Yotta, Mirror Security Launch Encrypted AI Inference-as-a-Service

Yotta Data Services and Mirror Security have partnered to jointly provide secure, encrypted AI inference-as-a-service on Yotta's Shakti Cloud, one of India's largest sovereign AI infrastructures, featuring over 16,000 GPUs. 

This collaboration combines Mirror Security's production-ready fully homomorphic encryption (FHE) engine with Yotta's Tier IV sovereign AI cloud infrastructure, allowing businesses to execute AI workloads on sensitive data that remains encrypted during processing. 

The partnership aims to deliver mission-critical AI applications across sectors, including banking, financial services, healthcare, legal services, and government agencies—fields where data privacy and regulatory compliance are essential.

"With Yotta's world-class Shakti Cloud and Mirror's FHE Engine, Indian enterprises can now leverage the full power of AI without compromising data confidentiality or sovereignty. Data stays encrypted, keys stay with the customer, and intelligence stays in India," said Pankaj Thapa, Co-founder and CEO, Mirror Security, in a statement.

As stated in the press release, the secure, encrypted AI Inference solution guarantees compliance with India's AI Governance Guidelines and data localisation requirements, aligning with the country's ambition for homegrown, autonomous AI capabilities.

“Today, Yotta powers nearly 75% of India’s AI compute capacity, reinforcing our leadership in enabling the nation’s AI ecosystem. With Mirror Security’s advanced privacy-preserving capabilities integrated into Shakti Cloud, we are further strengthening our role as the trusted foundation for India’s most critical and regulated AI deployments,” said Sunil Gupta, Co-founder, CEO and MD, Yotta Data Services. 

Last week, Yotta launched Bhashini's comprehensive sovereign AI cloud transformation on Yotta’s Government Community Cloud and Shakti Cloud. As a result of this advancement, Bhashini now operates entirely on Indian cloud and GPU infrastructure, ensuring that language datasets, models, and citizen interactions remain within India.