October 31, 2025
Zoom Leverages NVIDIA to Put Enterprise AI on the Fast Track
Zoom is stepping up its AI game by teaming up with NVIDIA to make enterprise AI faster, smarter and more adaptable to the way teams actually work.
The latest iteration of Zoom AI Companion introduces a hybrid language model framework designed to balance speed, precision, and complexity, a move that could boost productivity for business users.
At its core, the hybrid system combines Zoom’s proprietary Small Language Models (SLMs) with a fine-tuned Large Language Model (LLM).
The idea is to let SLMs handle routine or highly specialised tasks quickly, while the LLM kicks in for more complex reasoning. The system routes queries intelligently between the two, optimising both latency and output quality.
A New Way to Scale AI
The partnership brings NVIDIA’s Nemotron technology into Zoom’s federated AI architecture, enabling AI Companion 3.0 (announced at Zoomtopia 2025) to be customised to individual user behaviours without compromising data privacy. Businesses in sectors like finance, healthcare, and government can now deploy AI tools that adjust to specific workflows, learning from user interactions while remaining secure.
What sets this approach apart is its focus on adaptability and context. Zoom’s AI generate summaries, automates repetitive tasks, and learns how teams work and delivers outputs aligned with their habits and priorities.
Meanwhile, NVIDIA’s models ensure this intelligence runs efficiently, even under heavy enterprise workloads, keeping performance high while managing costs.
This hybrid model offers a way for organisations to scale AI. Instead of relying solely on large, resource-heavy models, Zoom’s framework leverages smaller, specialised engines to handle everyday queries quickly, reserving the LLM for situations that demand deeper analysis. The result is a system that feels responsive, capable, and tuned to real-world enterprise needs.
X.D. Huang, chief technology officer at Zoom, said: “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.”
