GPU.Net
GPU.Net is a decentralized GPU compute marketplace operating 18 datacenters across 8+ countries on its proprietary GAN Chain L1 blockchain, serving AI developers, enterprises, and sovereign AI programs with on-demand GPU rental, managed services, and an ASTRA AI suite at 73-97% lower cost than hyperscalers.
- Company typePrivate
- Founded2022
- HeadquartersDelhi, India
- Headcount11–50
- GTM typeB2B and B2C
- OfferingSoftware
What GPU.Net does
GPU.Net operates a decentralized GPU compute marketplace that aggregates idle and partner-supplied GPU capacity from 18 datacenter sites across 8+ countries, settling transactions on its proprietary GAN Chain Layer 1 blockchain. Founded in 2022 in Delhi, India, the platform connects GPU providers with consumers through a self-serve dApp (dapp.gpu.net) where customers rent compute by the hour at $0.10-$0.50/hour — advertised as 73-97% cheaper than hyperscaler alternatives. The core technology stack includes the GAN Chain L1 (PoS BFT consensus, ~4,820 TPS, 1.2s finality, 799 validators), a Subnet architecture for self-governing specialized GPU marketplaces, and a Compute Score benchmarking system (CU x CM x CH) that determines provider rewards. The company has built strategic partnerships with NVIDIA (since 2022), Northern Data/Tether (25,000+ GPUs in Europe), G42 (UAE), and others, while operating owned infrastructure including a 5MW Bangalore facility hosting H200 GPUs and a $3M Hong Kong deployment.
Revenue is generated through five primary streams: usage-based GPU rental (the core marketplace), enterprise cluster reservations (quote-based, multi-year), managed services (cluster management, data preparation, model optimization targeting $0.5-10M pain points), $GPU token transaction fees across the GAN Chain ecosystem, and Solana-based Barrel of Compute NFTs. Customer segments span AI/ML developers and researchers (primary), enterprise AI companies and sovereign AI programs (primary), Web3/blockchain projects, VFX/animation studios, and GPU providers/miners monetizing idle hardware. Go-to-market combines product-led growth (self-serve dApp), community-led growth (350,000+ Questboard participants, 12,000+ Discord members, 150+ KOLs), enterprise field sales for large cluster bookings, and an event-driven motion via the GPU x branded event series across APAC and the Middle East.
The AI surface sits atop the infrastructure layer: the ASTRA suite offers Image generation, Video generation (using third-party Seedance 2.0 model), and AI Agents (supporting MCP/A2A/ANP/ACP protocols) accessible via API and the gRouter inference gateway. GPU.Net has raised $5.75M total funding (a $500K seed in August 2023 and a $5.25M Series A in April 2024), self-reported $10M revenue in 2024, and is preparing for a Token Generation Event (TGE) for its native $GPU token that will activate token-based settlement, staking, and governance across the ecosystem.
GPU.Net firmographics
Firmographics- Name
- GPU.Net
- Legal name
- GPU.net
- Website
- https://gpu.net
- Company type
- Private
- Founded year
- 2022
- Operating status
- Operating
- Headcount range
- 11–50 employees
- Short description
- GPU.Net is a decentralized GPU compute marketplace operating 18 datacenters across 8+ countries on its proprietary GAN Chain L1 blockchain, serving AI developers, enterprises, and sovereign AI programs with on-demand GPU rental, managed services, and an ASTRA AI suite at 73-97% lower cost than hyperscalers.
