Zoom and NVIDIA announced a strategic partnership to jointly deliver faster, higher-quality, and more customizable AI services to enterprise users. This collaboration will enhance the application capabilities of Zoom AI Companion 3.0 in data security-critical industries such as finance, healthcare, and government by integrating NVIDIA Nemotron open-source technology.
Hybrid AI Architecture: Intelligent Allocation of Computing Resources
Zoom’s innovative federated AI architecture will take this to the next level, intelligently distributing queries across different models:
• For simple tasks: Use Zoom’s proprietary Small Language Models (SLMs) to achieve low latency and high quality performance
• Complex inference requirements: Invoking finely tuned large language models (LLMs), such as the 490 billion parameter model developed based on NVIDIA Nemotron.
• This hybrid architecture strikes the best balance between cost, quality, and speed, bringing a more efficient collaboration experience to enterprises.
Technology Integration Highlights
Zoom's Chief Technology Officer, XD Huang, stated, "With the support of NVIDIA Nemotron open-source technology, we are enhancing our enterprise-grade search augmentation generation (RAG) capabilities, enabling AI Companion to seamlessly integrate with platforms such as Microsoft 365, Google Workspace, and Salesforce."
Key areas of cooperation include:
• Develop custom models using NVIDIA NeMo tools
• Improve AI reasoning speed and decision-making efficiency
• Strengthen cross-platform integration capabilities
AI deployment that prioritizes privacy and security
Zoom places particular emphasis on responsible AI development principles:
• Never use customer audio, video, chat, or other content to train AI models.
• Provide security deployment solutions for sensitive industries such as finance, healthcare, and government.
Safety and privacy are considered core principles of generative AI.
Kari Briski, NVIDIA's Vice President of Generative AI Software, stated, "Our collaboration with Zoom is just beginning, and we will continue to bring smarter and more efficient AI inference services to our customers in the future."
This collaboration marks the beginning of a new era of customization for enterprise-grade AI collaboration tools, providing high-performance and secure intelligent solutions for various industries through a combination of hybrid architecture and open-source technology.
