Zoom Communications, Inc. and NVIDIA are collaborating to provide businesses with AI that is quicker, better, and more adaptable. NVIDIA Nemotron open technologies are added to Zoom’s federated architecture to enable AI Companion 3.0 in sectors like government, healthcare, and finance.
This framework is growing to include a next-generation hybrid language model approach, a ground-breaking AI architecture that can intelligently route queries between a refined Large Language Model (LLM) for complex reasoning and Zoom’s proprietary Small Language Models (SLMs), which are optimized for low latency and quality for a specific skill or tasks.
The new hybrid model will maximize cost effectiveness, quality, and latency while accelerating enterprise productivity and collaboration experiences.
A federated architecture is used by Zoom’s AI framework to autonomously choose the optimal AI model for every task. By using this strategy, Zoom AI Companion may take advantage of NVIDIA’s cutting-edge AI software, services, and infrastructure in addition to other models, providing customers with improved capabilities at the lowest possible cost.
This includes Zoom’s new 49-billion-parameter LLM, which was created using NVIDIA NeMo tools and is based on NVIDIA Nemotron to provide the best possible balance between accuracy, speed, and cost.
With the help of this cutting-edge architecture, government agencies and business clients may take advantage of both open and closed model innovation, which results in increased cost effectiveness, quicker AI workflows, deeper reasoning capabilities, and more effective cooperation inside AI Companion.
The next iteration of Zoom’s federated AI architecture, which dynamically integrates optimal models—such as the Llama Nemotron Super-based reasoning model—to provide the best possible mix of accuracy, performance, and cost, will be powered by this endeavor.
High-quality performance for real-time transcription, translation, and summarization has already been demonstrated by Zoom’s patent-pending federated AI method. With the NVIDIA open model developments, Zoom’s remarkable performance is extended by this new method.
“We’ve increased our speed and enhanced lower-cost model decision making using NVIDIA GPUs and AI software stack, helping to optimize AI Companion’s core capabilities and enable faster go‑to‑market timelines,”
said X.D. Huang, Chief Technology Officer at Zoom.
“With the help of NVIDIA Nemotron open technologies, we’re accelerating the development of our enterprise retrieval-augmented generation (RAG) capabilities, allowing AI Companion to work seamlessly with Microsoft 365, Microsoft Teams, Google Workspace, Slack, Salesforce, and ServiceNow. This partnership allows us to deliver powerful, security-focused, and scalable AI experiences to our customers at rapid speed.”