Huo Xing Cai Jing reports that on August 12, AI security startup Attestable officially launched and completed a $20 million seed round led by Jamin Ball of Altimeter Capital and Yonatan Mandelbaum of TLV Partners, with participation from Halcyon Futures, Cerca Partners, and several other investors. Attestable’s founder, Yogi, stated that as AI increasingly integrates into critical infrastructure, national security, and enterprise systems, verifying the trustworthiness of AI operations has become a global priority. The company aims to build a universal verification layer for cutting-edge AI labs, critical infrastructure, and national-level applications. Attestable leverages zero-knowledge proof (ZKP) technology to shift trust in AI systems from data centers to mathematical verification. This approach can prove that a specific output was generated by an approved model, model weights, input data, and execution policy—without exposing model parameters or user privacy data—and without requiring the model to be re-run. Attestable reports that its technology has already achieved verification inference for Meta’s Muse Glimmer 30B model on a single NVIDIA H100 GPU, reaching a speed of 85 tokens per second. The generated proofs are compact and quantum-resistant, enabling rapid verification. In response, Ethereum co-founder Vitalik Buterin published a post stating that this achievement signifies that the performance overhead of zero-knowledge proofs for large language models has now approached single-digit percentages. He noted that the next challenge is further reducing the computational cost of technologies such as fully homomorphic encryption (FHE).
AI security startup Attestable completes $20M seed round, achieves ZK breakthrough
MarsBitShare
AI + crypto news broke on August 12, 2026, when Attestable, a new AI security startup, announced a $20 million seed round led by Jamin Ball of Altimeter Capital and Yonatan Mandelbaum of TLV Partners. The company is building a general verification layer using zero-knowledge proofs to secure AI systems. The technology verifies model outputs without exposing data or parameters. A recent test on Meta’s 30B model achieved 85 tokens per second on a single H100 GPU. Vitalik Buterin noted that ZKP performance overhead is now nearing single digits, with FHE remaining a challenge. The firm’s focus on mitigating security risks in AI infrastructure has attracted significant attention from key investors.
Source:Show original
Disclaimer: The information on this page may have been obtained from third parties and does not necessarily reflect the views or opinions of KuCoin. This content is provided for general informational purposes only, without any representation or warranty of any kind, nor shall it be construed as financial or investment advice. KuCoin shall not be liable for any errors or omissions, or for any outcomes resulting from the use of this information.
Investments in digital assets can be risky. Please carefully evaluate the risks of a product and your risk tolerance based on your own financial circumstances. For more information, please refer to our Terms of Use and Risk Disclosure.



