Irys partners with Warden Protocol to build an AI-native on-chain memory infrastructure.

It is reported that the data infrastructure public chain Irys has announced a strategic partnership with the AI-native Blockchain platform Warden Protocol. The two parties will jointly provide on-chain memory support for over 5 million users running AI models, and work together to create the next generation of data interaction infrastructure for intelligent applications.

As artificial intelligence models continue to increase in complexity and autonomy, the role of data has evolved from “passive invocation” to “core driving force.” Every inference output, user interaction, and environmental feedback forms the basis for the model’s ongoing learning, and this “contextual data” constitutes the true memory of AI. However, existing data systems—whether centralized cloud services or mainstream Blockchain storage protocols—are mostly built for static content, lacking real-time response, programming capabilities, and long-term verifiability, making it difficult to meet the structural needs of AI.

Irys, as the first Layer 1 Blockchain to support programmable data, natively integrates the smart contract execution environment IrysVM, a flexible multi-ledger architecture, a stable cost mechanism, and strong on-chain verification capabilities, providing a new data infrastructure pathway for AI models. Developers can not only store data permanently or temporarily on-chain but also embed permission management, invocation rules, and pricing logic through contracts, enabling the data to truly possess the characteristics of “readable, usable, and tradable.”

Warden Protocol focuses on creating a blockchain system that serves smart applications, currently supporting over 5 million users running various AI models. While the model layer continually upgrades, the underlying data system has always constrained performance release. Through this integration with Irys, Warden will gain native on-chain memory capabilities for the first time, allowing models to automatically invoke historical context in each inference, inherit behavioral logic, and build a long-term evolving intelligent system.

This cooperation not only represents an important leap for Web3 data infrastructure towards AI-native directions, but also validates Irys’s technological leadership in programmable data, native execution environments, and value capture mechanisms. In the future, both parties will continue to deepen their cooperation around on-chain model asset management, cross-model data reuse, and the commercialization of AI applications.