NEAR AI and Sigma Software have formed a strategic partnership to launch confidential AI infrastructure aimed at enhancing security for enterprise and government AI workloads. This collaboration will leverage NEAR AI’s hardware-enforced confidential inference technology, which operates within Trusted Execution Environments (TEEs) to ensure that sensitive data is processed securely and encrypted throughout. The partnership’s initial pilot deployments will occur in the financial services sector and will include the use of IronClaw AI agents to automate complex workflows with strict access controls. This approach not only addresses the growing need for robust data privacy in AI applications but also offers verifiable proofs of secure computation, enhancing trust beyond traditional contractual guarantees.

NEAR AI: NEAR AI delivers confidential computing infrastructure leveraging Trusted Execution Environments with Intel TDX for CPUs and NVIDIA Confidential Computing for GPUs to enable private inference and verifiable AI execution. The platform supports encrypted model execution and multichain actions for production workloads. In partnership with Sigma Software, NEAR AI provides its confidential inference stack and IronClaw agents for pilot deployments in enterprise and government sectors, starting with financial services.
IronClaw: IronClaw is an open-source AI agent runtime developed by NEAR AI, implemented in Rust with features like encrypted credential vaults, WebAssembly sandboxing, and hardware-secured execution in TEEs. It emphasizes verifiable security to prevent data leaks in autonomous agents. Enterprises in the Sigma Software partnership will use IronClaw to automate sensitive workflows with strict access controls and cryptographic proofs.
Sigma Software: Sigma Software is a global technology consulting firm offering custom software development, systems integration, and engineering services for enterprises in industries including finance, healthcare, automotive, and public sector. The company focuses on building complex, security-critical digital systems. Through the new partnership with NEAR AI, Sigma Software integrates confidential AI infrastructure into client environments to deploy secure inference and agentic workflows.
Illia Polosukhin: Illia Polosukhin is co-founder of NEAR Protocol and an AI researcher focused on verifiable privacy and user-owned AI systems. He advocates for cryptographic verification over trust-based security in enterprise AI. In the partnership announcement, he highlighted NEAR AI’s proofs ensuring no unauthorized data access in confidential inference and IronClaw agents.
Katherine Tuluzova: Katherine Tuluzova serves as CEO of Sigma Software Group, bringing extensive experience from starting as a developer to scaling the company’s US operations. She oversees delivery of secure digital solutions for critical sectors. She noted that the NEAR AI partnership embeds security and transparency into AI architectures for business and public-sector processes.

`json
{
“Verifiable Proofs”: “Every inference generates hardware attestation and signed outputs confirming secure, unmodified execution.”,
“Hardware-Enforced Privacy”: “NEAR AI employs TEEs to encrypt prompts and outputs, decrypting only within isolated enclaves during computation.”
}
`