What is 0G Labs?

0G Labs, formerly known as Zero Gravity Labs, is a modular Layer 1 blockchain purpose-built for on-chain AI. Its mission is to create an ecosystem where data, AI computation, and inference all take place directly on-chain, fully transparent and easily verifiable.

Since launching in 2024, 0G has secured over $35 million in pre-seed funding from top-tier investors including Hack VC, Delphi Digital, and Stanford Builders Fund, fueling the development of its testnets and long-term technology roadmap.

In recent testnet releases like Newton v2 and Galileo v3, 0G has demonstrated impressive performance: reaching 2,500 transactions per second, sub-second finality, and attracting participation from thousands of validators. These results highlight a robust and scalable architecture, purpose-built to power next-generation AI applications on-chain.

Architecture

0G Labs was built from the ground up to power a new wave of applications: AI on-chain. To make this possible, it leverages a multi-layered architecture consisting of four key components: 0G DA, OG Storage, 0G Chain, and dAIOS. These layers are tightly integrated to ensure high performance, robust security, and seamless scalability for both AI and Web3 workloads.

0G Storage

The foundation of 0G’s architecture is 0G Storage, which handles data using erasure coding – a technique that splits and encodes data into fragments, allowing it to remain recoverable even if parts of the network fail or go offline. This approach ensures high availability without the heavy data duplication required by traditional systems.

What truly sets it apart is the Proof of Random Access (PoRA) mechanism. Instead of passively storing data, each storage node must actively prove that it holds the requested data fragments. These challenges are issued randomly and automatically, and only nodes that respond correctly are rewarded. This decentralized verification method enforces data integrity without relying on centralized oversight.

0G DA

Building on top of the storage layer is the Data Availability (DA) layer, which ensures that stored data remains accessible to relevant parties, particularly rollups and AI workloads. Unlike traditional DA solutions that simply verify data presence, 0G DA uses a lightweight quorum-based validation system powered by Verifiable Random Functions (VRFs). This allows for randomly selecting a subset of validators to confirm data availability.

Rather than requiring the entire network to participate in verification, only a sufficient quorum is needed. This dramatically reduces latency and increases throughput – a crucial advantage for AI applications that demand access to large volumes of input data, such as inference logs or fine-tuning datasets.

A key benefit of this DA layer is its shareability: it can be utilized by multiple chains or rollups simultaneously. This not only conserves validator resources but also supports horizontal scalability across the broader ecosystem.

0G Chain

Next is the 0G Chain – a modular blockchain purpose-built to handle the transaction throughput and logic required for AI workloads. Unlike traditional chains, 0G separates the three core functions of execution, storage, and data availability into distinct layers. This modular design allows for greater performance optimization and customization.

One of its most powerful features is AI-native interoperability: smart contracts on the 0G Chain can directly invoke inference tasks from AI nodes, verify results via cryptographic proofs, and integrate those outputs into on-chain state. This enables entirely new types of use cases, such as smart contracts that react autonomously to machine learning outputs, without relying on third-party intermediaries.

dAIOS

At the top of the stack sits dAIOS – the decentralized AI operating system – which orchestrates and deploys AI services without intermediaries, all while remaining verifiable. In this layer, each AI model operates as an independent smart contract. Whenever a user runs an inference, the system automatically generates cryptographic proofs to confirm that the model executed correctly, on the right data, and produced verifiable results. This ensures full transparency and tamper-proof integrity throughout the process.

Unlike today’s “black-box” AI approach, dAIOS offers programmable infrastructure for building AI pipelines that draw directly from the underlying layers (storage, DA, and chain) and tie into decentralized application logic. Notably, inference services can be priced dynamically, settled instantly, and integrated into a tokenized AI economy – where data, models, and computation themselves become tradable assets.

Ecosystem

The 0G Labs ecosystem is an open network of over 236 projects, featuring more than 350 real-world integrations across its core infrastructure layers – Chain, Storage, DA, and Compute. Among them are prominent AI-native projects focused on inference and autonomous agents, including:

• SentientAGI
• Eliza OS
• OpenLedger
• Talus
• PlayArts
• Virtuals

The ecosystem also includes key infrastructure, DeFi, and ecosystem support partners such as:

• Chainlink
• Aethir (AethirCloud)
• EigenLayer
• Union
• Blade Games

Core Team

0G Labs was founded by a team of seasoned veterans from the tech and Web3 industries. CEO Michael Heinrich is a serial entrepreneur, known for founding Garten and teaching at Stanford. CTO Ming Wu brings deep technical expertise from his time at Microsoft Research and previously served as the CTO of Conflux. Co-founders Thomas Yao and Fan Long both have strong academic backgrounds and extensive experience in blockchain system design and venture investment.

Beyond the founding team, 0G’s broader technical crew includes talent from top-tier companies like Apple, Amazon, Chainlink, and Avalanche – a reflection of the project’s deep expertise and long-term vision.