MOZAIK: A Privacy-Preserving Analytics Platform for IoT Data Using MPC and FHE

MOZAIK represents a classic academic approach to the 'Privacy-Enhancing Technology' (PET) stack, specifically targeting the IoT-to-cloud pipeline. While the combination of MPC (for interactive computation) and FHE (for computation on encrypted data at rest) is technically sophisticated, the project currently lacks any market signal, evidenced by its 0-star count despite having 8 forks—a pattern typical of a university research group or a peer-reviewed paper submission where co-authors/reviewers clone the repo. The defensibility is low because, while the math is hard, several specialized companies (e.g., Zama for FHE, Inpher or Cosmian for MPC) have much more robust, production-ready libraries that could implement this architecture more efficiently. The main 'moat' would be the specific protocol optimizations for resource-constrained IoT devices, but without an active community or high-performance benchmarks against industrial TEEs (Trusted Execution Environments) like Intel SGX or AWS Nitro Enclaves, it remains an academic exercise. Frontier labs (OpenAI/Anthropic) are unlikely to compete here as this is a niche infrastructure play, but cloud providers are a significant threat; they prefer hardware-level encryption (Confidential Computing) which offers significantly better performance than the software-defined FHE/MPC approach proposed here.