On June 9, Google officially rolled out a major update to its AI research and note-taking tool, NotebookLM. The underlying model of the tool has been natively switched to Gemini 3.5 and deeply integrated with the agent Antigravity, marking NotebookLM's evolution from an early "document reading and content organization tool" into a full-stack AI research assistant.

gemini, Google

The most core breakthrough in this upgrade is the introduction of a secure cloud computing environment, allowing NotebookLM to directly write and run code, perform in-depth data analysis, and generate charts in real time within the notebook. Previously, users such as market analysts and academic researchers had to frequently switch between multiple software applications when completing data cleaning, trend analysis, and visualization. The new version of the tool enables closed-loop operations throughout the entire process and supports direct output of over a dozen mainstream file formats, including PNG, SVG, PDF, Markdown, Excel spreadsheets, and PowerPoint presentations.

From a strategic perspective, Google currently adopts a tiered access strategy for payment. New features are initially available only to AI Ultra subscribers and some enterprise Workspace clients. Although internal assessments by Google show that the new system has an average success rate of over 65% in core dimensions such as large document analysis and advanced web research, the commercial trial of its advanced capabilities reflects how tech giants are accelerating the transformation of "lightweight free tools" into assetized, high-barrier paid office entry points. This move is not only a test of users' willingness to pay but also signals a deep evolution in the AI application layer from "light interaction Q&A" to "heavy workflow end-to-end production."