Pointing out that Social media video has become a dominant form of communication is not going to win me an “Insight of the Year” award. Whether it’s a direct video message or a social post, we press record and move on without thinking twice.
However, behind the simplicity of “record and send” lies a structural compromise that most platforms have quietly accepted.
When a user sends a video from their phone today, they create a single encoded file. In some cases that file must work everywhere, across a fragmented mix of devices, operating systems, and unpredictable network conditions. At minimum it provides the source and therefore maximum quality for any processing downstream of the phone.
In that split second a number of trade-offs are made. What upload bandwidth is available? How large might the audience be? What sort of screen will this end up on? How time-critical is the content? All of it decided once. Up front.
In practice, this often means sending 360p or 480p video at modest bitrates to guarantee fast upload and reliable delivery. Once that file is created, that is it. There is no easy way to improve it later.
The result is a permanent trade-off between visual quality, compatibility, and speed.
What if capture did not mean compromise?
Conversations I’ve had recently suggest there is another way to think about this.
While messaging is the most visible example, the same single-file assumption quietly shapes feed delivery, live workflows, AI pipelines, and storage architectures across the platform.
Instead of producing a single fixed-quality file, video can be structured in layers. A universal base stream guarantees compatibility and fast delivery. An additional enhancement layer allows higher resolution and improved quality for devices and applications that can actually make use of it.
Rather than generating one file for everyone, the device produces something much more flexible. The base guarantees speed and reach. The enhancement layer dynamically ensures an upgraded experience wherever possible.
Think streams that start instantly, and seamlessly switch to a higher resolution as soon as bandwidth permits.
For messaging, this translates into a more recipient-aware experience. Devices that cannot decode or display higher resolution simply receive the baseline. More capable devices reconstruct a higher-quality version as network conditions allow, without affecting anyone else.
Compatibility is preserved. Nothing breaks. Existing behavior stays intact. Quality just scales where it can.
Technologies such as MPEG-5 LCEVC enable this layered approach. They can be rolled out with an app update, working alongside existing “base” codecs rather than replacing them.
Beyond messaging: cloud and AI implications
Social platforms process enormous volumes of user-generated content. Only a small fraction of uploads attract the majority of views, yet every clip still must be stored, processed, and delivered across multiple resolutions and formats.
Introducing scalability at the source, sending a base layer with optional enhancement, changes that equation.
Platforms gain an additional lever to improve bitrate efficiency without creating parallel encodes, while still maintaining a reliable fallback for all users.
The same layered structure creates flexibility across both machine analysis and human consumption. Lighter representations can support many inference tasks and faster processing, while higher fidelity remains available wherever visual quality or precision makes a difference.
This is just one example of how sending two quality levels from the starts begins to unlock smarter ways to manage storage, compute and delivery costs at scale.
A broader architectural question
The real opportunity may lie in combining these ideas.
What happens if capture itself becomes scalable?
How should cloud workflows change if layered source material arrives directly from the device?
Could messaging, feed delivery, and AI processing share a more unified asset structure?
Social platforms are already balancing rising expectations for quality against upload speed, infrastructure cost, and device fragmentation. Those tensions are not going away.
The real shift may be rethinking the starting point. Does mobile video still need to begin life as a single fixed file?
In my role leading the Professional Services team at V-Nova I’m hearing more discussions about ways to innovate and improve the sophistication in how apps are sending video to deliver against increasingly demanding QoE and cost KPIs. If you are navigating these trade-offs, we would be happy to help you A/B test some of these latest approaches.
Sam Orton-Jay leads a professional services function supporting 3rd parties rolling out the V-Nova family of technologies across a broad range of video services and solutions. With over 20 years of experience in video tech, Sam has led the creation of industry-leading products in video compression and streaming from conception to commercial success.
