At NEARCON 2026, NEAR AI unveiled two major releases, IronClaw and the Confidential GPU Marketplace, introducing a privacy-focused infrastructure described as an AI Integrated Super App stack. The company stated that these launches arrive with new multimodal and privacy capabilities, presenting a system designed to be confidential by default and verifiable during execution. The announcement highlights a broader shift toward secure intelligent computing where trust, automation, and decentralized infrastructure operate together without compromise.
Source: X official
NEAR is a blockchain-based ecosystem built to run intelligent applications, agents, and services securely. It combines decentralized networks with cloud-style computing so developers can build advanced products without relying entirely on centralized providers. Through this architecture, automated agents can perform actions, while blockchain verifies activity for transparency and trust. The AI Integrated Super App vision focuses on creating a single environment where identity, payments, compute, and automation interact seamlessly, making it easier for companies to deploy intelligent solutions at scale.
One major launch is IronClaw, a safer version of OpenClaw designed so AI agents can complete real tasks without exposing sensitive data. OpenClaw is a framework that enables intelligent software to interact with applications and execute workflows across systems. IronClaw improves this approach by introducing encrypted environments that separate secrets from models, allowing automation without compromising privacy.
Key highlights of IronClaw include:
Secure execution of Artificial Intelligence workflows
Separation between sensitive data and Artificial Intelligence models
Ability for agents to access apps while keeping secrets protected
Verifiable logs showing actions performed
This development addresses a core challenge in modern automation, ensuring intelligent tools can act without risking confidential information.
Alongside IronClaw, NEAR introduced the Confidential GPU Marketplace, a platform where organizations can rent GPU computing power while keeping workloads private. The concept of confidentiality comes from Trusted Execution Environments, secure areas where code runs safely and data remains encrypted. NEAR DCML coordinates decentralized machine learning resources across this network, enabling secure distribution of computing capacity. Companies can therefore run training or inference without building expensive infrastructure, while ensuring that even infrastructure providers cannot access the underlying data.
Blockchain infrastructure has historically struggled to support enterprise-grade confidential computing. Many ecosystems offered storage and payments but lacked secure processing layers required for large-scale automation. These launches aim to close that gap by enabling trusted execution, scalable compute access, and runtime verification of automated actions.
Potential impact areas include:
Growth of decentralized Artificial Intelligence service marketplaces
Increased enterprise experimentation with private compute
Stronger infrastructure narratives across digital asset sectors
Expansion of secure agent-based applications
The Artificial Intelligence Integrated Super App direction signals convergence between Web3, artificial intelligence, and cloud infrastructure, which may influence long-term investment and product development trends.
NEAR AI’s IronClaw and Confidential GPU Marketplace mark progress toward secure decentralized automation. By combining private compute, verifiable execution, and scalable framework, the AI Integrated Super App approach could accelerate enterprise adoption and reshape Artificial Intelligence-blockchain development worldwide.
