Sigma Software and NEAR AI Partner to Deploy Confidential Inference Infrastructure for Enterprise and Government AI Workloads

Sigma Software, a global engineering company building complex, security-critical digital systems, today announced a strategic partnership with NEAR AI to deploy confidential AI inference and secure agent infrastructure for enterprise and government clients. The partnership combines Sigma Software’s systems integration and deployment capabilities with NEAR AI’s hardware-enforced confidential inference, enabling organizations to run AI workloads with execution integrity and strict confidentiality around sensitive data.

As AI adoption accelerates, organizations are looking for ways to leverage powerful models and leverage agentic workflows while maintaining full control over sensitive data. Under the partnership, Sigma Software will integrate NEAR AI Cloud’s confidential computing stack into enterprise environments, beginning with pilot deployments in financial services. The engagement also includes deployment of IronClaw, NEAR AI’s open-source agent runtime, to automate complex business workflows with strict access controls and verifiable execution.

Confidential Inference Infrastructure

At the core of this partnership is NEAR AI’s unique approach to confidential inference, where AI models run inside hardware-enforced Trusted Execution Environments (TEEs). These secure environments ensure that prompts and outputs are encrypted in transit and only decrypted within an isolated enclave during computation. No cloud provider, infrastructure operator, or third party can access the data at any point in the process.

What makes this approach unique is that privacy is not based on trust alone, but on cryptographic verification. Every inference request generates verifiable proofs, including hardware attestation and signed outputs, that confirm the computation was executed inside a secure and unmodified environment.

Organizations can independently validate this by:

  • Requesting attestation reports tied to specific models and environments
  • Verifying hardware integrity using NVIDIA and Intel attestation services
  • Validating signed request and response hashes to ensure results have not been tampered with

This allows enterprises and governments to move beyond policy-based guarantees and instead rely on provable privacy and execution integrity, giving them full confidence in how their AI workloads are handled.

Learn more details on how verification works in practice here.

Beyond secure inference, the partnership supports the development of intelligent, secure AI agents using the secure and open source Ironclaw agent framework to automate complex business workflows across internal systems. These agents are designed to operate with strict access controls and verifiable execution, unlocking a new level of productivity without compromising security.

“Most enterprise AI security today comes down to contractual promises that your data won’t be accessed. That’s not trustworthy verification,” said Illia Polosukhin, Co Founder of NEAR. “From confidential inference to secure AI agents running on IronClaw, NEAR AI proves cryptographically that no one touched the data. Sigma Software can put that proof into production for mission-critical enterprise systems.”
“AI is increasingly embedded in critical business and public-sector processes,” said Katherine Tuluzova, CEO at Sigma Software. “Our partnership with NEAR AI strengthens our ability to engineer AI solutions where security, transparency, and control are built into the system architecture from the start.”

The partnership will launch with a series of pilot deployments across enterprise clients, with plans to scale globally and support high-impact use cases in industries where data security is critical.

About Sigma Software: Sigma Software is a global technology company delivering complex digital solutions for enterprises across industries including automotive, telecom, media, finance, healthcare, and public sector. Learn more at https://sigma.software/.

About NEAR AI: NEAR AI is an artificial intelligence research, engineering, and product development company committed to building an AI future owned by everyone. Founded by AI pioneer and former Google Research team member Illia Polosukhin, NEAR AI’s confidential and verifiable AI infrastructure empowers developers and enterprises to deploy AI models with full control over their data. With hardware-backed private inference via a simple API, NEAR AI Cloud runs sensitive AI workloads securely and at scale, from privacy-critical consumer interactions to autonomous systems and critical infrastructure. Serving users across major platforms such as Brave Nightly, OpenMind, and Venice AI, NEAR AI is proven infrastructure for transforming sensitive data into secure intelligence and advancing a user-owned AI economy. Learn more at https://near.ai/.

Share article: