What Hyperbolic does

Hyperbolic Labs, Inc. operates an open-access AI cloud platform that aggregates global GPU supply — including a two-sided marketplace of data centers and individuals with idle hardware — into three core compute offerings: On-Demand Clusters (multi-tenant H100/H200 rentals at $1.49/hr with sub-minute deployment), Reserved Clusters (custom-configured dedicated capacity with InfiniBand/RoCE/NVLink and a 99.5% SLA, contracted at ~$10K/month and up), and Dedicated Hosting/Endpoints for 24/7 single-tenant inference. Layered on top is a Serverless Inference API that serves 25+ open-source generative models (Llama-3.1-405B-Base in BF16, Llama-3.3-70B, DeepSeek, Qwen, SDXL, Flux, plus audio and video) through an OpenAI-compatible endpoint with zero data retention, priced from $0.0001 per 1K tokens.

The company earns revenue across usage-based GPU rentals, per-token inference, multi-month reserved subscriptions (with 5-10% volume discounts and up to 40% on longer terms), higher-tier dedicated instance rates, marketplace service fees on supplier transactions, and AI consulting services. Distribution combines a self-serve product-led growth motion (web dashboard, Stripe and crypto payments, Basic/Pro freemium tiers, referral credits) with a sales-assisted enterprise motion for reserved clusters and dedicated endpoints; the developer surface extends into Hugging Face, VS Code/Cursor, Homebrew CLI, and an MCP server enabling AI agents to autonomously provision compute. Customers span individual developers, AI startups, research groups, enterprise AI teams, and infrastructure intermediaries, with notable enterprise contracts including QumulusAI (multi-year Blackwell GPU deployment) and WhiteFiber (H200 capacity supporting Modal Labs).

Founded around 2022 and headquartered in Irvine, California, Hyperbolic has raised approximately $19 million across a July 2024 seed round led by Polychain Capital and Lightspeed Faction and a December 2024 round with Variant and Polychain. The company serves a community of 200,000+ builders, pursues SOC 2 Type II and GDPR compliance (with HIPAA-ready configurations for qualified enterprise use cases), and operates data centers across North America, Europe, and Asia. Its stated 3-10x cost advantage versus inference competitors, combined with platform-level endorsements from figures such as Andrej Karpathy, positions it as a price-aggressive challenger to hyperscalers and specialty GPU clouds in the open-source AI serving and agentic compute segments.

Name

Hyperbolic

Legal name

Hyperbolic Labs, Inc.

Website

https://hyperbolic.ai

Company type

Private

Founded year

2022

Operating status

Operating

Headcount range

11–50 employees

Short description

Hyperbolic Labs operates an open-access AI cloud providing on-demand and reserved GPU rentals plus an OpenAI-compatible serverless inference API for 25+ open-source models, serving 200,000+ developers, AI startups, researchers, and enterprise AI teams with a stated 3-10x cost advantage.

Ownership category

akta.pro rank

Product category

GPU Cloud Infrastructure

NAICS

Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services (5182)

SIC

Services-Computer Processing & Data Preparation (7374), Services-Computer Programming, Data Processing, Etc. (7370)

akta.pro primary industry

AI Compute Cloud & GPU-as-a-Service (HDAAAAAK)

akta.pro secondary industries

Model Hosting, Serving & Inference Platforms (HDAAACAB), Decentralized GPU/AI Compute Networks (training/inference, GPU renting, model hosting) (FSAPALAC)

Keywords

• GPU cloud computing
• AI inference services
• Serverless inference API
• GPU marketplace platform
• AI infrastructure cloud

Headquarters

HQ city

Irvine

HQ country

United States

HQ region

North America

Offices3 records

Markets served

GTM type

B2B

Offering type

Software

Cost components

Technology or R&D, Infrastructure, Personnel, Marketing or Sales, Operations

Revenue model

1. On-Demand GPU rentals: Hourly GPU rental on multi-tenant clusters; users pay per-hour for H100, H200, and other GPUs from $1.49/hr; deploys in under a minute; no long-term commitment. Payment via credit card or crypto (USDC, USDT, DAI). Pay-as-you-go model.
2. Serverless Inference API: Usage-based pricing per 1K tokens for hosted open-source LLMs and image models; from $0.0001 per 1K tokens; OpenAI-compatible API; BF16 and FP8 precision tiers offered.
3. Reserved Clusters and Dedicated Endpoints: Multi-month to multi-year contracts (typically 3-12 months, starting at ~$10K/month) for dedicated GPU clusters with guaranteed uptime (99.5% SLA) and volume discounts (5-10% typical). Enterprise sales motion with custom quotes; ACH, wire transfer, or credit card. Custom model hosting also available.
4. Enterprise dedicated GPU instances: Higher-tier hourly rates for dedicated single-tenant GPU instances; H100 SXM at $3.20/hr, H100 PCIe at $3.00/hr, A100 SXM at $1.80/hr, A100 PCIe at $1.60/hr, RTX 4090 at $0.50/hr, RTX 3090 at $0.30/hr. Includes unlimited requests, dedicated support, custom SLAs.
5. Marketplace service fees (transaction fees): Hyperbolic charges Suppliers a service fee deducted from the Buyer purchase price; remainder is remitted to the Supplier. Functions as a limited payments agent between Buyers and Suppliers on the GPU marketplace.
6. AI Consulting / Professional Services: Engineering consulting for AI teams needing setup, scaling, and debugging help across training, fine-tuning, and inference workloads.

