Media over QUIC and the Fediverse: Towards a Federated Live Streaming Network

The Fediverse has proven that social networks do not have to be centralized. Millions of users communicate across independent servers without surrendering their communities or data to a single provider. ActivityPub enables posts, comments, and media to be shared between thousands of autonomous instances.

Yet while posts, images, and videos have been successfully federated, live streaming remains largely centralized. A handful of platforms control the infrastructure, the costs, and ultimately access to audiences.

Media over QUIC (MoQ) has the potential to change that.

The Fediverse is built around a simple principle: communities should remain independent.

Administrators run their servers primarily for their own users. Their goal is to serve their communities, not the entire world. This is why the Fediverse replicates content across instances. Users do not have to move to a centralized platform to participate in conversations. Instead, the content comes to the communities.

In other words, in the Fediverse, content moves closer to users rather than forcing users to move to the content.

This model has already proven itself for static or slowly changing content. Posts, images, and videos are replicated, cached, and served locally by participating instances.

Why shouldn’t the same principle apply to live media?

Media over QUIC is a new protocol designed for the distribution of real-time media. Instead of connecting every viewer directly to the origin server, streams can be forwarded through a network of relays.

A stream only needs to be transmitted once. From there, it can be replicated throughout the network. Each node can receive the stream and redistribute it locally to its own users.

This represents a fundamental shift.

Rather than having every viewer connect to a single service, live media itself moves through a network of independent nodes—much like posts and messages already move through the Fediverse via ActivityPub.

In many ways, Media over QUIC could become for live media what ActivityPub has become for social content.

If Fediverse servers also act as MoQ Relay-Server, something new emerges: a federated content delivery network.

Each instance remains independent and continues to serve its own community. At the same time, it can subscribe to streams originating from other instances and redistribute them locally.

The origin server no longer needs to deliver a stream to thousands or even millions of viewers directly. Instead, it only serves participating relays, which in turn serve their local communities.

As the network grows, so does its distribution capacity.

Scalability emerges not despite decentralization, but because of it.

One of the biggest advantages of this model is cost control.

Today, a successful livestream often means increasing infrastructure costs. If a stream goes viral, the origin server suddenly becomes responsible for viewers around the world.

In a federated model, each instance only pays for its own users.

An instance with 20 viewers pays for 20 viewers.

An instance with 500 viewers pays for 500 viewers.

No administrator is expected to finance bandwidth for the entire Internet. Costs remain local, predictable, and under control.

This makes live streaming economically viable even for small communities and independent operators.

Not every server needs to play the same role.

Some instances may choose to serve only their own communities. Others may decide to contribute additional relay capacity and become part of the wider distribution infrastructure.

This opens the door to entirely new business models:

Decentralization does not eliminate commercial opportunities. On the contrary, it enables a diverse ecosystem where private, non-profit, and commercial participants can coexist.

A federated live streaming network does not replace traditional CDNs.

Instead, Media over QUIC enables hybrid architectures.

A stream could simultaneously be distributed through Fediverse relays, regional hosting providers, and commercial CDN operators. Communities and service providers can decide how much infrastructure they want to contribute and where they want to rely on external services.

Decentralization and professional infrastructure are not competing ideas. They complement each other.

The true promise of Media over QUIC is not merely technical.

It is about ownership.

Today, live streaming depends on a small number of centralized platforms. Distribution infrastructure and audience reach are controlled by a few companies.

A federated live streaming network could break this dependency.

Thousands of independent communities could collectively build infrastructure that belongs to no single entity while still scaling globally.

Just as ActivityPub decentralized social networking, Media over QUIC could decentralize the distribution of live media.

Perhaps the future of live streaming will not be a single global platform.

Perhaps it will consist of thousands of independent communities sharing infrastructure, controlling their own costs, and together building an open network for live media.

A network where users no longer have to come to the platform.

Instead, the media comes to the communities.