Livepeer (@Livepeer): Livepeer Q4 2025 network report

Livepeer Q4 2025 network report Our friends at @MessariCrypto are back with a quarterly deep dive into how the Livepeer network is being used. Q4 was a confirmation quarter. Strong, sustained usage on the network reinforced growing demand for real-time AI video. 2025 in numbers Network usage reached 78.7M minutes, up 174% YoY Demand-side fees grew 162% YoY to $191.7k AI inference drives 70%+ of network revenue Over the past year, more apps have started applying AI inference directly to live video via generative effects, avatars and interactive experiences. When AI becomes part of a product’s core workflow, usage doesn’t spike and die down. It stays on. That’s what shows up in the year-over-year growth and the steadier usage pattern in the second half of 2025. @DaydreamLiveAI helped make this visible. By making it easier to build and run real-time AI video workflows under live conditions, it pushed projects past experimentation. As those workflows started running regularly, they exposed real constraints around latency, throughput and coordination. These are now shaping how the network continues to improve. Why Livepeer wins in real-time AI Our infrastructure is designed to process streams continuously under low-latency constraints. Real-time AI video fits naturally into that model. Real-time AI demand also behaves differently from traditional compute. It grows while apps are live and scales with user activity. You can’t pause it or plan it neatly in advance without breaking the experience. Supporting such demand requires infrastructure that can adapt in real time instead of relying on fixed capacity. Livepeer’s distributed GPU network and usage-based execution are built for this pattern. Developers can run real-time AI video applications continuously and expand usage as demand grows, without pre-provisioning or reworking their systems. That alignment is why demand has remained strong as more products move into production. Looking ahead 2025 was about validating real-time AI video demand through real network usage. Q4 showed the network operating at a higher baseline. As outlined in the roadmap, the focus now shifts from driving network performance to accelerating adoption. In 2026, the priority is to build on current demand as more real-time AI video products come to market and move further into production. You can check out the full report here