Hyperbolic Labs
Hyperbolic is an AI cloud platform that makes global GPU compute accessible, liquid, and productive. It unifies heterogeneous GPU capacity so developers, researchers, startups, enterprises, and infrastructure providers can provision compute, run inference, and manage AI workloads through one platform.
At a Glance
- AI startups and application builders
- Researchers, research teams, and AI labs
- Enterprise AI and platform teams
- AI infrastructure and developer platforms
- +2 more
AI Tools by Hyperbolic Labs
(1)
Hyperbolic
GPU Cloud for AI Workloads
Discussions
No discussions yet
Be the first to start a discussion about Hyperbolic Labs
Latest News
Announced $30M Series A1 led by Walden Catalyst; introduced Forge expansion, multi-cloud orchestration, and Global Compute Grid plans.
Introduced Flux, next-generation inference powered by multi-cloud orchestration.
Introduced Managed Kubernetes for GPU workloads in beta.
Introduced Forge, the infrastructure layer behind reliable GPU provisioning.
Products & Services
Self-serve rental of NVIDIA H100, H200, B200 and other GPUs for training, fine-tuning, inference, experiments, batch jobs, and production; supports single instances through 1,000+ GPU clusters.
Fixed-term GPU capacity reserved in advance at discounted rates for predictable, sustained workloads.
Dedicated GPU infrastructure through Hyperbolic's supplier network for enterprise, sensitive, production-scale, and long-term workloads.
Serverless, token-billed inference for leading/open models through an OpenAI-compatible API, without customers provisioning or operating GPUs.
Market Position
Hyperbolic positions itself as an open-access, multi-provider AI cloud and liquidity/orchestration layer rather than a single-cloud GPU lessor. It competes for GPU compute and inference workloads with hyperscalers and specialist providers such as AWS, CoreWeave, Lambda, RunPod, Vast.ai, Together AI, and similar GPU clouds, differentiating through supplier aggregation, flexible access, open-model inference, and a planned compute marketplace/grid.
Leadership
Founders
Jasper Zhang, Ph.D.
Mathematician and AI-infrastructure entrepreneur; completed a UC Berkeley mathematics Ph.D. in two years and previously worked in market making at Citadel Securities. He is Hyperbolic's CEO and co-founder.
Yuchen Jin, Ph.D.
AI researcher and engineer, described by Hyperbolic as an award-winning researcher from the University of Washington; CTO and co-founder.
Executive Team
Jasper Zhang, Ph.D.
CEO and Co-Founder
UC Berkeley mathematics Ph.D. and former Citadel Securities market maker; leads Hyperbolic's company and compute-market strategy.
Yuchen Jin, Ph.D.
CTO and Co-Founder
AI researcher and engineer; Hyperbolic identifies him as an award-winning AI/math researcher and technology co-founder.
Founding Story
Jasper Zhang and Yuchen Jin saw researchers, startups, and developers blocked by the cost and scarcity of GPU compute. Their initial insight was that large amounts of consumer and data-center GPU capacity sit idle; by pooling and making that capacity accessible, they could make AI development more open, affordable, and collaborative.
Business Model
Revenue Model
Usage-based GPU rental and infrastructure services, including on-demand hourly GPU billing, prepaid/fixed-term reserved capacity, custom-priced private cloud, and token-based inference API billing. Hyperbolic also sells orchestration and infrastructure services to GPU providers and enterprise customers.
Pricing Tiers
H100 SXM from $3.19/hour; H200 from $3.99/hour; B200 from $5.99/hour. No minimum term; rates are refreshed based on supplier availability.
Capacity is reserved and paid for upfront for a fixed term; exact rate depends on GPU, duration, and availability.
Dedicated infrastructure and long-term capacity arrangements.
Model-specific input, cached-input where supported, and output token rates; accessed with a Hyperbolic API key and prepaid balance.
Target Markets
- AI startups and application builders
- Researchers, research teams, and AI labs
- Enterprise AI and platform teams
- AI infrastructure and developer platforms
- Neoclouds and GPU cloud providers
- Model developers and AI-native companies
- AI experimentation and prototyping
- Model training and fine-tuning
- Inference and serving AI applications
- Research benchmarking and confidential/ privacy-sensitive AI workloads
- Production AI infrastructure and predictable capacity planning
- GPU cloud providers monetizing and orchestrating heterogeneous fleets
- RadixArk
- Daytona
- Reve AI
- Pika