LazAI is excited to introduce the launch of its pre-Testnet, marking a major milestone in building a decentralized and trustworthy AI infrastructure. This initial phase allows both developers and users to contribute privacy-conscious data and mint Data Anchoring Tokens (DATs), helping to establish a transparent and governed foundation for artificial intelligence. LazAI is placing data ownership, verifiable computation, and community participation at the heart of AI innovation.

Build With Verified AI Data, Not Blind Trust

With the launch of the pre-Testnet, developers are invited to begin testing core elements of the LazAI stack. These include the generation of Trusted Execution Environment (TEE) proofs and integration with on-chain data anchors. The Alith SDK enables builders to craft AI models and logic that can prove their inputs and data sources, eliminating the uncertainty and opacity of traditional centralized systems.

This phase prioritizes feedback from active developers to refine the architecture ahead of the upcoming full Testnet release.

Introducing Data Anchoring Tokens (DATs)

A key innovation of LazAI is the introduction of Data Anchoring Tokens, or DATs. These on-chain assets represent verified data contributions and can be minted through LazAI’s CorruptedAlith (Calith) mini-app. Contributors will be able to submit datasets ranging from personal metrics to niche domain information, anchoring them securely and transparently. Users will be able to contribute through Calith and other dApps built by developers. Developers engage directly with the flow to contribute and earn DAT.

DATs are more than a token. They allow users to stake their data’s authenticity and value, creating a new model for knowledge-based incentives. This is similar in spirit to liquidity mining in DeFi, but focused on information instead of capital, making data a participatory and composable building block of AI systems.

A Focused Builder Phase for High-Impact Testing

The current phase is deliberately focused on engaging serious developers, rather than broad public scaling. Participants can begin working with tools and features such as:

• High-integrity data tracking using DATs for traceability and quality assurance
• On-chain model referencing through the Alith SDK, streamlining AI inference and improving cost efficiency
• Trusted Execution Environments for secure and provable data ingestion without relying on centralized oversight

Built on a Scalable Foundation for Verifiable AI

LazAI, incubated by the Metis ecosystem, ensures that AI development can scale without compromising on decentralization or data accountability. By embedding verification at the infrastructure level, LazAI is laying the groundwork for a new era where AI is not just intelligent, but trustworthy and open to scrutiny.

LazAI is also featured in the HyperHack program, offering a dedicated $30,000 reward pool for projects that integrate the Alith SDK into their applications. Get to know Alith here.

This effort aligns with Metis’s broader Hyperion roadmap, which includes the launch of the public Testnet in May and mainnet rollout planned for August. These upgrades will support throughput levels of up to 2 gigagas per second, a critical benchmark for high-volume AI-powered applications.

A New Standard for AI Transparency and Ownership

LazAI sees verifiable data as the foundation for trustworthy AI systems. In today’s centralized models, data often enters black-box environments without proof of origin or quality. LazAI introduces a radically different framework where contributions are rewarded, provenance is preserved, and AI outputs are accountable to users and builders alike.

“AI is scaling fast, but without proof, it’s just a guess,” says Ming Guo, co-founder of LazAI. This pre-Testnet lays the groundwork for a builder-centric community driving the decentralized AI renaissance.

How to Get Involved

Developers interested in testing the Alith SDK and contributing to the evolution of verifiable AI can join the pre-Testnet by visiting the official Telegram bot. Technical documentation, test DAT minting tools, and faucet access are in continuous development to support onboarding.

Data contributors will also be able to visit the mini-app to upload datasets and mint DATs. In addition, developers can look forward to a new pre-Testnet Points system in the works. LazAI is not just building tools. It is shaping the future of AI with proof over promises. 

Join the Hyperion Forum to connect with builders, and follow the socials to stay up to date on every release and milestone.

‍