- Ownership category
- akta.pro rank
GPU.Net industry classification
Industry- Product category
- Decentralized GPU Compute
- NAICS
- Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services (5182), Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services (518), Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services (51821), Computer Systems Design and Related Services (5415)
- SIC
- Services-Computer Processing & Data Preparation (7374), Electronic Computers (3571)
- akta.pro primary industry
- Decentralized GPU/AI Compute Networks (training/inference, GPU renting, model hosting) (FSAPALAC)
- akta.pro secondary industries
- AI Compute Cloud & GPU-as-a-Service (HDAAAAAK), Decentralized Compute Markets (general-purpose compute, serverless, containerized workloads) (FSAPALAB), GPU-Accelerated HPC Systems (HDACAOAC)
Keywords
Where GPU.Net is headquartered
LocationHeadquarters
- HQ city
- Delhi
- HQ country
- India
- HQ region
- Asia
Offices8 records
Markets served
GPU.Net business model
Business model- GTM type
- B2B and B2C
- Offering type
- Software
- Cost components
- Infrastructure, Technology or R&D, Personnel, Marketing or Sales, Operations
Revenue model
- GPU Rental (On-Demand Compute): The primary revenue stream: customers rent GPU capacity by the hour from the global marketplace of 18 datacenter sites. Flexible, contract-free access to NVIDIA GPUs (RTX 4090, A100, H100, H200) across multiple regions. Pre-installed software stack includes NVIDIA drivers, Docker, CUDA, Python, and ML frameworks. Users configure GPU type, region, RAM, and storage, then SSH into running instances.
- Large Cluster Booking / Reserved Capacity: Enterprise customers can reserve dedicated GPU clusters for sustained workloads, with configurable cluster sizes and indicative pricing (final terms confirmed in quotation). Targets AI training, large-scale simulations, and research institutions requiring guaranteed capacity and data residency compliance.
- Managed Services (GPU Cluster Management, Data Preparation, Model Optimization): Professional managed services for enterprise deployments: GPU cluster management (NDA and deployment handling, 2x faster time to deployment, zero idle usage), data preparation for LLM training across multiple stages, and model optimization using memory management techniques. Pain points addressed: 15-20 days/$0.5-1M for cluster setup, 45-90 days/$1-3M for data prep, 180-360 days/$5-10M for model optimization.
- $GPU Token Transaction Fees: GAN Chain's $GPU token serves as the settlement asset for all on-chain transactions — subnet creation, staking, validator rewards, and fee payments. Users pay small $GPU fees for subnet services. Block rewards (fixed emissions distributed based on Queen votes and compute usage) create ongoing token utility. Fee burns and a token burn mechanism reduce circulating supply.
- Barrel of Compute NFT (Solana): NFTs minted on Solana (sol.gpu.net) that provide GPU credits, Queen NFT swaps, compute redemption, and are burnable for use on dapp.gpu.net. Listed on Magic Eden. New quests reward Barrel holders via token.gpu.net.
Pricing tiers
| Model | Billing | Price |
|---|---|---|
| Usage-based | Pay-as-you-go | Basic GPU tasks (light workloads) |
| Usage-based | Pay-as-you-go | High-performance AI workloads (intensive training/inference) |
| Usage-based | Multi-year contract | Reserved Capacity / Large Cluster Booking |
| Usage-based | Pay-as-you-go | Subnet Usage Fees |
Go-to-market motion4 records
Distribution channels6 records
Marketing channels9 records
GPU.Net product offering
Product offeringCore offering
GPU.Net operates a decentralized GPU compute marketplace built on its proprietary GAN Chain Layer 1 blockchain, enabling on-demand rental of high-performance NVIDIA GPUs (RTX 4090, A100, H100, H200) across 18 datacenter sites in 8+ countries via a self-serve dApp. The platform also delivers the Astra AI suite (text-to-image, text-to-video, AI agents, gRouter, Core API), Subnets as self-governing GPU marketplaces, GVEX decentralized exchange, and enterprise managed services including GPU cluster management, data preparation, and model optimization.
Product overview
GPU.NET is a decentralized GPU compute platform operating as a platform-plus-modules architecture. The core platform consists of GAN Chain (L1 blockchain for decentralized compute settlement) and the GPU.net dApp (marketplace for on-demand GPU rentals). ASTRA provides the AI intelligence layer with video generation, image generation, and AI agents. Additional modules include Subnets (user-created GPU-powered marketplaces), GVEX (verifiable exchange for $GPU/NODE trading), GPU Provider/Validator Nodes, GPU Cluster Management, Data Preparation, and Model Optimization services. The ecosystem is anchored by the $GPU token and the Questboard campaign driving community engagement. GPU.NET Enterprise offers dedicated infrastructure across 18 datacenters in 8+ countries with managed services for business customers.