Pricing tiers

Go-to-market motion1 record

Distribution channels7 records

Marketing channels11 records

Core offering

Hyperbolic operates an open-access AI cloud that combines a multi-provider GPU marketplace with three managed compute products (On-Demand Clusters, Reserved Clusters, and Dedicated Endpoints) and an OpenAI-compatible Serverless Inference API serving 25+ open-source models. Customers rent H100/H200 GPUs by the hour, run inference via a drop-in API, or book multi-month reserved capacity with InfiniBand networking and 99.5% uptime SLAs. The platform is delivered to developers through a self-serve dashboard plus CLI, VS Code/Cursor extensions, and an MCP server for AI-agent orchestration.

Product overview

Hyperbolic operates a unified AI cloud platform branded as 'The Open-Access AI Cloud' (Hyperbolic Labs, Inc.) that combines a GPU marketplace, three core compute offerings — On-Demand Clusters, Reserved Clusters, and Dedicated Hosting/Dedicated Endpoints — with a Serverless Inference API for accessing 25+ open-source generative models via an OpenAI-compatible endpoint. The platform is delivered to developers through a layered stack: the underlying GPU marketplace aggregates supplier capacity; the three GPU rental products expose that capacity as managed compute; Serverless Inference layers hosted model serving on top; and developer-facing tools — the Hyperbolic CLI, the VS Code/Cursor extension, and the Hyperbolic MCP Server for AI agents — let users provision, manage, and serve workloads directly from terminal, IDE, or autonomous agents. AI Consulting Services wrap the entire stack for fast-scaling teams needing hands-on help.

Differentiator

Problem solved

Functional benefit

Products and services

• On-Demand Clusters Multi-tenant hourly rental of H100 and H200 GPU clusters deployable in under a minute, with no quotas or sales calls; pricing from $1.49/hr for H100s; intended for AI developers and teams that need flexible compute for training, fine-tuning, and short-term inference workloads.
• Serverless Inference OpenAI-compatible pay-as-you-go inference API hosting 25+ open-source generative models (Llama 3.1-405B-Base, Llama 3.3-70B, Qwen, DeepSeek, SDXL, Flux.1 [dev], plus VLMs, TTS, and text-to-video) with BF16 and FP8 precision tiers, zero data retention, and pricing from $0.0001 per 1K tokens; intended for AI developers building applications on open-source models.
• Reserved Clusters Dedicated GPU cluster reservations with 3-12 month commitments, custom networking (InfiniBand 400Gb/s, RoCE 200Gb/s, NVLink), volume discounts up to 40%, and a 99.5% uptime SLA; contracts typically start around $10K/month with minimum 8-GPU configurations; intended for enterprise teams running 24/7 production training and inference workloads.
• Dedicated Hosting / Dedicated Endpoints Single-tenant GPU hosting with private endpoints for LLMs, VLMs, and diffusion models; unlimited requests, dedicated support, custom SLAs, and tiered hourly rates (H100 SXM $3.20/hr, H100 PCIe $3.00/hr, A100 SXM $1.80/hr, A100 PCIe $1.60/hr, RTX 4090 $0.50/hr, RTX 3090 $0.30/hr); intended for workloads exceeding 100K tokens/min or requiring 24/7 throughput.
• GPU Marketplace Two-sided marketplace that aggregates GPU supply from data centers and individuals and lets buyers rent that capacity on demand; Hyperbolic acts as a limited payments agent, deducting a service fee from each buyer transaction and remitting the remainder to suppliers via Stripe; intended for both GPU providers seeking to monetize idle hardware and developers/enterprises seeking flexible supply.
• AI Consulting Services Engineering services from Hyperbolic engineers to help fast-scaling AI teams with setup, scaling, and debugging across training, fine-tuning, and inference workloads; sold as a professional services engagement alongside the platform's compute products.
• Hyperbolic CLI Command-line interface (MIT-licensed) for browsing, renting, and managing Hyperbolic on-demand GPU VMs and bare-metal clusters (including InfiniBand configurations in multiples of 8 GPUs); intended for developers who prefer terminal workflows.
• VS Code / Cursor Extension (Hyperbolic AI Cloud) IDE extension for VS Code and Cursor that enables in-editor GPU provisioning, SSH connections, Jupyter notebook execution, and vLLM model serving; supports gated Hugging Face models; intended for ML engineers working inside their editor.
• Hyperbolic MCP Server Model Context Protocol server that exposes Hyperbolic GPU cloud capabilities (list-available-gpus, rent-gpu-instance, terminate-gpu-instance, ssh-connect, remote-shell, etc.) to AI agents such as Claude Desktop via natural language; intended for developers building agentic AI workflows that need on-demand GPU orchestration.

