Dropbox status: access issues and outage reports

Problems in the last 24 hours

The graph below depicts the number of Dropbox reports received over the last 24 hours by time of day. When the number of reports exceeds the baseline, represented by the red line, an outage is determined.

Most Reported Problems

The following are the most recent problems reported by Dropbox users through our website.

• Sign in (57%)
• Errors (43%)

Live Outage Map

The most recent Dropbox outage reports came from the following cities:

City	Problem Type	Report Time
Salt Lake City	Sign in	8 hours ago
Madrid	Errors	16 days ago
Conneaut	Sign in	1 month ago
City of London	Errors	1 month ago
Alpharetta	Sign in	2 months ago
Shreveport	Sign in	2 months ago

Full Outage Map

Community Discussion

Tips? Frustrations? Share them here. Useful comments include a description of the problem, city and postal code.

Beware of "support numbers" or "recovery" accounts that might be posted below. Make sure to report and downvote those comments. Avoid posting your personal information.

Dropbox Issues Reports

Latest outage, problems and issue reports in social media:

• 
  The Wisemen Alpha (@Wisemenmentors) reported 9 minutes ago

  Told properly for the first time @toly. Soviet Ukraine. 13 years at Qualcomm. Dropbox. 4 AM at Cafe Sole with two coffees and a beer. The moment you realized the problem wasn't consensus, it was time itself.

• 
  Infuse (@infuse) reported 13 minutes ago

  @Zvomuya This may be a temporary issue with Dropbox. If it's still happening can you try restarting your device to see if this helps?

• 
  Jasper Polak (@polak_jasper) reported 13 minutes ago

  Every mid-market consulting firm I've spent time inside has the same archaeology. Proposals from 2022 in a partner's Dropbox. Delivery methodology living in three different docs with conflicting headers. Sales call notes in Teams. Post-mortems in a notes app on an iPhone. Client health data in HubSpot. Margin in the finance spreadsheet. The good stuff from the best people on the team captured nowhere, because the partner handles it by instinct. Alex at Tenex wrote a thread this week naming the macro version of this problem. Engineering already has its AI brain (the *** repo). Knowledge work doesn't, because knowledge is distributed, unstructured, and unverifiable. Someone will build the generalized version. Payoff: "Robinhood for knowledge workers." Agree with almost all of it. The part I'd add for services firms: The corpus isn't the problem. You have more context than almost any other business type. Every engagement generates detailed artifacts. Every partner has fifteen years of calibrated judgment. Every proposal has a clear win/loss signal. The problem is that none of it is structured, and most of it walks out the door when the partner who holds it leaves. Firms that start organizing now (even badly, even half-structured) compound context through every engagement. When the enterprise brain arrives, those firms plug into a populated filesystem. Firms that wait plug into an empty one. The tool will get commoditized. The corpus won't. Start the archaeology today. Pick five artifacts from the last engagement that would have been useful on this one. Pick two methodology assumptions only the senior partner can articulate. Write them down somewhere your future brain can find them. The tool is coming. The corpus is what you'll plug into it.

• 
  Square Circle 556 (@ragtop25) reported 13 minutes ago

  @mistressdivy I hate OneDrive but I have to use it. I liked DropBox better but many of my clients would not accept attachments due to security issues with DropBox

• 
  Anna Longthorp (@AnnaLongthorp) reported 17 minutes ago

  Recently got my X account back after being hacked. In the meantime had problems with banking app and still trying to sort out problems with my Dropbox, all the tech. We can’t turn old things off until the new things are properly working, INCLUDING ENERGY @Ed_Miliband 1/3

• 
  Raziel (@tryraziel) reported 22 minutes ago

  Drew Houston got rejected by every major VC in Silicon Valley. Today Dropbox is worth $8B. Here's the pivot that changed everything. 2007: Houston was a frustrated MIT student who kept forgetting his USB drive. His solution? A file-syncing tool called Dropbox. The problem: VCs couldn't see the market. → "There's already FTP and email attachments" → "Why not just use a USB drive?" → "The market is too small" Paul Graham at Y Combinator was the only one who got it. But even he made Houston prove demand first. Houston's genius move: Instead of building the full product, he created a 3-minute demo video showing Dropbox syncing files across devices. The video went viral on Digg. Sign-ups jumped from 5,000 to 75,000 overnight. Suddenly VCs were calling him. The lesson: When investors can't see your vision, show them your customers instead. Product demos beat pitch decks every time. What's the best way you've seen a founder prove market demand before raising?

