and services, enterprise applications adoption, network traffic growth trends, data center growth, and our global team covering national/regional wireline and wireless broadband access proliferation. The forecast draws from these analyst teams’ knowledge, their industry surveys, and their market sizing and forecasts. Figure 5 shows some of the Omdia sources that helped contribute to the AI network traffic sizing model.
Figure 5: Omdia sources for AI network traffic forecasting
Source: Omdia
A summary of some of the factors underlying the model’s traffic growth assumptions include:
- Devices: Availability of enabling devices, increasing levels of preprocessing offload, and increasing potential for local AI processing on-device.
- Network and data center services: General availability of sufficient resources end-to-end (capacity, performance, space, power, systems, and storage) at competitive price points; continuous price erosion in service rates by volume, as volumes increase; transfer and storage costs low enough that visual data is the default format.
- AI software and hardware: Continued low price points for essential basic AI functions (e.g., processing of high-capacity feeds such as video/image recognition). AI applications mirror bandwidth profiles of today’s non-AI equivalents (e.g., NetOps to AIOps; visual processing to AI-enabled cognitive analytics). Growth moderates based on development and shipping volumes of more efficient and higher-end GPUs.
- Enablement: Proliferation of AI-enabled deployment and integration tools to lower integration costs and accelerate B2B transformation. ISVs, PaaS/SaaS, and others add increasing levels of AI capability into existing platforms and software.
- Non-commercial/black hat use: AI will be developed, and traffic will be generated, outside of conventional supplier ecosystems.
To better understand evolving pace of traffic growth, Omdia continues to monitor AI-related influences beyond business, including the legal, political, economic, environmental, and societal/labor implications of AI development.
Traffic definitions
Net new AI traffic applications
Net new AI traffic applications are enabled by, and exist because of, new capabilities unlocked by AI. These applications tasks might be simple, but implementing them with conventional application development and integration would have been prohibitively expensive. GenAI can write copy, generate images, and can act as a natural-language interface that supplies and organizes information. B2B AI examples include cameras installed in a warehouse to monitor and track inventory levels on shelves, and video for surveillance, security, and safety. Figure 6 provides examples of two twists on standard video surveillance that could provide unique industry value.
Figure 6: Applications ideas to leverage cellular and generate net new AI traffic
Source: Omdia (images sourced from Wikimedia Commons)
AI-enhanced traffic applications
AI-enhanced traffic applications existed prior to AI. These applications add AI features for new functionality and better experiences. Over time, AI features will take over the conventional application. Omdia divides AI enhanced traffic into AI and non-AI portions. Examples of AI enhanced applications include: development tools that enable low-code environments; image recognition and manipulation tools; e-commerce and media site recommenders based on users’ past preferences; and management systems that add AIOps. Figure 7 provides examples of two common applications enhanced with AI features.
Figure 7: Examples of enhanced AI – conventional applications adding AI functionality
Source: Omdia
Conventional traffic applications
Conventional traffic applications are not driven by AI, and do not incorporate AI in processing and interactions. This application type is included to provide a complete view of all categories of network traffic.
Further reading
AI Network Traffic Forecast: 2023–33 (January 2025)
AI Network Traffic Report Forecast: 2022–31 – Analysis Update (March 2025)
Author
Brian Washburn, Service Provider Enterprise & Wholesale