Differentiator
Problem solved
Functional benefit
Brands
- Astra: AI product suite including Astra Video Generation, Astra Image, Astra Agents, and Astra Core API for AI model access.
- GAN Chain
- Subnets
- GVEX
- GRouter
- Barrel of Compute
Products and services
- GPUNET Enterprise Enterprise GPU compute infrastructure offering access to GPU clusters across 18 datacenters in 8+ countries, with managed services for GPU cluster management, data preparation for LLM training, and model optimization. Targets AI companies, research institutions, and sovereign AI programs requiring reserved capacity, data residency, and compliance.
- GPU.net dApp Decentralized application for renting GPU resources on-demand from the global GPU network, connecting users to available GPUs and enabling marketplace browsing by GPU type and region with hourly rental and pre-installed software stack (NVIDIA drivers, Docker, Python, CUDA, ML frameworks).
- ASTRA AI Platform Unified AI intelligence platform providing access to AI models and inference services, including video generation, image generation, AI agents, and LLM inference capabilities accessible via API or web interface.
- ASTRA Video Generation Text-to-video generation service using the Seedance 2.0 model, supporting durations up to 15 seconds, resolutions up to 1080p at 24 FPS, with audio sync for matching audio generation.
- ASTRA Image Text-to-image generation service with prompt enhancement capabilities that auto-improve prompts before generation.
- ASTRA Agents Autonomous AI agent platform for building and deploying AI agents that can perform tasks, interact with users, and automate workflows, with protocol support for MCP, A2A, ANP, and ACP.
- GAN Chain Layer 1 blockchain serving as the settlement layer for decentralized compute and marketplaces, where every subnet is a self-governing marketplace for GPUs, AI, storage, gaming, and social applications, secured by validators with PoS BFT consensus.
- Subnets User-created specialized networks on GAN Chain that leverage decentralized GPU power for tasks like AI bots, creative tools, and token launches, with creators (Kings), users, and investors (Queens) staking $GPU tokens.
- GVEX (GPUNET Verifiable Exchange) Ownerless peer-to-peer decentralized exchange with an order-book model for trading $GPU and NODE tokens natively on GAN Chain, featuring local stablecoin trading and deep liquidity.
- GPU Provider Nodes Node infrastructure for GPU providers to contribute surplus GPU capacity to the network and earn $GPU token rewards based on their compute score (CU x CM x CH).
- GPU Validator Nodes Validator infrastructure for securing GAN Chain, with validators staking $GPU, finalizing blocks, and earning $GPU emissions. Requires NFT license for authority.
- GPU Cluster Management Managed service for GPU cluster deployment and management that reduces deployment time from 15-20 days to faster implementation, handles NDA and model deployment, and enables easier workload scaling across distributed GPU clusters with zero idle usage.
- Data Preparation Managed service for LLM training data preparation covering pretraining, instruction-tuning, alignment, in-context learning, and task-specific fine-tuning stages.
- Model Optimization Optimization service enabling larger models on affordable GPUs through cutting-edge memory management techniques and efficient GPU resource utilization.
- API Access Public API for accessing AI models and inference services, with API pricing, key management for developers, and documentation at docs.gpu.net.
Quantifiable outcome
- 73-97% cost savings versus purchasing GPUs: 100 hours/month at $0.50/hour = $600/year vs. $2,250 purchase cost ($1,650 savings = 73% less), with savings reaching 97% for smaller workloads
- +5 more outcomes
Companies that use GPU.Net
Customer profileSegments5 records
Ideal customer profiles4 records
GPU.Net technology and API
TechnologyTechnology focussed Yes
API detail
- Has API
- Yes
- API docs
- API detail
Core technology
AI maturity
App detail
AI capability14 records
Feature8 records
GPU.Net partnerships and signals
Strategic signalPartnerships
Twelve partnerships are on record, tiered core, major and minor.