• 
  Abhishek Singh (@0xlelouch_) reported 27 minutes ago

  Super chad legendary interviwer at dropbox: You need to store 10 billion small files (1-10KB each). Block storage costs are $100K/month. How will you reduce storage costs? [Real problem at Dropbox]

• 
  ᴾᵒᵗ ᵒᶠ ˢⁿᵉᵉᵈ (@P0tofSn33d) reported 27 minutes ago

  @Revolution61858 @Liliyalyv @2WBIA_Reformed ***** y dont u got yoself a dropbox or getchu a link tree wit all da links to download or some shieet so dat when dey take down 1 link u gots all sorts of avenues? Hustler Mindset *****.

• 
  Luna (@_LunarLunaa) reported 30 minutes ago

  @ilovetmrmygffr did the dropbox link work? got taken down a bit ago

• 
  𝕲𝖎𝖌 𝕯𝖎𝖌𝖌𝖊𝖗 (@Gig_Digger) reported 31 minutes ago

  @unusual_whales The problem is several years back when it started was peak Bidenflation era. It reset all prices higher, but thats also Biden should have never been put in office with dropbox stuffing.

• 
  Nyan Nyan (@CatNyanpital) reported 34 minutes ago

  @thescepticalre1 Somewhat. But that's purely speculative and outside the scope. If you can put a price on the quality of data on Reddit vs Dropbox vs Snapchat, you will have solved a multi-billion dollar problem. **** you could start your own company based on this. To me, data is like commodities in that regard. Sort of like gold. You have a hole in the ground (or company) that you want to mine. The quality and amount that is able to be mined out is anyone's guess. I will say that there is an advantage to understanding how AI models mines data.

• 
  The Insight Brief (@SimpleTech247) reported 34 minutes ago

  You pay Google $10/month to store your files. On Google's servers. Where Google can read them. You pay Dropbox $12/month. On Dropbox's servers. Where Dropbox can read them. You pay Apple $10/month. On Apple's servers. Where Apple can read them. Dropbox was breached in 2024. User emails, hashed passwords, API keys, and OAuth tokens were exposed. There is a tool that syncs your files directly between your own devices. No cloud. No server. No middleman. Ever. It's called Syncthing. 81,900+ stars on GitHub. Your files go directly from one device to another. Peer-to-peer. They never touch a third-party server. Not even Syncthing's. Here's what it does: → Syncs files between any number of devices in real-time. → Peer-to-peer. No central server. Your files go directly between YOUR devices. → TLS encryption with perfect forward secrecy on every connection. → Every device authenticated with a strong cryptographic certificate. → Works over LAN and internet. No port forwarding needed. → Selective folder sharing. Sync different folders with different people. → File versioning. Deleted or changed something? Roll it back. → Runs on Windows, Mac, Linux, Android, FreeBSD, Solaris, and more. → Web dashboard to monitor everything from your browser. → No account. No sign-up. Install it. Share a device ID. Done. Here's the wildest part: There is no Syncthing server. There is no Syncthing cloud. There is no company storing your data. The protocol is open and documented. There is nothing between your devices except an encrypted tunnel. Google has shut down 293 products. Dropbox has been breached. iCloud photos have leaked. Every cloud service is one policy change away from scanning everything you store. Syncthing can never shut down your files. Because your files were never on their servers. Dropbox Plus: $12/month. $144/year. Google One 2TB: $10/month. $120/year. iCloud+ 2TB: $10/month. $120/year. Syncthing: $0. Unlimited devices. Unlimited storage. Your hardware. Your files. Forever. 349 contributors. 464 releases. 5,000+ forks. Battle-tested since 2013. Run by the Syncthing Foundation. A Swedish non-profit. MPL-2.0 licensed. Open protocol. Peer-to-peer. Free forever. 100% Open Source.

• 
  avrl ☘ (@avrldotdev) reported 36 minutes ago

  Applied System Design (Real Scale) 9 How Dropbox syncs files across devices? Problem You & a colleague are offline. You both move 10,000 files into different subfolders. When you both go back online at the same time, how does Dropbox prevent a total file-system meltdown?

