NEAR AI used NEARCON 2026 to introduce IronClaw and the Confidential GPU Marketplace, framing both releases as building blocks for its AI Integrated Super App strategy. The stated goal is to let software agents carry out real tasks without exposing sensitive data, while keeping those actions verifiable.
One stack for identity, payments, compute, and automation
NEAR AI is presented as a blockchain-based system for smart apps, agents, and computing tools. Its pitch goes beyond onchain payments or storage. The platform combines blockchain rails with cloud-style services so developers can build modern products without relying fully on large centralized platforms.
In that model, software agents can take action and the network can help verify what happened. That is the core of the AI Integrated Super App idea: identity, payments, compute, and automation brought into one system. The trust problem is central here. Many software systems can perform tasks, but users and companies cannot always inspect what took place, and NEAR says it wants to close that gap.
IronClaw focuses on private agent execution with auditability
IronClaw is one of the headline launches from the event. NEAR describes it as a safer version of OpenClaw, designed for software agents that complete tasks while keeping sensitive information out of view. Its main feature is separation, with secret data kept apart from the model.
IronClaw also produces logs that can be checked later. So the underlying actions stay hidden, but the record can still be verified after the fact. That design targets two practical concerns for companies using automation tools: protecting confidential data and retaining a way to confirm what the agent did. In NEAR AI’s broader product map, IronClaw serves as the secure execution layer inside the AI Integrated Super App concept.
Confidential GPU Marketplace targets protected access to compute
The second major release is the Confidential GPU Marketplace. It is built for companies that need computing power and want to rent GPU capacity while keeping data protected. The source material points out a familiar issue: modern software often needs heavy compute, but many firms do not want to build that infrastructure themselves or expose data during training and inference.
NEAR’s answer is to use secure environments so workloads can run while data remains protected. In that sense, the marketplace is not just a GPU rental channel. It is also the compute layer in the AI Integrated Super App stack, aimed at making protected processing easier to access. Cost matters too, as the product is described as a way to reduce infrastructure burdens for companies.
NEAR is aiming at the gap between blockchain and secure computing
The message behind both launches is straightforward. Many blockchain systems support payments and storage, but secure computing is still less developed. NEAR AI is positioning IronClaw and the Confidential GPU Marketplace around that missing layer.
The source says these tools may make safer automation easier for companies to try and may also support growth in decentralized service platforms and agent-based applications. The larger direction is that blockchain, intelligent software, and cloud tools are moving closer together. NEAR wants its platform to sit at that intersection, with protected compute, trusted execution, and smart agents inside the same system.