OpenRelay
YC S26Seattle, US · Founded 2026 · 2 employees · 1 known investors
OpenRelay operates a distributed overlay network that connects consumer GPUs into a fault-tolerant mesh for inference workloads, automatically routing traffic around failed nodes without manual intervention. The platform allows anyone to contribute compute resources and deploy inference jobs via a single API call.
Also known as VectorLay
AI & Machine LearningCloud ComputingData & InfrastructureDeveloper ToolsEnterprise SoftwareAIUnited States of AmericaAmerica / Canada
Founders & leadership· Y Combinator alumni (S26)
OpenRelay was founded in 2026 by Prashant Patel and Jaden Wang.
PP
Prashant PatelinFounderHe previously served as Staff Engineer at Voltagepark, where he worked on managed inference and orchestration software, and was a founding member of Amazon Bedrock at AWS, where he led the Custom Model Import initiative and scaled inference infrastructure for large language models. He holds an MS in Computer Science from NYU.
JW
Jaden WanginFounderJaden Wang previously worked in high-performance computing roles at Voltage Park Engineering and TensorDock, and founded Heaviside Compute before starting OpenRelay; he attended the University of Washington.
Investors · 1
Competitors · 8
by search overlap
Runpod16 shared keywordsRunpod is an AI developer cloud that provides on-demand GPU infrastructure for experimenting with, training, fine-tuning, deploying, and scaling machine learning models. It offers autoscaling serverless GPU endpoints, persistent GPU pods, and multi-GPU clusters for AI developers and teams running inference and training workloads.
Vast.AI GPU compute cloud13 shared keywordsVast.ai operates a marketplace for on-demand GPU compute, offering pricing set by supply and demand across 20,000+ GPUs and 40+ data centers with API, CLI, and Python SDK access. It provides GPU Cloud instances, serverless inference endpoints, and dedicated multi-node clusters for developers and AI teams running training, inference, fine-tuning, and rendering workloads.
Northflank13 shared keywordsNorthflank is a developer platform for deploying and running application workloads across cloud providers (AWS, Azure, Google) and bare-metal without infrastructure management or vendor lock-in. It targets software engineers and ops teams building and shipping cloud-native applications.
Salad12 shared keywordsSalad operates a peer-to-peer cloud computing network where users share spare computing resources for rewards while businesses run workloads like AI and gaming applications on the affordable, distributed infrastructure. The company uses a proprietary trust system to manage node performance and optimize hardware allocation across its community-powered cloud platform.
Lambda Labs11 shared keywordsLambda provides cloud GPU infrastructure for AI training and inference, offering single-tenant clusters and instances built on NVIDIA GPUs such as GB300 NVL72, HGX B200, and H200. It serves AI labs and enterprises with managed superclusters, 1-Click Clusters, and on-demand instances featuring SOC 2 Type II compliance and hardware-level isolation.
Hive9 shared keywordsHivenet provides cloud infrastructure for AI and general workloads, including an inference API for open-weight models, GPU and CPU compute (RTX 4090/5090 and vCPU instances), S3-compatible storage, and private AI paths for RAG and fine-tuning. It targets researchers, startups, studios, schools, and technical teams that want cost control and regional/sovereignty choice over where workloads run.
Thunder Compute9 shared keywordsThunder Compute develops GPU virtualization software that improves utilization of graphics processing units in data centers. The company targets cloud providers and enterprises seeking to increase computational capacity without additional hardware investment.
Spheron8 shared keywordsSpheron operates a GPU cloud marketplace that aggregates compute capacity from certified data centers worldwide, offering NVIDIA GPUs (H100, H200, B200, A100) at 40-60% below hyperscaler pricing. The platform serves AI training teams and ML builders with on-demand GPU instances, reserved capacity, and custom cluster sourcing across multiple providers through a single dashboard.
Companies competing with OpenRelay for the same Google search keywords, organic and paid, via search-intersection analysis.
Frequently asked questions
- What does OpenRelay do?
- OpenRelay operates a distributed overlay network that connects consumer GPUs into a fault-tolerant mesh for inference workloads, automatically routing traffic around failed nodes without manual intervention. The platform allows anyone to contribute compute resources and deploy inference jobs via a single API call.
- Who founded OpenRelay?
- OpenRelay was founded by Prashant Patel, Jaden Wang in 2026.
- Who are OpenRelay's investors?
- OpenRelay's investors include Y Combinator.
- Where is OpenRelay headquartered?
- OpenRelay is headquartered in Seattle, US.