BlockBeats reported on August 12 that AI security startup Attestable officially launched and completed a $20 million seed funding round. The round was led by Jamin Ball of Altimeter Capital and Yonatan Mandelbaum of TLV Partners, with participation from Halcyon Futures, Cerca Partners, and other investors.

Yogi, founder of Attestable, stated that as AI increasingly penetrates critical infrastructure, national security, and large enterprise systems, verifying the credibility of AI operations has become a crucial global challenge. The company aims to build a universal verification layer for cutting-edge AI labs, critical infrastructure, and national-level applications.

Attestable utilizes zero-knowledge proof technology to shift trust in AI systems from data centers to mathematical verification mechanisms. This technology can prove that an approved model, model weights, input data, and operating strategy do indeed generate a specific output, without revealing model parameters or user privacy data, and verification can be completed without rerunning the model.

Attestable states that its technology has already achieved verification inference of the Meta Muse Glimmer 30B model on a single NVIDIA H100 GPU, reaching a speed of 85 tokens per second. The generated proof files are small, quantum-safe, and can be verified quickly.

In response, Ethereum co-founder Vitalik Buterin stated that this achievement means the performance overhead of zero-knowledge proofs for large language models has approached single-digit levels. He indicated that the next challenge is to further reduce the performance overhead of technologies such as fully homomorphic encryption (FHE).