The deal started the way a lot of software acquisitions do: with a customer who could not stop using the product. Inside Databricks, the finance team had begun running its models on Row Zero, a Seattle startup that built a spreadsheet able to hold more than a million live rows without crashing. On Sept. 24, the company decided to buy the product outright rather than keep renting it.
Databricks said it has acquired Row Zero, a company founded by former AWS and Tableau engineers, and will fold its technology into Genie, the AI coworker Databricks sells to help business teams turn data into answers. The companies did not disclose terms. Row Zero raised $10 million in May 2025 at a valuation of roughly $40 million, according to PitchBook, from a group that included Wes McKinney, the creator of the pandas data library.
The logic of the deal is straightforward. Genie can answer questions and run analysis, but business users still do most of their work in a grid. Row Zero gives Genie a spreadsheet interface where finance, operations, sales and marketing teams can explore data, model scenarios and collaborate, all inside the same governed environment Genie already draws on. Genie, which Databricks has positioned as an agent that works on a company’s own data rather than a general-purpose chatbot, is the product the acquisition is meant to strengthen.
Row Zero was founded in 2022 and has spent four years on the problem of making a browser spreadsheet behave like a database. It works like Excel or Google Sheets but connects directly to a data warehouse. When a source record changes, the spreadsheet refreshes; when a user finishes an analysis, the results can be pushed back to the cloud platform. A built-in AI assistant can summarize and visualize data without the user writing a formula. The pitch is a familiar tool that holds far bigger data than the spreadsheet on most desktops.
Breck Fresen, Row Zero’s founder and chief executive, described the acquisition as a step toward “the spreadsheet they deserve: secure, connected, scalable, and deeply integrated with AI agents.” Databricks chief executive Ali Ghodsi wrote on X about how the deal came together, recounting the version of events in which his own finance team discovered the product first.
The purchase is the latest move in Databricks’ effort to round out Genie, and the company said it is still scouting for more acquisition targets. That fits a broader pattern in which data-platform vendors buy small, focused startups to fill gaps in their agent products. Spreadsheets, in that picture, are the missing piece between an AI that answers questions and a tool businesses actually use to run their numbers.
Why spreadsheets matter is no mystery to anyone who has sat through a finance meeting. Excel and Sheets break down when a workbook runs past a few hundred thousand rows, forcing analysts to export flat files and pass them around. Row Zero’s argument is that the same formulas can work at a thousand times the scale, with cloud security baked in, and that the interface people already know is the one most likely to survive.
Governance is the part Databricks emphasized. Because Row Zero sits on top of the warehouse and inherits its permissions, the numbers an agent touches stay auditable, which matters for the finance and regulated teams Genie is aimed at. Databricks frames the result as a “live, governed spreadsheet” experience tied to its Genie Ontology, the shared business context the agent uses to ground its answers.
The startup’s technology also plugs into other AI systems. Row Zero’s workbooks implement MCP, the model-connection standard, and let enterprise customers bring their own key to attach the spreadsheet to AWS Bedrock, OpenAI, Anthropic or Google Gemini. Data, the company has said, never sits outside the customer’s own cloud.
The move reads as a response to a crowded field. Snowflake, Microsoft and Google are each pushing combinations of agents and business-intelligence tools, all chasing the same outcome: an AI that runs an analysis and then hands the result to a business user in a form that user already understands. Databricks is betting that the spreadsheet, not the chat window, is where that handoff happens.
Analysts described the purchase as a way to close the loop between an agent and the documents business teams actually touch. An AI that answers a question in a chat window is useful; an AI that can drop its answer into a live, auditable spreadsheet and let a controller sign off is something a finance department can put in front of an auditor.
Databricks has spent years trying to own the layer between raw data and the people who use it. The company, one of the most valuable private software firms, built its name on the lakehouse and then pushed into AI with Genie as the front door. Buying a spreadsheet maker extends that door into the grid where much of corporate finance still actually happens.
For now, the terms stay hidden and the product stays small. Row Zero’s team will join Databricks to build the spreadsheet experience into Genie across the major clouds. The test is whether a humble grid can carry the weight of an AI coworker, and whether finance teams trust an agent enough to let it write in the margins.
