TL;DR: Google is rolling out Gemini for BigQuery data in Google Sheets. In a sheet connected to BigQuery through Connected Sheets, you can ask questions in plain English, and Google says “Gemini queries run against the full dataset in BigQuery, not just what’s visible in your spreadsheet.” Gemini can also build refreshable dashboard tabs. It’s on by default, and Google says “there is no admin control for this feature.” BigQuery’s default on-demand pricing charges by the data each query processes, so if your team has Connected Sheets, check who can query which tables and set a daily query quota before the feature reaches everyone.

What happened

Google announced the feature on its Workspace Updates blog on October 19, 2026. It works in spreadsheets that are already connected to BigQuery through Connected Sheets. Per Google’s post, you can:

  • Ask in plain English. For example, “What were total sales in the Northeast last quarter?”, with the answer based on your complete BigQuery data.
  • Ask follow-ups. For example, “How does that compare to the same quarter last year?”
  • Go from question to dashboard. Describe a report, and Gemini “will share its plan for you to review, then build a new spreadsheet tab with native, connected, refreshable objects like database formulas, pivot tables, and charts.”

Who gets it: Business Standard and Plus, Enterprise Standard and Plus, Google AI Pro and Ultra, Google AI Pro for Education, and the AI Expanded Access add-on. Business Starter isn’t on the list.

When: a gradual rollout for both Rapid Release and Scheduled Release domains that started October 19, 2026. Google says it can take “up to 15 days for feature visibility.”

Controls: “This feature will be ON by default. There is no admin control for this feature.” It shows up for “users with permission to query the connected data,” in the Gemini side panel of a connected spreadsheet.

Why it matters

Connected Sheets already lets people analyze BigQuery tables with pivot tables and charts. As the feature rolls out, anyone with query access and an eligible plan can simply ask Gemini, and Gemini works out the queries. That’s the point of the feature, but it also makes it easier for more people to run more queries.

Queries cost money on the default plan. Google’s BigQuery pricing page says on-demand pricing, the default, charges “for the number of bytes processed by each query.” The first 1 TiB each month is free, and the page lists $6.25 per TiB after that in the US and EU multi-regions; some regions cost more. With Connected Sheets, Google says, queries run “on your behalf either when manually requested or on a defined schedule,” and Connected Sheets needs “a project with billing setup in BigQuery.” Google’s post doesn’t say how Gemini’s queries are billed or how many queries a single question can trigger, so assume they count like any other Connected Sheets query until Google says otherwise.

Refreshable dashboards keep querying. A dashboard tab that Gemini builds is made of connected, refreshable objects. If someone puts it on a refresh schedule, it runs queries on that schedule.

The default caps are high. Google’s quota docs say each project can process 200 TiB a day by default, and there’s no per-user limit unless you set one. By our math, 200 TiB at the listed $6.25 per TiB would be about $1,250 a day for one project. That’s a per-project ceiling, not a likely bill: it’s higher in regions that charge more, and each project has its own limit. It’s also the only built-in stop if nobody sets a lower one.

If your company pays for BigQuery capacity (slots) instead of on-demand, the cost works differently, and Google notes that query usage quotas apply only to on-demand pricing.

What to do

  1. Find out if you use Connected Sheets. If nobody on your team has connected a sheet to BigQuery (Data > Data connectors > Connect to BigQuery), there’s nothing for Gemini to query yet. Anyone with BigQuery access can connect one, though, so if you use BigQuery at all, steps 2 and 3 still apply.
  2. Check who can query what. Google says the feature is available to “users with permission to query the connected data,” so review who has query access to your large tables. There’s no Workspace admin switch to turn it off.
  3. Set a daily query quota. On the Quotas & System Limits page of the Google Cloud console, someone with the Quota Administrator role can lower QueryUsagePerDay (the whole project) and set QueryUsagePerUserPerDay (one limit that applies to every user and service account in the project). Google says the quota is “proactive, so you can’t run an 11 TB query if you have a 10 TB quota,” and daily quotas reset at midnight Pacific Time.
  4. Track Connected Sheets usage. Google says every Connected Sheets query has a job ID starting with sheets_dataconnector, and its docs include a ready-made query on BigQuery’s INFORMATION_SCHEMA jobs view (JOBS_BY_ORGANIZATION) that ranks users by bytes billed over the last 30 days (replace REGION_NAME with your region, for example region-us). Run it weekly during the rollout.
  5. Review scheduled refreshes. Ask people to check the refresh schedule on any dashboard tab Gemini builds, and to turn off schedules nobody reads.
  6. Know where the audit trail is. Google says access to BigQuery data from Connected Sheets is recorded in Cloud Audit Logs, which show who accessed the data and when.

Teams weighing other AI tools that bill by usage can find which settings actually stop spending, and which only alert you, in our guide to hard spending limits for AI agents. For another AI billing change that admins need to act on this month, see our story on Notion’s new AI credits.

Sources & method

Sources

Method. Last checked October 20, 2026. Features, plans, dates, prices, quotas and quotes come from Google’s Workspace Updates post, Help Center and BigQuery documentation. The $1,250-a-day, per-project figure is our arithmetic from Google’s default quota and listed price. We haven’t tested the feature ourselves. No affiliate links on this page. How we work.

Spotted an error? See our corrections policy and tell us.