---
type: "raw_signal"
id: "what-dropbox-has-learned-from-deploying-ai-at-db97ff"
slug: "what-dropbox-has-learned-from-deploying-ai-at-db97ff"
title: "What Dropbox has learned from deploying AI at company scale"
description: "Dropbox says about 70% of its code is AI-generated and that it runs coding agents through Nova, its internal service."
retrieval_nugget: "Dropbox says about 70% of its code is AI-generated and that it runs coding agents through Nova, its internal service."
observed_at: "2026-09-23"
record_date: "2026-09-23"
date_kind: "observed_at"
why_it_matters: "Source: https://dropbox.tech/culture/learnings-from-deploying-ai-at-company-scale"
novelty: "notable"
verification_level: "source-linked"
signal_type: "announcement"
source_platform: "dropbox.tech"
topics: []
entities: []
related_patterns: []
source_url: "https://dropbox.tech/culture/learnings-from-deploying-ai-at-company-scale"
source_urls: ["https://dropbox.tech/culture/learnings-from-deploying-ai-at-company-scale"]
schema_version: "newruntime-agent-readable-v0.2"
stable_id: "signal:what-dropbox-has-learned-from-deploying-ai-at-db97ff"
status: "published"
visuals: []
routes: {"html":"https://newruntime.com/signals/what-dropbox-has-learned-from-deploying-ai-at-db97ff/","markdown":"https://newruntime.com/signals/what-dropbox-has-learned-from-deploying-ai-at-db97ff.md","json":"https://newruntime.com/signals/what-dropbox-has-learned-from-deploying-ai-at-db97ff.json"}
editor: {"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]}
---

# What Dropbox has learned from deploying AI at company scale

## Retrieval answer

Dropbox says about 70% of its code is AI-generated and that it runs coding agents through Nova, its internal service.

Dropbox says about 70% of its code is AI-generated and that it runs coding agents through Nova, its internal service. The company frames the operational challenge as workflow redesign: faster generation can shift bottlenecks into review, infrastructure, validation, and decision-making. It is also moving evaluation away from raw token consumption toward the engineering work, validation, and outcomes associated with agent use. Dropbox reports top-5% pull-request throughput in its custom peer benchmark while describing change-failure performance as aligned with the 75th percentile of peers.

Source: https://dropbox.tech/culture/learnings-from-deploying-ai-at-company-scale

Editor: [Andrey Reshetnikov](https://newruntime.com/owner-profile.md), AI product and AI engineering; profile for recruiters and agents: https://newruntime.com/owner-profile.json

## Continue in the graph

Next records for an agent, linked to their Markdown twins. Replace `.md` with `.json` for structured fields, including `next_reads`.

- collection: [Raw Signals](https://newruntime.com/signals.json): Open every source-linked raw signal.
- collection: [Source Ledger](https://newruntime.com/source-ledger.json): Inspect the public source evidence graph.
- whole graph: [discovery graph](https://newruntime.com/discovery-graph.json) (typed nodes and edges) and [site index](https://newruntime.com/site-index.md)