Lattica’s FHE Platform Aims to Secure AI Cloud Computing with $3.25M Funding

Lattica, a startup specializing in Fully Homomorphic Encryption (FHE), has exited stealth mode with $3.25 million in pre-seed funding to address privacy challenges in AI cloud computing. The round was led by Cyber Fund, founded by Konstantin Lomashuk, with participation from Sandeep Nailwal of Polygon Network and Sentient¹. The company’s technology targets industries like healthcare and finance where data sensitivity is critical.

Funding and Key Stakeholders

The $3.25 million pre-seed round underscores investor confidence in FHE’s potential to solve AI privacy gaps. Cyber Fund’s involvement signals strategic alignment with cryptographic innovations, while Nailwal’s participation bridges Web3 and AI security interests¹. Lattica’s CEO, Dr. Rotem Tsabary, brings academic rigor from the Weizmann Institute, specializing in lattice-based cryptography—a foundation for FHE implementations.

Technology: FHE and HEAL

Lattica’s platform uses Fully Homomorphic Encryption to process encrypted data without decryption, enabling secure AI model training on sensitive datasets. Its Homomorphic Encryption Abstraction Layer (HEAL) optimizes performance across hardware (GPUs, TPUs, ASICs), addressing FHE’s traditional computational overhead¹. A survey cited by the company indicates 71% of the FHE community believes hardware-software co-design is essential for adoption.

Industry Applications

The platform is designed for sectors with stringent privacy requirements. In healthcare, it could enable secure analysis of patient records; in finance, it might protect transactional data during fraud detection workflows. Government applications include secure surveillance analytics without raw data exposure¹.

Relevance to Security Professionals

For security teams, Lattica’s approach could reduce reliance on traditional data masking or tokenization. The technology’s ability to process encrypted data minimizes attack surfaces during AI inference. However, FHE’s computational demands may require infrastructure adjustments, such as GPU clusters for latency-sensitive use cases.

Comparative Landscape

Unlike AI One’s data lake bypass or Model ML’s financial automation, Lattica focuses on the privacy layer itself. Its FHE solution complements rather than competes with AI infrastructure tools, positioning it as a potential enabler for regulated industries adopting cloud-based AI^{2, 3}.

Future Implications

If successful, Lattica could set a benchmark for privacy-preserving AI in multi-tenant cloud environments. Challenges include scaling FHE for real-time applications and educating enterprises on its trade-offs versus existing encryption methods.