SpeedyBird Orchestrator Update: Scaling GPU Power & Network Contributions
Greetings Livepeer community,
SpeedyBird has been building in Livepeer for three years now, and I’m excited to share the next stage of growth. Delegators often ask how additional stake supports the network. The answer is simple:
- More GPUs online to increase capacity for transcoding and AI jobs
- More time invested into development, research, and ecosystem contributions
This dual focus ensures that delegating to SpeedyBird strengthens both the infrastructure of the network today and the innovation needed for tomorrow. Additionally, this powers contributions to the Cloud SPE (see below for more details).
As stake grows, SpeedyBird continues to expand GPU capacity while dedicating more time to Livepeer software development, research, and Cloud SPE initiatives that strengthen the broader ecosystem.
Current Operations
- 6 GPUs Globally Distributed:
- 1× RTX 5090
- 1× RTX 4090
- 1× RTX 3090
- 3× GTX 1080s
- Enterprise-grade East Coast server (owned): Datacenter-level hardware → Supermicro motherboard, 32-core AMD EPYC 7302, 128GB ECC RAM, redundant Intel P5500 NVMe RAID, solar-backed + 1Gbps fiber.
- Additional datacenter nodes provide uptime, redundancy, and global coverage.
Why Delegators Should Choose SpeedyBird
- Active & Committed: Currently ranked #44 by stake, yet consistently one of the most engaged and contributing operators — running the latest updates, deploying new GPUs, and supporting Cloud SPE initiatives month after month.
- Proven Contributions: From building the free Cloud Gateway to contributing in the early stages of Live AI Video (Showcase/Daydream), the AI Leaderboard, advanced Metrics stack, and Cloud SPE tooling, SpeedyBird has consistently delivered value beyond node operations.
- Enterprise Reliability: Redundant design, enterprise hardware, and sustainable energy.
- Global Datacenter Presence: Capacity across multiple regions.
- Long-Term Commitment: 3 years operating with a strong focus on ecosystem growth. Consistently active in community calls and Discord with impeccable uptime and reward calls.
Cloud SPE Focus
Beyond individual operations, I’m also part of the Cloud SPE, where we’re pushing forward several ecosystem-wide initiatives:
- Public gateway with free access
- GPU pool development & fair revenue sharing
- Job tester with AI Support (including live video)
- Robust demos, docs, and public APIs
- Open metrics and testing capabilities
Recent Cloud SPE Technical Updates
In recent weeks, we’ve focused on expanding the Cloud SPE’s ability to handle Live Video AI jobs, including:
- Enhancing the AI Job Tester to support live video stream inputs and playback quality detection.
- Adding scoring logic for new job types that were previously unsupported in the testing framework.
- Integrating MediaMTX and managing multiple service URIs not yet supported in the onchain registry.
- This work extends prior completed (and funded) proposals and required significant development and testing beyond simple configuration updates — all completed without additional funding. In fact the funding for these services ran out months ago.
Read the full Cloud SPE update here →
Looking Ahead
Delegating to SpeedyBird doesn’t just put GPUs online — it ensures continued innovation, transparency, and ecosystem growth. With your support, I’ll continue scaling both infrastructure and development to keep Livepeer at the forefront of AI-powered video.
Thank you for your trust and continued delegation.
— SpeedyBird