Dislaimer: The information presented here reflects the Mission and Roadmap of the AI SPE. While it closely aligns with the broader mission of Livepeer Inc. and the wider community, please note that the AI SPE operates as a separate entity, and individual community groups may have distinct missions and priorities.
Explore the glossary below to look up common terms frequently used in this document, related proposals, and other materials.
The AI SPE, founded in January 2024 through this treasury proposal, emerged from growing concerns about increasing centralization in AI, where a few major tech companies control AI research, data access, models, infrastructure, and deployment ecosystems. It is a collaboration between an open-source R&D group—comprising long-term Livepeer contributors and orchestrators—and the ecosystem team.
Since its inception, the AI SPE has driven Livepeer’s AI efforts by building the core network software needed to run AI compute jobs on the Livepeer network. This foundational work enabled the ecosystem team to validate the impact, potential, and competitiveness of bringing AI based video compute jobs onto the Livepeer network, supported by initiatives such as hackathons and the first cohort of the startup program.
The AI SPE has successfully completed two treasury-funded development phases, focusing on creating and stabilizing the core network software required to onboard the first AI orchestrators and enable them to serve both generative AI jobs (such as image creation with diffusion models) and batch AI jobs (such as data analysis and labeling tasks like transcription and segmentation). As part of this effort, the team developed the first AI pipelines and delivered comprehensive documentation, SDKs, detailed guides for orchestrators and developers, and robust monitoring tools to track performance and reliability through orchestrator and gateway metrics, among other key deliverables.
These efforts, carried out in collaboration with the broader Livepeer open-source community, have empowered multiple startups to perform hundreds of thousands of AI compute jobs on the Livepeer network, generating more than $23,000 in fees. Detailed summaries of all the group’s deliverables, along with a fully transparent financial overview, are available in our first-phase and second-phase retrospective forum posts.
Given the substantial work still required to achieve our overarching mission—to establish Livepeer as the premier decentralized platform for media-related AI compute jobs within the next three years—we are now preparing for the next treasury proposal in collaboration with the ecosystem team. Below, you’ll find more details about our mission and a high-level roadmap outlining the key deliverables and major features planned for the upcoming proposal phase to help us realize this vision.
The AI SPE’s greatest asset is its people—a collective of dedicated community members, contributors, and collaborators whose passion and expertise fuel our efforts. From seasoned Livepeer engineers at the forefront of core developments to open-source developers contributing through the Bounty Program, everyone plays a crucial role in advancing decentralized AI infrastructure. Below, we introduce the core team members of our open-source contributor group, collaborators from the ecosystem team (not funded by the SPE), and spotlight some of the most impactful contributions from the broader Livepeer community.
https://contributors-spotlight.vercel.app/widget