Dropbox is buying cloud storage startup Tigris Data for $350 million, the biggest acquisition in the company’s history, and using the deal to launch a new storage platform built specifically for AI agents instead of regular files. The announcement came Thursday alongside “DBX Agent Cloud,” Dropbox’s bet that the next wave of cloud storage customers will be bots, not people. The deal is expected to close in the fourth quarter.
Dropbox confirmed the acquisition in a release posted Thursday morning, calling it part of an “AI-Powered Product Evolution.” Tigris Data was founded by a group of former Uber engineers who helped build the ride-hailing company’s internal storage infrastructure, and it has spent the last few years building cloud storage designed to sit physically closer to the people and machines using it, cutting down on the lag that comes from pulling data across long distances. The startup raised a $25 million Series A last year and had largely stayed under the radar until now.
That makes $350 million a striking number for a company most people have never heard of, and it says a lot about how much storage infrastructure has become a battleground in the AI race. Dropbox told reporters the deal is expected to close in the fourth quarter of this year and said flatly that it has no plans to cut staff because of the acquisition, a notable thing to clarify given that the company trimmed roughly 6% of its global workforce just last month.
What DBX Agent Cloud actually does
The product at the center of the announcement, DBX Agent Cloud, is not aimed at the people who already use Dropbox to back up photos or share work files. It is built for developers and companies building AI models and autonomous agents, and it is launching first as an object storage product compatible with tools already built for Amazon S3, which should make it easy for developers to plug in without rewriting their existing code.
Dropbox says the roadmap goes further than basic storage. It plans to add support for documents, images, audio, and video, along with filesystem features that let AI agents manage their own files and keep continuity across tasks and sessions. That last part matters more than it might sound. One of the quiet problems with agentic AI right now is that most agents forget everything the moment a session ends, which makes it hard to trust them with anything that spans more than a single conversation. A storage layer that lets an agent pick up where it left off, days or weeks later, is a real engineering gap, not a marketing buzzword.
It also fits into Dropbox’s broader AI push, which already lets users ask an AI assistant to work directly with their files, choose which underlying model handles a given task, and connect their Dropbox content to outside tools like ChatGPT, Claude, and Muse. Dropbox says more than 70,000 teams are now using this updated AI experience. The initial rollout of the new storage features will go to eligible Standard and Advanced customers in the US, UK, and Australia, with a wider release planned for 2027.
A pattern, not a one-off
This is not Dropbox’s first AI-related purchase. The company has previously picked up Reclaim.ai, Promoted.ai, and Mobius Labs, each a smaller, more targeted deal than this one. The Tigris purchase is a different scale entirely, and it reads less like a bolt-on feature and more like Dropbox trying to reposition itself as infrastructure for the agent economy rather than just a place to store a backup of your laptop.
That instinct is not unique to Dropbox, and the timing is not a coincidence either. The same morning Dropbox announced the Tigris deal, cloud storage rival Box was separately unveiling its own agent-focused retrieval system built on Nvidia’s Nemotron embedding technology, with Box CEO Aaron Levie telling reporters the goal was to send agents back the exact chunks of a document they need instead of forcing them to read entire files. Security vendor CrowdStrike made a similar calculation when it spent nearly $500 million buying Fiddler AI and Dreadnode to build out tools for policing AI agents rather than just the networks around them. Three different companies, three different parts of the stack, all racing toward the same conclusion at once: whoever controls the plumbing underneath AI agents, whether that’s security, storage, or retrieval, ends up with a much stickier product than whoever just sells a chat interface on top of someone else’s infrastructure.
And agents are getting expensive to run at scale, which is exactly why storage efficiency is becoming its own competitive front. Y Combinator’s Garry Tan said earlier this year that he personally spends up to $200,000 a year on AI agent tokens, a figure that sounds absurd until you consider how many startups are now running fleets of agents around the clock. Every inefficiency in how an agent stores, retrieves, or re-fetches data turns directly into a bigger bill. A storage platform purpose-built to keep agents from re-reading the same files over and over, or losing their place between sessions, is a cost argument as much as a technical one.
There’s also a simpler reason Dropbox needed a swing this big. The company has spent the past few years fending off a narrative that it’s a legacy file-sync business in a world that has mostly moved on to Google Drive, Microsoft OneDrive, and a dozen AI-native note and knowledge tools. Buying a storage startup built from the ground up by engineers who scaled infrastructure at Uber, then wrapping it in a developer-facing product with its own brand name, is a way of telling investors and customers that Dropbox still has a technical edge worth paying for. Whether $350 million was the right price for that story is something only the next few quarters of DBX Agent Cloud adoption numbers will answer.
The bottom line
Dropbox spending its largest acquisition ever on a two-year-old storage startup is a bet that the real money in AI infrastructure over the next few years will not be in building bigger models, but in building the boring plumbing those models and their agents need to actually get work done reliably. That is a reasonable bet. Storage, permissions, and continuity are unglamorous problems, but they are exactly the kind of unglamorous problems that make or break whether a company trusts an AI agent with anything that matters. The real test is not the press release. It is whether developers actually show up to build on DBX Agent Cloud instead of just defaulting to whatever storage their cloud provider already sells them, and whether Dropbox can convince the market it is an infrastructure company now, not just the place people dump old files they forgot to delete.
The Teck Hustlers Editorial Team covers technology, business, finance and gaming news for readers who want the story behind the headline. Every article is researched, written and edited by our in-house team, checked against company announcements, filings and reputable reporting, and updated when new facts emerge. We focus on what each development means for founders, professionals, investors and everyday users. Have a tip or a correction? Reach us through our contact page.
