30+ sources. Zero spin.
Cross-referenced, unbiased news. Both sides of every story.
Google Upgrades NotebookLM with Gemini 3.5, Cloud Computing, and Web-Based Source Discovery

What Google Actually Shipped
Google updated NotebookLM on Monday, June 8, 2026 — and it's a meaningful upgrade, not a marketing rebrand.
The biggest changes, according to The Verge and TechCrunch: the app now runs on Gemini 3.5 instead of its previous model, it connects every notebook to a "secure cloud computer" via Google's Antigravity platform, and it can now help users build a source library from scratch using Google Search — instead of requiring you to manually import everything yourself.
Before this update, NotebookLM required you to bring your own documents, PDFs, or YouTube links. You couldn't start with nothing. Now you can open a chat, describe a research topic, and NotebookLM will suggest sources from the web that you can then choose to import. TechCrunch notes this could help users surface sources in other languages or find work from related authors they didn't know existed.
The Antigravity Cloud Computer
Each notebook is now connected to what Google calls a "secure cloud computer" running on Antigravity — the company's agentic coding platform. That means NotebookLM can write and execute actual code as part of the research process, not just summarize text.
This represents a shift from a tool that reads your documents to one that can analyze data inside them.
The Verge reports that NotebookLM can now output research in a significantly expanded set of formats: PDFs, Word documents, Markdown, CSV sheets, JSON files, Excel spreadsheets, PowerPoint files, PNG and SVG data visualizations, and images generated through Google's Nano Banana image tool. You can also edit outputs after the fact.
TechCrunch adds that NotebookLM will now show users the detailed reasoning steps behind its answers — a transparency feature that lets you audit how it arrived at a conclusion rather than just accepting the output.
Who Gets It — and Who Doesn't
All of these features — Gemini 3.5, the cloud computer, web-based source discovery, the expanded export formats — are available today only to Google AI Ultra subscribers and Workspace business customers with AI Ultra Access or AI Expanded Access, according to both sources.
Google AI Ultra runs $249.99 per month.
Google says it plans to expand access to other plans "in the future." No timeline given. No specifics offered.
For a tool that Google has marketed as a democratizing research assistant since its 2023 launch, locking its best features behind a $250/month paywall is a notable tension. The average student, journalist, or independent researcher won't see these features anytime soon.
What the Coverage Is Missing
Both The Verge and TechCrunch covered the features adequately. Neither spent meaningful time on the pricing wall or what it means for NotebookLM's original positioning as an accessible research tool.
The coverage also doesn't address the competitive parallel: Google is essentially building what Microsoft has been trying to do with Copilot in Office 365 — AI-assisted document creation with cloud execution. The difference is that Microsoft baked Copilot into tools millions of businesses already pay for. Google is charging separately, and charging a lot.
There's also no independent verification of Google's "more accurate and reliable" claim about Gemini 3.5. That's a Google blog post assertion, not a benchmark result. Take it with appropriate skepticism until third-party evaluations exist.
What This Means
NotebookLM's June 8 update is substantive. Moving from passive document summarization to active web research and code execution is a real capability leap.
But Google is charging enterprise prices for what it once framed as a tool for everyone. If you're paying $250 a month, this update gives you something useful. If you're not, you're still waiting.
Google built something interesting. Then it put it behind a velvet rope.