• 
  Ezaz (@0xEzaz) reported 40 minutes ago

  “Delete Your Dropbox.” Sounds extreme until you realize how much of your life sits on someone else’s server, quietly monitored, limited, and one policy change away from disappearing. This isn’t just a challenge. It’s a wake-up call. The idea is simple: 24 hours. Move your files out of centralized storage and into the BitTorrent ecosystem. No gatekeepers. No single point of failure. Just your data, distributed across a network that doesn’t need permission to exist. We turn it into a movement. A live leaderboard tracking how much data people “liberate” from traditional cloud silos. A real-time counter ticking upward gigabytes, terabytes, petabytes each number representing users taking back control. Not just deleting accounts, but changing how they think about ownership. Because that’s what this is really about. Centralized platforms trade convenience for control. They decide uptime, access, even what’s allowed to exist. The BitTorrent ecosystem flips that model. Your files don’t sit in one place waiting to fail they live everywhere, secured by participation, not policy. So yeah, delete your Dropbox or don’t. But understand the difference. One system rents you space. The other gives you sovereignty. And once you see that, it’s hard to go back. @BitTorrent @justinsuntron #TRONEcoStar

• 
  Jainam Parmar (@aiwithjainam) reported 42 minutes ago

  Fix 3: Turn on two-factor authentication on every important account Even if someone finds a password in your notes, 2FA stops them at the door. They need the password AND a code from your phone to get in. Set it up on: → Email (Gmail, Outlook, Yahoo) → Banking apps → Social media (Instagram, X, Facebook) → Cloud storage (iCloud, Google Drive, Dropbox) → Crypto exchanges Use an authenticator app (Google Authenticator, Authy) not SMS. Text codes can be intercepted through SIM swapping.

• 
  Neural Insights (@neural_insights) reported 43 minutes ago

  Netflix, Google, Anthropic, Meta, xAI, Perplexity, DeepSeek, Instagram, Spotify, Dropbox, Reddit, Pinterest, Uber, Airbnb, Quora all use Python. But sure—Python is “too slow” for your project.

• 
  Grok (@grok) reported 44 minutes ago

  @shravanrayhaan @LeoNelissen Dropbox has ~700 million registered users globally. Latest reported: 18.08 million paying users as of Q4 2025 (flat-to-slightly down YoY, per their FY2025 results). Q1 2026 earnings due May 7. Paying users drive the bulk of their ~$2.5B ARR.

• 
  CPASteve (@BuffCpa) reported 44 minutes ago

  @AccountingAsArt @cordes_tax We use UltraTax on a remote server. We print PDFs, move them to DropBox (our internal storage). Admin puts them through the Tax Return Deljvery system in TaxDome. Easy.

• 
  Anirudh Sharma (@anirudhology) reported 46 minutes ago

  @0xlelouch_ This is a classic small-file storage problem and Dropbox actually solved it by building "Magic Pocket": an in-house object store. The issue with block storage is that it wastes massive space on per-file metadata and partial-block fills. For 10B files, this overhead can cost up to $ 90K/month. Pragmatically, we can pack small files into large (~4MB) immutable blobs which are called "extents" or "superblobs". We write them sequentially, index with a content-hash to blob-offset mapping stored separately. This removes per-file filesystem overhead, reduces metadata pressure, and allows efficient erasure coding across whole blobs. We should also have a tiering policy where we compress blobs and move cold data to low-cost deep archive. Dedup at file level can also be incorporated. This combination can cut costs by 70-90% while keeping latency bounded by the index lookup.

• 
  Glitch Truth (@glitchtruth) reported 46 minutes ago

  the bottleneck is OSC 52 only forwards text, not image bytes. workaround most people land on: screenshot to a synced folder (Dropbox/iCloud) and reference the path, or scp it over before pasting. iTerm2's imgcat works the other direction but not for input. real fix would need a custom escape sequence nobody's shipped yet.

• 
  No (@cameraplan7) reported 48 minutes ago

  @itskinkerbell drive. If people try to download a photo off of Dropbox in a browser/link, the quality actually goes down. I’ve tested on multiple photos and it doesn’t happen with drive

• 
  Frankly Frank (@NoFrankingWay) reported 49 minutes ago

  @ElSotong Little of each. I have: - 28 TB SSD RAID 5 - speed backup - 43 TB UNRAID with HDDs, 1024 Gb is SSD cache - everything BKUP - 12 TB HDD RAID 5 - file BKUP - iCloud - work file BKUP - DropBox Personal Photography BKUP - 250 TB server for Chia - a 2020 project gone defunct. So now I have hella storage

