Irys continues to expand AI data infrastructure; will programmable data become the new direction in the next phase?

Since 2026, AI Agents, automated workflows, and on-chain AI narratives have continued to expand. The market's focus on AI infrastructure has gradually shifted from solely model capabilities and GPU computing power to how data is called, verified, executed, and collaborated on. Against this backdrop, Irys's ongoing strengthening of its AI Datachain and "programmable data" approach has begun to re-enter discussions around AI infrastructure and developer ecosystems.

Compared to traditional decentralized storage projects that primarily address "how data is stored long-term," Irys is now attempting to answer a more complex question: when AI Agents start participating in on-chain transactions, automated execution, and cross-protocol collaboration, is data still just a static storage object, or does it need to become a resource that can be called, verified, and involved in on-chain logic execution by AI? This shift has also led Irys's market positioning to gradually move from Storage Infrastructure toward the AI data execution layer.

Irys's Recent Focus on AI Datachain and Programmable Data

The core change in Irys over the past year is that its narrative has clearly shifted from traditional storage infrastructure to AI data infrastructure.

In early 2025, Irys launched a testnet for its Programmable Datachain tailored for AI scenarios and began continuously updating its roadmap around AI-native infrastructure, verifiable AI, and on-chain data execution capabilities. The official emphasis is no longer just on data upload and long-term preservation but on whether data can become an on-chain resource that smart contracts can directly call, verify, and execute.

This is where the concept of "programmable data" becomes truly important.

In the past, on-chain data was mainly recorded and stored, but with the rise of AI workflows, data itself has begun to assume more functions. If AI Agents want to participate in on-chain automated trading, content generation, state judgment, and cross-protocol collaboration, they must access trusted data in real-time and trigger subsequent actions based on data results. This means the data layer is shifting from "passive storage" to "active execution."

What Irys aims to promote is fundamentally a data structure capable of participating in AI workflows.

This directional shift also makes Irys's approach markedly different from traditional Storage Chains. Compared to projects that emphasize storage capacity and long-term preservation, Irys now places greater emphasis on data execution capabilities, data verifiability, and on-chain automation and collaboration.

After the Expansion of AI Agent Popularity, the Market Begins to Focus on Data Execution Capabilities

Following the surge in AI Agent popularity, the market’s discussion focus on AI infrastructure is undergoing a clear change.

In early 2024, during the initial AI market boom, discussions mainly centered around model capabilities, inference performance, and GPU power. Whether it was NVIDIA, TSMC, or cloud giants, the core logic revolved around "expanding AI training demands." But as AI Agents and automated workflows gradually entered on-chain scenarios, developers began to realize that having only AI models was insufficient to support complex AI workflows.

For AI Agents to truly participate in on-chain tasks, several key issues must be addressed:

• Is the data source trustworthy?
• Can the data be verified in real-time?
• Can AI call data across protocols?
• Does the data support on-chain collaboration and execution?

This indicates that the integration of AI and crypto is shifting from "model competition" to "data structure competition."

Especially in scenarios like automated trading, prediction markets, AI collaboration networks, and on-chain identity systems, data is no longer just input material but directly influences the execution results of AI Agents. If data cannot be verified, traceable, or participate in on-chain logic, AI Agents are likely to remain at the conceptual demonstration level.

The data execution capabilities that Irys emphasizes now are a response to this context. Compared to traditional Web2 AI workflows, on-chain AI scenarios demand higher transparency, verifiability, and cross-application collaboration capabilities, which is precisely the direction Irys is trying to penetrate.

Why Programmable Data Is Entering Developer Ecosystem Discussions

The reason "programmable data" is entering developer ecosystem discussions is not just because of conceptual updates, but because AI workflows themselves are becoming increasingly complex.

In the past, blockchain infrastructure competition mainly focused on:

• Consensus efficiency
• Data availability
• Storage capacity
• Scalability performance

But with the expansion of AI scenarios, developers are discovering that data itself needs stronger interaction capabilities.

If AI Agents are to operate long-term, they need continuous access to on-chain and off-chain data; to automate tasks, they must verify data authenticity; to collaborate with other Agents, data must be composable and support state synchronization. This means data is no longer just "read," but will be integrated into the entire execution process.

The proposed programmable data route by Irys essentially aims to enable data to participate in smart contract logic, rather than just stay at the storage layer. If this approach succeeds, the value of the data layer will extend beyond "storing information" to enhancing AI workflow trustworthiness, automation, and cross-protocol collaboration.

This is also why more developers are beginning to revisit data structure issues.

A key change in the current AI infrastructure track is that the market is starting to reassess: whether future AI applications truly need not only models and computing power but also new data execution structures.

Changes in the Competitive Landscape Between Irys, Arweave, and Celestia

Irys’s current competitive direction is markedly different from traditional storage chains and modular DA projects.

Historically, the market often discussed Irys alongside Arweave because both involve data storage and on-chain data structures. But as Irys continues to strengthen its AI Datachain approach, its competitive logic has begun to diverge from traditional Storage Infrastructure.

Arweave leans more toward long-term data storage, Celestia focuses on modular data availability, and projects like EigenDA and Avail emphasize rollup data availability. In contrast, Irys now emphasizes:

• AI data calling
• Data execution capabilities
• Verifiable AI
• On-chain automation workflows

This difference indicates that Irys is aiming to enter a more AI-native infrastructure domain.

