Nvidia discusses Confidential Computing for secure AI data processing

Summary

NVIDIA and VAST Data are emphasizing the importance of protecting enterprise data while leveraging AI capabilities, as highlighted in a discussion featuring NVIDIA's @j_boitano and VAST Data's Renen Hallak. They introduce the concept of NVIDIA Confidential Computing, which allows enterprises to maintain confidentiality over their sensitive data and proprietary models during AI processing. This initiative aligns with VAST Data's recent launch of DataEnclave, designed to enhance their VAST DataEngine and AI Operating System by enabling secure, hardware-protected AI inference.

Tokens

$NVDA

Analysis

NVIDIA: NVIDIA develops accelerated computing platforms and software for AI workloads, including Confidential Computing features that use encrypted memory and trusted execution environments to protect data and models during processing. In the context of this news, NVIDIA's technology is central to enabling enterprises to run AI inference on sensitive data without exposing proprietary information or model weights. Justin Boitano, NVIDIA's Vice President of Enterprise AI, participated in the discussion highlighting these capabilities alongside VAST Data. VAST Data: VAST Data provides data infrastructure positioned as an AI Operating System for handling large-scale AI workloads, including storage, databases, and now confidential computing capabilities. In the context of the news, VAST Data announced DataEnclave, a new runtime built on NVIDIA Confidential Computing that creates hardware-isolated environments for running AI models against sensitive data while protecting both business information and model IP. The company collaborates with NVIDIA and partners to deliver this in customer data centers and trusted clouds. Renen Hallak: Renen Hallak is the Founder and CEO of VAST Data, where he leads development of the company's AI Operating System platform. In this news, Hallak discusses with NVIDIA executives how VAST DataEnclave leverages NVIDIA Confidential Computing to allow secure processing of enterprise data with proprietary AI models. His comments emphasize managing models as core operating system resources within trusted boundaries. Collaboration: NVIDIA and VAST Data are collaborating to enable secure, hardware-protected AI inference that keeps enterprise data and proprietary models confidential during processing. Product Launch: VAST Data introduced DataEnclave as the confidential AI capability of its VAST DataEngine and AI Operating System, built directly on NVIDIA Confidential Computing. Enterprise Focus: The discussion highlights solutions for enterprises to bring AI capabilities to sensitive data in protected environments without exposing business information or model intellectual property.

Categories

techaimachine_learning

Related sources

View Original Tweet