Quantifiable outcome

• 3-10x less expensive than inference competitors
• +5 more outcomes

Named customers6 records

Segments4 records

Ideal customer profiles4 records

Technology focussed Yes

API detail

Has API

Yes

API docs

API detail

Core technology

AI maturity

App detail

Integration9 records

AI capability12 records

Feature9 records

Partnerships

Six partnerships are on record, tiered core infrastructure partner, supporting vendor, flagship distribution partner, standard app store partner and ecosystem partner.

• Stripe, Inc.core infrastructure partnerTechnology or IntegrationHyperbolic's third-party payment service provider for card acceptance, merchant settlement, and related services. All payments for the Services are processed through Stripe; Hyperbolic is bound to Stripe's Privacy Policy and Terms of Service.
• PostHogsupporting vendorTechnology or IntegrationPostHog provides session replay technologies and analytics on Hyperbolic's Service. Replays record users' clicks, mobile app touches, mouse movements, scrolls, and keystrokes to diagnose usability issues.
• Hugging Faceflagship distribution partnerTechnology or IntegrationHyperbolic's inference API is live on Hugging Face, enabling developers worldwide to access open-source models through Hugging Face's ecosystem. Co-marketing testimonial from Hugging Face CEO Clém Delangue featured on Hyperbolic's homepage.
• Apple Inc.standard app store partnerChannel Partner/ Reseller/ DistributorApple is a third-party beneficiary of Hyperbolic's Terms of Service for App Store Sourced Applications; Hyperbolic distributes its mobile application through the Apple App Store subject to Apple Media Terms of Service.
• Google (Google Play)standard app store partnerChannel Partner/ Reseller/ DistributorHyperbolic distributes its mobile application via the Google Play store under Google Play Sourced Application license terms.
• Polychain Capital (crypto VC ecosystem)ecosystem partnerGTM or Marketing PartnerPolychain Capital is recognized as one of the top crypto VCs aggressively investing in 'Crypto+AI' projects in 2024, with approximately 100% YoY financing growth and 138% more funding rounds in the sector.

Scale indicators13 records

Recent moves6 records

Expansion highlights6 records

Direct peers

• RunPod: GPU cloud offering on-demand and reserved GPU rentals plus serverless endpoints for AI inference. Direct competitor in the developer/individual tier of GPU rental.
• Akash Network: Decentralized GPU compute marketplace using blockchain-based coordination. Direct peer on the decentralized GPU aggregation thesis that Hyperbolic was originally built around.
• Modal Labs: Developer-focused serverless compute platform for AI/ML workloads with Python SDK and on-demand GPU access. Comparable to Hyperbolic's Serverless Inference and developer toolchain (CLI, IDE extensions) targeting AI engineers.
• CoreWeave: Specialized GPU cloud provider offering on-demand and reserved NVIDIA GPU clusters for AI training and inference. Closest direct competitor to Hyperbolic's reserved clusters and dedicated endpoint offerings, operating at much larger scale.
• Vast.ai: Peer-to-peer GPU marketplace aggregating idle consumer and data-center GPUs. Closest direct competitor to Hyperbolic's marketplace model with similar two-sided supply/demand dynamics.
• Crusoe: GPU cloud provider purpose-built for AI workloads, with large-scale H100/B200 clusters and competitive pricing. Competes with Hyperbolic's reserved cluster and dedicated endpoint offerings.
• io.net: Decentralized GPU network aggregating underutilized GPUs for AI training and inference. Closely parallels Hyperbolic's blockchain-based idle GPU aggregation origin story and target market.
• Together AI: AI cloud platform combining GPU compute with hosted open-source model inference (similar OpenAI-compatible API for Llama, Qwen, DeepSeek). Directly competes with Hyperbolic's Serverless Inference and On-Demand GPU products with overlapping developer audience.
• Lambda Labs: GPU cloud provider offering on-demand H100 instances, reserved clusters, and hosted inference for AI workloads. Direct competitor in the developer-focused GPU rental market.

Broad incumbents

• Amazon Web Services (EC2 GPU instances): Hyperscale cloud with broad portfolio including P-series and G-series GPU instances, Bedrock for hosted models, and global data center footprint. Competes broadly on price, scale, and enterprise compliance rather than specialist AI developer experience.

Market position

Strengths5 records

Weaknesses5 records

Competitive moat6 records

Key risks6 records

Key highlights7 records

Customer concentration

• Xx.com/hyperbolic_labs
• GitHubgithub.com/HyperbolicLabs
• YouTubeyoutube.com/@hyperbolic-labs
• LinkedInlinkedin.com/company/hyperbolic-labs

Compliance4 records

Revenue estimate

Valuation estimate

Number of profiles

Profiles2 records

Funding overview

Funding rounds4 records

Investors24 records

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about Hyperbolic