Skip to main content

Gonka

Category: AI Infrastructure

Gonka is a decentralized AI compute network designed to maximize global GPU efficiency for AI model training and inference, providing permissionless access to distributed computing resources Gonka was founded in 2023. The company is led by David Liberman. Based in Los Angeles, United States. Team size: 10-50. Total funding raised: $80 million. Latest round: Strategic Investment (Bitfury Group $50M November 2025). Key investors include Bitfury Group, Coatue Management, Slow Ventures, K5, Insight, Benchmark Partners, Web3.com Ventures.

Founded
2023
Headquarters
Los Angeles, United States
Team size
10-50
Total funding
$80 million

Value proposition

Democratizes access to high-performance AI compute by transforming idle GPUs into a shared resource pool, offering cost-effective and sustainable alternatives to traditional cloud providers like AWS and Google Cloud

Products and solutions

Decentralized GPU Network, GNK Utility Token (for compute resource payments), OpenAI-compatible API for seamless integration, Decentralized AI Training Fund (20% inference revenue allocation for open-source LLM training)

Unique value

First decentralized network specifically optimized for productive AI workloads, with 98%+ computational efficiency compared to traditional Proof-of-Work systems

Target customer

AI developers, machine learning researchers, enterprises requiring scalable compute resources, and individual GPU contributors

Industries served

Artificial Intelligence, Blockchain Technology, Cloud Computing Infrastructure, Distributed Systems

Technology advantage

Transformer-based Proof-of-Work 2.0 protocol + Sprint-based consensus mechanism that ties mining rewards directly to meaningful computational output rather than energy waste

How they differentiate

Gonka emphasizes community-aligned computation and avoids design features that concentrate power (e.g., delegation). It focuses on efficient resource allocation for AI workloads while maintaining decentralization, differentiating itself from competitors like Bittensor (incentivized model training) and Prime Intellect (collaborative intelligence).

Main competitors

Bittensor, Prime Intellect, Akash Network

Key partnerships

Bitfury Group ($50M 2025 investment), House of ZK (technical collaboration), Coatue Management & Benchmark Partners (strategic investors), Gcore, Hyperfusion, 6blocks (hosting partners)

Notable customers

Enterprise AI labs, Research institutions, Individual GPU contributors

Major milestones

2023: Founded by David and Daniil Liberman, 2023: Secured $18M initial funding from Coatue, Slow Ventures, etc., 2025: Raised $50M from Bitfury Group, August 2025: Official launch of decentralized AI compute network, 2025: Achieved 6,000+ NVIDIA H100-series GPUs in network, 2025: Bitfury announces $50M strategic investment

Growth metrics

Network capacity of 6,000+ H100-equivalent GPUs, launched mainnet in August 2025

Market positioning

Positioned as a cost-competitive, decentralized alternative to centralized cloud providers (AWS, Google Cloud) and other blockchain-based compute networks. Targets AI researchers, developers, and enterprises seeking transparent, scalable, and affordable GPU resources.

Geographic focus

Global, with infrastructure designed to leverage distributed hardware contributions worldwide. No regional concentration specified, but founders are based in Los Angeles, USA.

Patents and IP

Pending IP in decentralized compute verification and AI workload optimization algorithms

About David Liberman

Serial entrepreneur and former core product director at Snap Inc. Co-founded ProductScience.ai and Humanism Co. Currently CTO of Product Science Inc (the incubator of Gonka protocol)

Official website: