Near Protocol price

Irys: The Core Infrastructure of a $3 Trillion AI Data Economy @irys_xyz is a project aiming to establish itself as a "Data Refinery" for the AI era, going beyond a simple decentralized storage solution. While traditional storage has merely focused on the simple storage of data, Irys expands data into Permanent Storage, Verifiable Compute, and Programmable Data, providing a new data infrastructure required by AI. This represents a paradigm shift that transforms data from a mere asset into an activated asset that operates and interacts on its own. The advancement of AI relies heavily on data. By 2025, the global volume of data is expected to exceed 180 zettabytes, and without infrastructure capable of efficiently storing and utilizing this data, the innovative potential of AI will inevitably be constrained. Irys targets this issue by presenting a query speed of around 200ms and a cost structure that is 16 times cheaper than Arweave, redefining data storage as AI-optimized infrastructure. The Limitations of Centralized Storage and the Need for Decentralization Today, most AI teams and Web3 startups still rely on centralized clouds like AWS, GCP, and Azure. However, this structure inherently has three limitations. 1. Data Silos Companies and institutions store data in distributed silos. This hinders the acquisition of integrated training datasets required by AI, resulting in the production of incomplete or biased models. In fact, a survey found that 70% of Web3 startups still depend on centralized clouds. 2. Security and Privacy Risks Centralized servers are prime targets for hacking. In just 2023, over 130 million health records were leaked, with the average cost per data breach incident reaching $10 million. Particularly, the leakage of training datasets for AI models can severely damage corporate value and trust. 3. Scalability Bottlenecks AI requires data-intensive operations. Currently, 80-90% of total AI computing is used for inference, placing a massive burden on data storage. Existing centralized systems cannot handle this demand, causing bottlenecks in terms of cost, energy, and outages. In contrast, decentralized storage structurally addresses these issues. Data is shared through a P2P network, breaking down silos, and security is enhanced through encryption and sharding structures. Additionally, it can leverage unused storage resources worldwide to achieve both cost savings and infinite scalability. Projects like 0G Labs and Shelby (in collaboration with NEAR/Aptos) have already successfully demonstrated this, and the DePIN market has grown to a scale of $30 billion, proving the feasibility of this model. Irys's Differentiation: AI-Centric Design Irys takes a fundamentally different approach compared to existing projects like Arweave or Filecoin. - Real-time Performance: While Arweave takes several minutes for block confirmations and Filecoin experiences delays of 2-5 seconds, Irys offers query speeds of under 200ms. This is a critical differentiator in real-time AI inference pipelines. - Programmable Storage: Irys can embed logic, permissions, and royalties directly into the data itself, allowing the data to interact directly with AI agents rather than just being stored. - Cost Structure: While Arweave's permanent storage costs over $7 per GB and Filecoin ranges from $4 to $5, Irys provides overwhelming cost efficiency with predictable fixed pricing and a free upload policy for data under 100KiB. - Independent L1 Architecture: Unlike Filecoin or Arweave, which rely on upper layers, Irys is designed as an EVM-compatible L1 chain, processing storage and execution in a single native environment. Use Cases: Irys Across the AI Stack 1. Provenance of Training Data and Models Irys stores datasets in an immutable form and records all modification histories on the blockchain. This enables data tampering prevention, protection against data poisoning attacks, and collaboration between researchers and DAOs. 2. Real-time Inference Pipeline The entire flow from input to computation to output in the AI inference process is transparently recorded, and smart contracts can be automatically executed upon data input. Additionally, privacy-preserving operations based on zero-knowledge proofs (ZKP) can also be supported for sensitive data. 3. Fusion of DePIN and AI DePIN devices like sensor networks and smart energy grids can permanently store data in Irys, and AI can analyze this data in real-time to build autonomous optimization systems. This can be utilized in predictive markets, traffic control, and IoT management. Key Partnerships and Cases - Surf Copilot (2025.09.16): A cryptocurrency AI analysis agent stores real-time data in Irys to provide tamper-proof market analysis. - Kite AI (2025.08.28): Collaborating with Irys to build an "app store" infrastructure for AI agents, establishing a comprehensive AI ecosystem covering data, identity, and governance. - SCAI Academic Collaboration: Permanently storing 100TB of scientific research data in Irys, building an intelligent paper search and review system. Alignment with the Web3 Narrative Irys is directly aligned with the core narrative of Web3 in 2025. - AI × Storage: Treating data as programmable assets for immediate use in AI inference. - DePIN: Connecting physical sensor data with AI to provide real-time intelligence. - ComputeFi: Combining storage and computation to eliminate the "double tax" problem. With 855M transactions recorded (Q3 2025) and a $10 million Series A investment (2025.08), along with a community growth evident in its 289,000 Twitter followers, Irys is showing significant momentum. Risks and Challenges - Scalability Validation: While testnet performance is excellent, it remains uncertain whether it can handle AI-scale data flows on the mainnet. - Privacy Limitations: The default public storage method may conflict with regulations like GDPR. - Adoption Barriers: For enterprises, initial costs and an immature ecosystem can be burdensome. - Intensifying Competition: Arweave has captured the permanent storage market, and Filecoin has a stable large-scale infrastructure. Centralized services like AWS remain cheaper and more mature. - Regulatory and Token Risks: The possibility of the $IRYS token being classified as a security and supply chain instability are also variables. Strategic Assessment - Irys: A performance leader optimized for real-time AI and programmable data. - Arweave: Specializing in archiving with a stronghold on permanent storage. - Filecoin: Enterprise infrastructure meeting large-scale corporate demand. The community refers to Irys as "Onchain Memory," highlighting its importance as a core infrastructure for AI agents. However, the key to success lies in validating mainnet scalability and securing a developer and enterprise ecosystem. Overall, Irys can provide high confidence in its technical differentiation and narrative fit, but only medium confidence in short-term adoption speed and competitive landscape. Nevertheless, the new perspective that Irys presents—data as programmable assets—will be a significant attempt to determine the direction of the evolution of AI×Web3 infrastructure.

Irys: The Core Infrastructure of a $3 Trillion AI Data Economy @irys_xyz is a project aiming to establish itself as a "Data Refinery" for the AI era, going beyond a simple decentralized storage solution. While traditional storage has merely focused on the simple storage of data, Irys expands data into Permanent Storage, Verifiable Compute, and Programmable Data, providing a new data infrastructure required by AI. This represents a paradigm shift that transforms data from a mere asset into an activated asset that operates and interacts on its own. The advancement of AI relies heavily on data. By 2025, the global volume of data is expected to exceed 180 zettabytes, and without infrastructure capable of efficiently storing and utilizing this data, the innovative potential of AI will inevitably be constrained. Irys targets this issue by presenting a query speed of around 200ms and a cost structure that is 16 times cheaper than Arweave, redefining data storage as AI-optimized infrastructure. The Limitations of Centralized Storage and the Need for Decentralization Today, most AI teams and Web3 startups still rely on centralized clouds like AWS, GCP, and Azure. However, this structure inherently has three limitations. 1. Data Silos Companies and institutions store data in distributed silos. This hinders the acquisition of integrated training datasets required by AI, resulting in the production of incomplete or biased models. In fact, a survey found that 70% of Web3 startups still depend on centralized clouds. 2. Security and Privacy Risks Centralized servers are prime targets for hacking. In just 2023, over 130 million health records were leaked, with the average cost per data breach incident reaching $10 million. Particularly, the leakage of training datasets for AI models can severely damage corporate value and trust. 3. Scalability Bottlenecks AI requires data-intensive operations. Currently, 80-90% of total AI computing is used for inference, placing a massive burden on data storage. Existing centralized systems cannot handle this demand, causing bottlenecks in terms of cost, energy, and outages. In contrast, decentralized storage structurally addresses these issues. Data is shared through a P2P network, breaking down silos, and security is enhanced through encryption and sharding structures. Additionally, it can leverage unused storage resources worldwide to achieve both cost savings and infinite scalability. Projects like 0G Labs and Shelby (in collaboration with NEAR/Aptos) have already successfully demonstrated this, and the DePIN market has grown to a scale of $30 billion, proving the feasibility of this model. Irys's Differentiation: AI-Centric Design Irys takes a fundamentally different approach compared to existing projects like Arweave or Filecoin. - Real-time Performance: While Arweave takes several minutes for block confirmations and Filecoin experiences delays of 2-5 seconds, Irys offers query speeds of under 200ms. This is a critical differentiator in real-time AI inference pipelines. - Programmable Storage: Irys can embed logic, permissions, and royalties directly into the data itself, allowing the data to interact directly with AI agents rather than just being stored. - Cost Structure: While Arweave's permanent storage costs over $7 per GB and Filecoin ranges from $4 to $5, Irys provides overwhelming cost efficiency with predictable fixed pricing and a free upload policy for data under 100KiB. - Independent L1 Architecture: Unlike Filecoin or Arweave, which rely on upper layers, Irys is designed as an EVM-compatible L1 chain, processing storage and execution in a single native environment. Use Cases: Irys Across the AI Stack 1. Provenance of Training Data and Models Irys stores datasets in an immutable form and records all modification histories on the blockchain. This enables data tampering prevention, protection against data poisoning attacks, and collaboration between researchers and DAOs. 2. Real-time Inference Pipeline The entire flow from input to computation to output in the AI inference process is transparently recorded, and smart contracts can be automatically executed upon data input. Additionally, privacy-preserving operations based on zero-knowledge proofs (ZKP) can also be supported for sensitive data. 3. Fusion of DePIN and AI DePIN devices like sensor networks and smart energy grids can permanently store data in Irys, and AI can analyze this data in real-time to build autonomous optimization systems. This can be utilized in predictive markets, traffic control, and IoT management. Key Partnerships and Cases - Surf Copilot (2025.09.16): A cryptocurrency AI analysis agent stores real-time data in Irys to provide tamper-proof market analysis. - Kite AI (2025.08.28): Collaborating with Irys to build an "app store" infrastructure for AI agents, establishing a comprehensive AI ecosystem covering data, identity, and governance. - SCAI Academic Collaboration: Permanently storing 100TB of scientific research data in Irys, building an intelligent paper search and review system. Alignment with the Web3 Narrative Irys is directly aligned with the core narrative of Web3 in 2025. - AI × Storage: Treating data as programmable assets for immediate use in AI inference. - DePIN: Connecting physical sensor data with AI to provide real-time intelligence. - ComputeFi: Combining storage and computation to eliminate the "double tax" problem. With 855M transactions recorded (Q3 2025) and a $10 million Series A investment (2025.08), along with a community growth evident in its 289,000 Twitter followers, Irys is showing significant momentum. Risks and Challenges - Scalability Validation: While testnet performance is excellent, it remains uncertain whether it can handle AI-scale data flows on the mainnet. - Privacy Limitations: The default public storage method may conflict with regulations like GDPR. - Adoption Barriers: For enterprises, initial costs and an immature ecosystem can be burdensome. - Intensifying Competition: Arweave has captured the permanent storage market, and Filecoin has a stable large-scale infrastructure. Centralized services like AWS remain cheaper and more mature. - Regulatory and Token Risks: The possibility of the $IRYS token being classified as a security and supply chain instability are also variables. Strategic Assessment - Irys: A performance leader optimized for real-time AI and programmable data. - Arweave: Specializing in archiving with a stronghold on permanent storage. - Filecoin: Enterprise infrastructure meeting large-scale corporate demand. The community refers to Irys as "Onchain Memory," highlighting its importance as a core infrastructure for AI agents. However, the key to success lies in validating mainnet scalability and securing a developer and enterprise ecosystem. Overall, Irys can provide high confidence in its technical differentiation and narrative fit, but only medium confidence in short-term adoption speed and competitive landscape. Nevertheless, the new perspective that Irys presents—data as programmable assets—will be a significant attempt to determine the direction of the evolution of AI×Web3 infrastructure.

I won’t lie I am so bullish on @idOS_network Backed by @fabric_vc @arbitrum @circleminiapp @RippleXDev @NEARProtocol and @Starknet A whole community of power house This is an alpha