Mirror Security Launches Encrypted AI Inference

Mindflair plc, the AIM-quoted company focused on investing in Artificial Intelligence ("AI") and related technology, notes that Mirror Security Limited ("Mirror"), a Dublin-based cybersecurity company specialising in data protection across the generative AI ecosystem, has announced full production availability of its Encrypted AI Inference platform utilising NVIDIA Accelerated Computing, solving a problem that has blocked AI adoption in healthcare, finance, government, and defence. Mirror is a portfolio company held within Sure Valley Ventures' ("SVV") third fund ("SVV3"), in which Mindflair holds an interest.

Mirror, spun out of University College Dublin, enables AI systems to run inference on encrypted data - processing sensitive information without ever decrypting it. Mirror's platform keeps data encrypted throughout computation. The platform integrates NVIDIA AI inference capabilities across NVIDIA's accelerated computing infrastructure, enabling Secure AI Inference and Secure AI Memory, where data remains encrypted throughout computation and storage.

Mirror operates across Ireland, the USA and India, working with enterprises, government agencies and defence organisations that require AI deployment without compromising confidentiality or regulatory compliance.

This launch supports India's ambitious IndiaAI Mission to build comprehensive indigenous AI capabilities. Yotta Data Services, NVIDIA's partner in India and operator of the 16,000+ GPU Shakti Cloud, will be the first to deploy Mirror's platform for Indian enterprises, government, and defence organisations. Mirror believes that this launch supports regulated organisations in deploying AI at scale while maintaining encryption and data control.

Nicholas Lee, Director of Mindflair plc, commented:

"Mirror's full production launch using NVIDIA AI Infrastructure is a major milestone for the company-taking encrypted AI inference into full production and enabling regulated organisations to adopt AI at scale without decrypting sensitive data or compromising privacy and compliance."