Modelyo Technologies

Modelyo's platform builds a hardware-rooted trust layer on confidential computing architectures, using trusted execution environments and confidential VMs on Intel (SGX, TDX), AMD (SEV-SNP), Arm (CCA) and NVIDIA confidential GPUs (Hopper H100/H200, Blackwell B200/GB200, and planned Vera Rubin support). Instead of protecting individual compute instances one at a time, it provides a mesh of confidential managed services — Kubernetes, networking, observability, storage, databases, cache and queue, backup and disaster recovery, and AI/ML services — that all inherit the same security and trust architecture. Encryption keys are held customer-side in the customer's own HSM or key management service and are never held by the cloud provider or Modelyo; access to sensitive data is bound to environments verified against those keys. Continuous attestation checks infrastructure and workloads against trusted policies (every 60 seconds, per the company) and automatically moves workloads if integrity changes. For AI, the platform seals prompts, memory, reasoning traces and retrieval lookups, verifies each agent tool call and retrieval against customer keys, cryptographically separates tenants sharing GPU infrastructure, and returns cryptographic proof of which model ran on which data, at near-native throughput on production GPUs. Deployment uses existing Kubernetes tooling and identity providers.

Governments, defense organizations, financial institutions, healthcare and pharmaceutical companies, critical infrastructure operators and large enterprises seeking to adopt cloud and AI services while retaining control over their data and intellectual property.

Benny Meir - CEO & Co-Founder

Michael Kiperberg - CTO & Co-Founder