Across the mobile ecosystem, one reality is no longer up for debate: video dominates the network.

Industry estimates consistently show that 70–80% of mobile traffic is now driven by streaming and social video. While this shift has unlocked new levels of engagement, it has also introduced a structural challenge for operators. More video doesn’t automatically translate into more revenue—but it almost always results in higher delivery costs.

For years, operators attempted to manage this imbalance using network-based traffic management techniques, including deep packet inspection (DPI) and, in some cases, decryption. These approaches aimed to “look inside” video streams and reshape delivery within the network.

That era is now coming to a close.

Growing encryption standards, evolving privacy concerns and expectations, and regulatory pressure have effectively sunset decryption-based video management as a viable long-term strategy. Even where elements of DPI remain, their ability to meaningfully influence modern adaptive video streams has diminished.

At the same time, Video itself has evolved.

From Traffic Growth to Experience Intelligence

Today’s video delivery is governed by application-layer logic such as HTTP Adaptive Video Streaming (HAVS). These systems dynamically adjust video quality in real time, continuously seeking to deliver the highest possible resolution based on perceived network conditions.

The intent is to maximize user experience. The unintended consequence is often over-delivery, and more data consumed than necessary for the device, screen size, or viewing context.

This creates a hidden inefficiency: networks are working harder, and operators are paying more, to deliver experiences that are not meaningfully better for the subscriber.

AI changes this equation by shifting the focus from traffic management to experience intelligence.

AI at the Application Layer: A New Control Point

Rather than attempting to manage video from within the network, applied AI is increasingly being used at the application and device level to influence how video is requested and delivered in the first place.

Operating at this layer, AI enables real-time, context-aware decisions about bitrate, resolution, and delivery patterns—before unnecessary data ever traverses the network.

The result is a more precise and efficient model:

• Improved Customer Experience: Video streams align more closely with device capabilities, real-world conditions, and operator policies; reducing buffering and stabilizing playback.
• Preserved Quality, Reduced Waste: Subscribers continue to enjoy high-quality video, faster loading (especially with live video programming) without excess data that provides no perceptual benefit.
• Lower Delivery Costs: Eliminating unnecessary data flow directly reduces network strain and, for many operators, wholesale data expense.

This is not about restricting usage or degrading service. It is about precision—ensuring that every bit delivered contributes to a better experience.

Subscriber Behavior Intelligence: Turning Usage into Insight

As AI optimizes video delivery in real time, it also enables a powerful secondary benefit: the generation of anonymized Subscriber Behavior Intelligence.

By observing how video is consumed— across content platforms, session durations, time of day, and network conditions—operators gain a clearer picture of usage patterns without collecting personally identifiable information

.

When analyzed with AI, this data becomes a strategic asset:

• More effective plan design and targeted promotions
• Faster identification of experience issues and proactive customer care
• Earlier detection of churn risk based on behavioral shifts

This intelligence layer transforms video from a cost center into a source of insight—linking network performance directly to customer engagement and lifetime value.

The Power of Small Gains in a Video-First World

In a network where video represents the majority of traffic, even modest efficiency gains compound quickly.

A 10% reduction in video data usage—achieved without degrading perceived quality—can deliver meaningful improvements in both network performance and operating margins. At scale, this becomes one of the few levers available to operators that simultaneously enhances customer satisfaction and profitability.

AI is uniquely suited to unlock these gains, continuously learning and adapting in real time to optimize outcomes.

A Smarter, More Sustainable Model

As networks evolve toward 5G and beyond, the industry faces a critical challenge: how to support continued growth in video consumption without proportionally increasing cost and complexity.

The answer is not simply more capacity. It is smarter video delivery.

The shift away from legacy approaches like decryption-based traffic management—and toward AI-driven, application-layer control—marks a fundamental turning point. Operators now have the opportunity to align video delivery with both user experience and economic efficiency.

In a video-first world, success will not be defined by how much data a network can carry, but by how intelligently that data is delivered.

For operators, the question is no longer whether video will continue to grow—it’s whether the strategies in place today are designed to manage that growth efficiently or simply absorb it. ---

To contribute content, please contact comms@ccamobile.org. (CCA members only.)