Databricks Bought Row Zero for the Screen Where People Check an Agent's Work

The Tool Databricks' Own Finance Team Was Already Using

Databricks announced on September 24 that it has acquired Row Zero, a Seattle spreadsheet startup. It did not disclose the price or terms. Row Zero was founded in 2021 by former AWS and Tableau engineers. Two of its founders, CEO Breck Fresen and Nick End, had earlier built Shoefitr, a shoe-fitting startup that Amazon bought in 2015.

Row Zero's pitch is scale. A Microsoft Excel worksheet tops out at 1,048,576 rows. Databricks says Row Zero lets business users work with billions of rows at interactive speed. TechCrunch reported that Row Zero raised $10 million in May 2025 at a valuation PitchBook estimated at about $40 million. GeekWire described that round as a Series A.

The deal started inside Databricks. According to TechCrunch, the company's finance team had taken to Row Zero because it could handle live spreadsheets well past a million rows, and was pairing it with Genie, the AI agent Databricks built in-house. When executives asked the finance team what the tool was, the idea took shape. CEO Ali Ghodsi wrote on X that spreadsheets are an interface every business analyst loves, so marrying BI, agents and spreadsheets makes sense.

A Place for People to Recheck What Agents Touched

The problem Databricks put first in its announcement is the loose spreadsheet file. Business teams have long parked data in ungoverned spreadsheet files, and people and agents pass those files around. Every export, the release says, adds another unmanaged data pipeline inside the company. Databricks' argument is that once agents analyze data and act on it, a company can't let them run on data of unknown lineage and quality.

Row Zero replaces the download with a spreadsheet wired straight to the source. Every query honors the individual user's permissions, and data refreshes automatically from the live source. Exports can be locked down, and results can be written back to the source. Every interaction can be audited.

The release singles out one more point about agents. Because agents work on the spreadsheet directly through Row Zero's tools, business teams can read and check each agent action using the same formulas and processing model they already know. In an email to GeekWire, End said that even as AI agents take on more analysis, people still need a familiar interface to verify the results.

Databricks co-founder Patrick Wendell said that as more finance, operations and go-to-market teams adopt Genie, a seamless spreadsheet experience has become a must-have. Row Zero will be built into Genie on the web, desktop and mobile, so users can move from a chat to hands-on pivoting and modeling.

A Notebook in March, a Spreadsheet in September

On its own, this is one spreadsheet company changing hands. Line it up with Databricks' other purchases this year and a pattern shows. SiftD.ai, a deal that closed on March 23, had only launched its product in November 2025. TechCrunch described that product as an interactive notebook, similar to Jupyter, where people and agents work together. Row Zero's own release calls its product the spreadsheet built for humans and agents to work together with data. Within six months, Databricks bought two tools that put people and agents on one screen.

Agents run through the other deals too. Panther, which Databricks agreed to buy on June 16 and closed on August 3, is an AI security operations platform that runs agentic detection and response. Electric, which joined on August 11, is the team behind PGlite, a Postgres database small enough to run inside a browser tab or an agent sandbox. Databricks says PGlite's weekly downloads grew from 1 million to 13 million in twelve months.

The money to keep going is there. On August 13, Databricks said it had passed a $7 billion revenue run-rate, growing more than 80% year over year in its second quarter. The same day it said it had closed a $5 billion round at a $190 billion valuation and had delivered positive adjusted free cash flow over the prior twelve months. Ghodsi told TechCrunch the company intends to do many more acquisitions like this one.

How Far the Standalone Promise Reaches

Row Zero's current customers will want to know what happens to the product. GeekWire reported that Fresen wrote on LinkedIn that the standalone product will remain available to existing users. Databricks' release says Row Zero will keep supporting data sources beyond Databricks. All Databricks customers will be able to use it across the major clouds.

The wording is narrow. It covers existing users and gives no end date. The team's center of gravity is moving, too. Fresen said the team will focus on embedding its technology across Databricks, and End said most of the team will work out of Databricks' Seattle office. A company running Row Zero alongside a different data platform should ask, before its next renewal, which side will get new features first.

The price is no benchmark either. It wasn't disclosed, and the only public figure is PitchBook's estimate of about $40 million from the 2025 round. That is not the purchase price. What founders running similar companies can take from this deal is not a number but what the buyer set out to get.

The Acquirer Was a Customer First

The SiftD.ai and Row Zero deals share one more trait: Databricks knew the team or the product before it bought them. TechCrunch reported that the Databricks team knew SiftD.ai's CEO, Steve Zhang, from his years as chief scientist at Splunk. Row Zero was a tool Databricks' finance team was using in its daily work.

For teams building tools that sit on top of a platform, that order of events is worth noting. Once a department at the platform company relies on your product every day, the conversation starts with the buyer already knowing what the product is worth to them. Acquisition candidates may surface in a platform's internal workflows before they show up in a pitch meeting.

Where the product sits points the same way. Databricks bought Row Zero so that business teams could read and check what an agent did. Startups selling agents to enterprises can treat the question of where, and in what format, a customer's staff reviews the agent's output as one axis of product design. Row Zero answered it with the format nearly every office worker already knows: the spreadsheet.