Especially as AI Agent popularity continues to grow, the market is re-evaluating: does the future of AI require an independent data execution layer? If AI workflows increasingly depend on on-chain verification and automation, traditional storage or DA structures may not fully meet these needs. This is one of the key reasons Irys’s current approach is gaining attention.

However, challenges remain.

Irys is still in an early stage; whether its AI Datachain can truly form an independent ecosystem requires more developer engagement and real-world application validation. Compared to mature storage and DA projects, AI data execution layers are still a nascent exploration in the market.

Why On-Chain AI Workflows Need New Data Infrastructure

The increasing complexity of on-chain AI workflows is a significant backdrop for the renewed activity in the AI data infrastructure sector.

Many past AI + crypto projects remained conceptual, but as AI Agents begin to attempt automated trading, governance, and on-chain collaboration, the market faces a real question: how can AI operate securely, transparently, and verifiably on-chain?

For on-chain AI scenarios, model capabilities alone are insufficient; data execution and verification are equally critical.

In automated trading, on-chain analytics, multi-agent collaboration, and AI-driven content scenarios, AI needs to access on-chain states in real-time, verify data authenticity, and execute complex logic. This suggests that future on-chain AI workflows will demand data layers far beyond traditional DeFi applications.

Irys’s ongoing emphasis on AI Datachain aims to serve as the data collaboration layer within AI workflows.

According to previously disclosed data, the network has processed over 600 million data transactions and covers more than 4 million active wallets. While this does not yet prove that the AI Datachain ecosystem is mature, it indicates that Irys has established a certain scale of infrastructure operation.

Meanwhile, in 2025, Irys completed a Series A funding round of $10 million, with investors including CoinFund, Hypersphere, Amber Group, Breed VC, and WAGMI Ventures. The AI data infrastructure remains early-stage, but institutional capital has already begun to position itself in the "AI + data layer" direction.

The market’s current focus is less on whether Irys can store data, and more on whether future AI workflows truly require a new on-chain data execution structure.

Risks Emerging as Competition in the AI Data Layer Intensifies

Although the narrative around AI data infrastructure is expanding, market opinions on this direction are also divided.

Currently, the AI infrastructure race is highly competitive, with Arweave, Celestia, EigenDA, Filecoin, and Avail all attempting to enter AI and data-related domains. At the same time, AI + crypto still lacks a truly large-scale killer app; most AI Agents and on-chain automation scenarios remain experimental.

This means market attention on Irys is more about "future infrastructure expectations" rather than mature commercial applications.

The biggest current debate is whether on-chain AI workflows truly need an independent data execution layer.

Bullish perspectives argue that as AI Agents and automation workflows grow more complex, traditional static data structures will no longer suffice, and data execution capabilities could become a key competitive advantage in the next phase of AI infrastructure.

Bearish perspectives believe that most AI Agents still lack genuine user demand, and the integration of AI and crypto has yet to produce large-scale applications, so AI Datachain may remain at the conceptual stage.

This divergence also explains why Irys remains a high-volatility, high-expectation AI infrastructure project.

Can Irys Expand Its Ecosystem Impact Post Mainnet Launch?

The future influence of Irys largely depends on the mainnet ecosystem and developer adoption.

For infrastructure projects, narrative can attract short-term market attention, but long-term value ultimately depends on developer engagement and real application needs. The current focus on the programmable data approach will need to be validated by whether developers actually build applications around the AI Datachain.

As we move into 2026, Irys continues to update its GitHub with IrysVM, multi-ledger architecture, and Bundler infrastructure, indicating a shift from pure narrative to foundational development tools.

If AI Agents and on-chain automation workflows continue to expand, the demand for data verification and execution capabilities could further grow. Conversely, if AI + crypto interest wanes or developers stick with existing storage and smart contract solutions, Irys’s differentiated approach might weaken.

Therefore, Irys’s real challenge is not just proposing the "programmable data" concept but ensuring that data truly integrates into developer workflows and on-chain AI scenarios.

Summary

The recent year’s strategic shift of Irys reflects a broader change in the AI infrastructure market’s focus.

Previously, the market mainly emphasized data storage and availability. But as AI Agents and on-chain automation workflows expand, data execution, verification, and collaboration capabilities are increasingly entering developer discussions.

Irys’s ongoing emphasis on AI Datachain and programmable data is an attempt to tap into this new direction.

In the short term, the AI data infrastructure sector remains in early development, with developer ecosystems, real demand, and workflow scale still needing validation. In the long run, if AI Agents evolve from interactive tools to on-chain execution entities, the data layer could become a critical competitive frontier in AI infrastructure.

FAQ

What does "programmable data" mean in Irys?

Irys’s programmable data refers to on-chain data that can not only be stored but also called, verified, and participate in AI workflows and on-chain automation.

Why does Irys emphasize AI Datachain?

Irys emphasizes AI Datachain mainly because, as AI Agents and on-chain automation scenarios expand, the market is increasingly focused on data execution and verification capabilities.

What is the difference between programmable data and traditional decentralized storage?

Programmable data emphasizes not only data preservation but also the ability for data to participate in on-chain logic, AI calls, and automated task execution.

How does Irys differ from Arweave and Celestia?

Irys currently emphasizes AI data calling and on-chain automation capabilities, whereas Arweave focuses on long-term storage, and Celestia on modular data availability.

What are the biggest risks in the current AI data infrastructure sector?

The sector is still in early stages; genuine AI workflow demands, developer adoption, and long-term ecosystem development still require further validation.