Discussion - Topic: Confidential Computing for AI Model Weights

About this event

This week's topic: Confidential Computing for AI Model Weights Discussion resources to help guide the conversation will be posted below a few days before the meetup. Zoom link will be added about 5 min before the event starts. We’ll discuss confidential computing for AI model weights—technologies that protect proprietary models while they are being loaded, processed, and used, not just when stored or transmitted. The conversation will explore trusted execution environments, encrypted memory, remote attestation, key management, and emerging GPU-based protections. We’ll also examine practical tradeoffs: performance, deployment complexity, cloud-provider trust, intellectual-property protection, and whether these approaches can genuinely prevent model theft or unauthorized access. Discussion Resources : Protecting Sensitive Data and AI Models with Confidential Computing By Uttara Kumar and Emily Sakata https://developer.nvidia.com/blog/protecting-sensitive-data-and-ai-models-with-confidential-computing (https://developer.nvidia.com/blog/protecting-sensitive-data-and-ai-models-with-confidential-computing?utm_source=chatgpt.com) Confidential Computing on NVIDIA H100 GPUs for Secure and Trustworthy AI By Emily Apsey, Phil Rogers, Michael O'Connor, and Rob Nertney https://developer.nvidia.com/blog/confidential-computing-on-h100-gpus-for-secure-and-trustworthy-ai/ (https://developer.nvidia.com/blog/confidential-computing-on-h100-gpus-for-secure-and-trustworthy-ai/?utm_source=chatgpt.com) Building a Zero-Trust Architecture for Confidential AI Factories By Hema Bontha, Manuel Huber, and Matheen Raza https://developer.nvidia.com/blog/building-a-zero-trust-architecture-for-confidential-ai-factories/ (https://developer.nvidia.com/blog/building-a-zero-trust-architecture-for-confidential-ai-factories/?utm_source=chatgpt.com) Hardware-Rooted AI Security That Won't Slow You Down By Sheel Pethe, Vidhya Krishnan, Aruna Manjunatha, Matheen Raza, and Jamie Li https://developer.nvidia.com/blog/?p=119469 (https://developer.nvidia.com/blog/?p=119469&utm_source=chatgpt.com) Hardware-backed Security at Global Scale: Confidential G4 with NVIDIA Blackwell Is GA By Meghna Sreenivasan and Sam Lugani https://cloud.google.com/blog/topics/ai-infrastructure/confidential-g4-gpus-are-ga (https://cloud.google.com/blog/topics/ai-infrastructure/confidential-g4-gpus-are-ga?utm_source=chatgpt.com) Verifiable, Private AI: Google Cloud Expands Confidential Computing Frontiers By Sam Lugani and Ranjit Narjala https://cloud.google.com/blog/products/identity-security/verifiable-trust-in-the-ai-era-whats-new-in-confidential-computing/ (https://cloud.google.com/blog/products/identity-security/verifiable-trust-in-the-ai-era-whats-new-in-confidential-computing/?utm_source=chatgpt.com) Attestation-Gated Key Release for Confidential Computing Workloads By Confidential AI Engineering Team https://confidential.ai/blog/attested-key-release (https://confidential.ai/blog/attested-key-release?utm_source=chatgpt.com) How to Secure Edge AI in Customer-Owned Environments By Microsoft Security Team https://www.microsoft.com/en-us/security/blog/2026/09/04/secure-edge-ai-customer-owned-environments/ (https://www.microsoft.com/en-us/security/blog/2026/09/04/secure-edge-ai-customer-owned-environments/?utm_source=chatgpt.com) Framework for Continuous Remote Attestation By MITRE Confidential Computing Layered Attestation Working Group https://www.mitre.org/news-insights/publication/framework-continuous-remote-attestation (https://www.mitre.org/news-insights/publication/framework-continuous-remote-attestation?utm_source=chatgpt.com) Confidential Inference Systems: Design Principles and Security Risks By Anthropic https://assets.anthropic.com/m/c52125297b85a42/original/Confidential_Inference_Paper.pdf (https://assets.anthropic.com/m/c52125297b85a42/original/Confidential_Inference_Paper.pdf?utm_source=chatgpt.com) TEE-Assisted Computation over Ciphertext: A Review By Yunyi Fang, Hui