Pricing snapshot
Sigma Computing is a spreadsheet interface on top of your cloud data warehouse. Plan its cost in two parts: what you pay Sigma, and the warehouse and AI model usage it generates on your own accounts.
Scroll sideways to see all columns. With a keyboard, focus this table and use the arrow keys.
| Item | What applies |
|---|---|
Sigma price | Ask the Sigma team through its contact page or a demo request. |
Free trial | Offered on Sigma's homepage. |
Queries | Run live on your cloud data warehouse; Sigma keeps no separate data store. |
AI processing | Runs on your warehouse compute, Sigma states, using AI models you connect. |
How usage reaches your warehouse
Sigma's architecture page puts it plainly: "You pay for queries. Not curiosity." When someone opens or changes a workbook, Sigma tries cheaper paths first: results already in the browser session, then re-sorts, filters and pivots on data already loaded, then a cached result fetched by warehouse query ID if your platform caches results, then the warehouse's own result cache, which most platforms hold for up to about 24 hours. Only then does warehouse compute run, and Sigma merges a page's tables, charts and pivots into fewer warehouse hits instead of one per element.
In budget terms, repeat views of unchanged data can avoid new warehouse compute; new questions and changed data are what reach it.
What adds usage
Agents. Sigma Agents can run on a schedule, answer questions from your governed warehouse and, with your approval, write results back to it. Admins see every agent people build and the tokens it uses, so AI model spend can be traced to individual agents.
Materialized tables. For repeat-heavy logic, Sigma suggests materializing expensive datasets into reusable warehouse tables refreshed on a schedule. That swaps repeated query compute for scheduled refreshes and stored tables in your warehouse.
Write-back. Input tables save edits to warehouse tables with INSERT, UPDATE or DELETE, so data entry in Sigma is also warehouse work.
Ask before you sign
Embedded use. Sigma offers white-label embedded analytics. If you plan to embed Sigma in your own product, ask how embedded use is priced.
Trial terms. Confirm the free trial's length and conditions before you plan a test workload around it.
Contract scope. Get the price in writing, with the features and services it includes, any minimum commitment and the renewal terms.
Measure usage during the trial
Query History shows the SQL Sigma generated for each workbook element, with queue time, Sigma runtime, warehouse runtime, result fetch time and request IDs for warehouse tuning. Usage dashboards show which workbooks and people drive the load. Run representative workbooks, schedules and agents during the trial, then compare these figures with your warehouse and AI model bills for the same period.
Take
Choose Sigma when your data already lives in a cloud warehouse and you want spreadsheet-style analysis, apps and agents on it without copying tables out. If your data is not in a warehouse yet, budget for one first: Sigma runs on the warehouse, not on a data store of its own.
Evaluate this price
Sources
Sigma homepage, contact page and architecture page, observed October 26, 2026.