Map / Model Security / Attestable
Attestable
attestable.com ↗ · LinkedIn ↗ · Open on the map
Capabilities
Zero-knowledge proof platform that cryptographically verifies AI system integrity, letting organizations confirm a model's outputs were correctly computed from a specified model and inputs without trusting hardware vendors, secure enclaves, or privileged operators.
Sensors / integration
Generates succinct zero-knowledge proofs alongside inference that attest to which model, weights, and inputs actually produced a given output, working across CPUs, GPUs, and accelerators without specialized trusted hardware.
Protections
Mathematically binds AI outputs to a verifiable computation trail, making it computationally infeasible to tamper with inputs, swap models, or fake results undetected, while preserving confidentiality of proprietary model weights and architecture.
Differentiator
Software-only, hardware-agnostic zero-knowledge cryptography instead of trusted-hardware attestation (TEEs/secure enclaves), giving mathematically absolute integrity guarantees that don't depend on trusting a vendor or operator; launched from stealth with a $20M seed at a $71M valuation.
Team & funding
- Founders
- Yogev Bar-On, Shahar Papini, Shahar Samocha
- Total raised
- $20M
- Investors
- TLV Partners, Altimeter Capital, Netz Capital, Cerca Partners
Other Model Security companies
ActiveFence · Advai · Adversa AI · AI Aramid · Alice · AptaSentry · Arize AI · Bosch AIShield · Citadel AI · Cranium AI · DataKrypto · Deepchecks · See all →
Last updated · Get the “Listed on” badge