- G42coreGPU.Net partnered with G42, a leading AI and cloud infrastructure provider in the UAE, to manage datacenter operations and support regional GPU deployment. G42's AI expertise and UAE market presence complement GPU.Net's infrastructure in the Middle East region.
- Northern DatacoreStrategic tie-up with Northern Data (backed by Tether) gives GPU.Net access to over 25,000 GPUs across high-density facilities in Europe. Partnership announced Q4 2024. Northern Data's GPU infrastructure complements GPU.Net's decentralized marketplace with additional compute supply.
- NVIDIAcoreOfficial partnership with NVIDIA since 2022 to deploy and leverage high-performance GPUs (H100, H200, RTX 4090) across GPU.Net's ecosystem. NVIDIA GPU technology is the foundational infrastructure of the decentralized compute network, with NVIDIA explicitly listed as a strategic partner and GPU partner on the enterprise page.
- AMDmajorGPU.Net is integrating directly with AMD's virtualized cloud and engineering team, unlocking up to 2.4x HPC performance compared to NVIDIA H100. AMD integration planned to provide GPU diversity and sovereign AI options for customers seeking non-NVIDIA hardware.
- ViettelminorVietnam's largest telecommunications company. GPU.Net is in active discussions with Viettel for potential colocation and infrastructure partnerships in Vietnam, targeting Southeast Asian market expansion.
- Techcom BankminorMajor Vietnamese financial institution. GPU.Net is in active discussions with Techcom Bank for potential infrastructure partnerships in Vietnam, complementing Viettel discussions.
- Yotta InfrastructuremajorIndia's largest datacenter operator. GPU.Net continues infrastructure negotiations with Yotta for GPU deployment partnerships in the Indian market.
- NTT DatamajorGlobal IT infrastructure provider. GPU.Net is engaged in ongoing infra negotiations with NTT Data for enterprise GPU deployment and managed services partnerships. Sharad Sanghi, former CEO of NTT India, is supporting GPU.Net's growth and partnerships.
- CDAC (Centre for Development of Advanced Computing)minorPartnership with CDAC, India's leading supercomputing institution, to explore decentralized compute applications. Established in 2023.
- Taanga StudiosminorStrategic backer providing support to enhance GPU.Net's infrastructure for generative AI, Web3 Metaverses, graphics rendering, and cryptocurrency mining. Named alongside NVIDIA as a strategic backer at Testnet launch.
- U2U NetworkminorDePIN partner for Alliance Quests on the Road to TGE campaign. U2U token holders receive bonus GXP, strengthening cross-protocol community engagement.
- GaianetminorDePIN partner for Alliance Quests on the Road to TGE campaign. Gaianet token holders receive bonus GXP, fostering ecosystem integration between decentralized AI compute projects.
Scale indicators12 records
Recent moves6 records
Expansion highlights6 records
GPU.Net competitors and assessment
Company assessmentDirect peers
- Akash Network: Decentralized compute marketplace using Cosmos SDK; allows providers to lease GPU and CPU capacity. Most direct comparable to GPU.Net's decentralized GPU rental model and DePIN positioning, with an established AKT token economy and similar target customers.
- io.net: Decentralized GPU network purpose-built for AI/ML workloads, aggregating idle GPUs globally. Competes head-to-head with GPU.Net for AI developer and researcher customers needing cheap inference and training compute, with comparable Solana-based architecture.
- Render Network: Decentralized GPU rendering network originally focused on 3D/VFX but expanding into AI compute. Comparable token-incentivized GPU marketplace model with established RNDR token, significant artist/creator community, and overlap with GPU.Net's rendering use case.
- Gensyn: Decentralized protocol for machine learning compute that distributes training across a network of devices. Targets the same AI training compute wedge as GPU.Net's enterprise offering, with strong cryptographic verification positioning.
- Spheron Network: Decentralized compute and inference network for AI/ML workloads, aggregating idle GPUs from data centers and individuals. Closely comparable business model and customer segment to GPU.Net, with community-led growth and token incentives.
- Vast.ai: Marketplace for renting GPU compute from individuals and smaller providers; founded 2018 and a more established peer-to-peer GPU rental incumbent. Direct competition for GPU.Net's self-serve developer customer base with a more mature user experience.
Emerging players
- Flux: Decentralized cloud computing network offering GPU compute, hosting, and node infrastructure. Overlaps with GPU.Net's decentralized compute marketplace but with broader hosting use cases and an earlier-mover advantage in Web3 infrastructure.
Broad incumbents
- CoreWeave: Specialized GPU cloud provider offering NVIDIA H100/H200 clusters at scale, serving major AI labs and enterprises. While centralized rather than decentralized, it competes for the same enterprise GPU cluster reservation business that GPU.Net targets.
- Lambda Labs: GPU cloud provider offering reserved instances and on-demand GPU rental for AI training and inference. Comparable managed services (cluster deployment, model optimization) and customer overlap with GPU.Net's enterprise product line.
- Together AI: AI cloud platform providing GPU compute for inference and training with optimized inference stacks. Competes with GPU.Net's ASTRA AI inference layer and enterprise GPU business with a more software-forward, higher-margin positioning.
Market position
Strengths5 records
Weaknesses5 records
Competitive moat5 records
Key risks6 records
Key highlights7 records
Customer concentration
GPU.Net social profiles
Digital presenceGPU.Net financial estimates
Financial estimateRevenue estimate
Valuation estimate
GPU.Net leadership team
Management profileNumber of profiles
Profiles1 record
GPU.Net funding detail
Funding detailFunding overview
Funding rounds2 records
Investors12 records
Funding detail is available on the Subscription and Enterprise plan.Contact sales →
GPU.Net M&A and investment
M&A and investmentM&A
Investments
M&A and investment is available on the Subscription and Enterprise plan.Contact sales →
Frequently asked questions about GPU.Net
What does GPU.Net do?
GPU.Net operates a decentralized GPU compute marketplace built on its proprietary GAN Chain Layer 1 blockchain, enabling on-demand rental of high-performance NVIDIA GPUs (RTX 4090, A100, H100, H200) across 18 datacenter sites in 8+ countries via a self-serve dApp. The platform also delivers the Astra AI suite (text-to-image, text-to-video, AI agents, gRouter, Core API), Subnets as self-governing GPU marketplaces, GVEX decentralized exchange, and enterprise managed services including GPU cluster management, data preparation, and model optimization.
Is GPU.Net a public or private company?
GPU.Net is a private company. It is classified as venture growth investor backed and is currently operating.
When was GPU.Net founded?
GPU.Net was founded in 2022. It employs 11 to 50 people.
Where is GPU.Net based?
GPU.Net is headquartered in Delhi, India, in the Asia region.
How does GPU.Net make money?
Five revenue lines are on record. GPU Rental (On-Demand Compute) is the primary driver. The others are large Cluster Booking / Reserved Capacity, managed Services (GPU Cluster Management, Data Preparation, Model Optimization), $GPU Token Transaction Fees and barrel of Compute NFT (Solana).
Who are GPU.Net's main competitors?
Direct peers on record are Akash Network, io.net, Render Network, Gensyn, Spheron Network and Vast.ai. Flux is listed as an emerging player. Broad incumbents are CoreWeave, Lambda Labs and Together AI.
Does GPU.Net have an API?
Yes. GPU.NET offers a public API for AI model access and inference. The platform provides API pricing at /api-pricing and developer documentation at docs.gpu.net. Users can generate API keys via the dashboard at /models/keys. The ASTRA platform provides AI inference capabilities via API including video generation (Seedance 2.0), image generation, and AI agents. Developer documentation is at docs.gpu.net.
What industry is GPU.Net in?
GPU.Net's product category is Decentralized GPU Compute. Its primary akta.pro industry code is FSAPALAC, Decentralized GPU/AI Compute Networks (training/inference, GPU renting, model hosting), with a secondary code of HDAAAAAK, AI Compute Cloud & GPU-as-a-Service. Its NAICS code is 5182 and its SIC code is 7374.