• 
  John Cartwright°͌͌͌͌͌͌͌͌͌͌͌͌͌͌͌͌͌͌͌͌͌͌͌͌͌ 🐈 🐈 🐈 (@bejiitas_wrath) reported 52 minutes ago

  Windows Defender, the built-in antivirus running on every Windows machine, has a working zero-day exploit with full source code sitting on GitHub. No patch, no CVE, and confirmed working on fully updated Windows 10 and 11. A researcher who says Microsoft went back on their word just handed every attacker paying attention a privilege escalation that takes any low-privileged account straight to NT AUTHORITY\SYSTEM. On Windows Server, the result is different but still serious: a standard user ends up with elevated administrator access. The vulnerability is called BlueHammer. On April 2nd, the researcher posted the public disclosure on a personal blog, and on April 3rd, the full exploit source code went live on GitHub. Both were published under the alias Chaotic Eclipse, also known as Nightmare Eclipse, with a message to Microsoft's Security Response Centre that comes down to: I told you this would happen. In late March, the same researcher opened a blog with a single post explaining that they never wanted to come back to public research. Someone had agreed with them and then broken it, knowing exactly what the consequences would be. The post says it left the researcher without a home or anything. A week later, BlueHammer went live on GitHub, with a message specifically thanking MSRC leadership for making it necessary. That is not someone annoyed with a slow review process. That is someone with nothing left to lose. BlueHammer is not a traditional bug, and it does not need shellcode, memory corruption, or a kernel exploit to work. What it does is chain five completely legitimate Windows components together in a sequence that produces something their designers never intended. Those five components are Windows Defender, Volume Shadow Copy Service, the Cloud Files API, opportunistic locks, and Defender's internal RPC interface. One practical limitation worth knowing: the exploit needs a pending Defender signature update to be available at the time of the attack. Without one in the queue, the chain does not trigger. That makes it less reliable than a push-button exploit, but it does not make it safe to ignore. When Defender runs an antivirus definition update, part of that process involves creating a temporary Volume Shadow Copy, which is the same snapshot mechanism Windows uses for backup and restore. That shadow copy contains files that are normally completely locked during regular operation, including the SAM database, which stores the password hashes for every local account on the machine. BlueHammer registers itself as a Cloud Files sync provider, the same kind of thing that OneDrive or Dropbox uses to sync files. When Defender touches a specific file inside that folder, the exploit gets a callback and immediately places an opportunistic lock on that file. Defender stalls, blocked, waiting for a response that is never coming. The shadow copy it just created is still mounted. The window is open. With Defender frozen in place, the exploit reads the SAM, SYSTEM, and SECURITY registry hives directly from the snapshot. It decrypts the stored NTLM password hashes using the boot key pulled from the SYSTEM hive, changes a local administrator account's password, logs in with that account, copies the administrator security token, pushes it to the SYSTEM level, creates a temporary Windows service, and spawns a command prompt running as NT AUTHORITY\SYSTEM. Then, to cover its tracks, it puts the original password hash back. The local account password looks completely unchanged. No crash, no alert, nothing. The Cloud Files provider name hardcoded in the exploit source code reads IHATEMICROSOFT. The administrator password used during the escalation is hardcoded as $PWNed666!!!WDFAIL. These are not bugs left by accident. They are messages, written directly into the code, and there is only one intended reader.

• 
  Grok (@grok) reported 52 minutes ago

  @richardroop @ihtesham2005 Syncthing pros: Completely free/open-source, zero subscriptions. Private P2P sync—files never touch a company server. Strong encryption, cross-platform (PC, phone, etc.), and reliable once running. Cons: Setup is more technical than Dropbox/iCloud. Devices need to be online (or use relays) for sync to happen. No easy web access or share links. UI feels basic. Not using it: You either pay for clouds (with their rules/privacy tradeoffs) or manually copy files—losing automatic, seamless access across devices. Great if privacy/control matters more than convenience.

• 
  neo_relic (@neo_relic123) reported 54 minutes ago

  @Ashleydoncare @rachallison1 ... IF A MAN LEAVES HIS CHILD AT A FIRE STATION OR HOSPITAL OR DROPBOX OR WHATEVER THEY WILL HUNT HIM DOWN!!! NOT THE SAME FOR WOMEN!!!! WAKE ******** UP!!!! 🤬🤬🤬🤬🤬🤬🤬

• 
  John Lamont is making waifus (@jtl2dotnet) reported 56 minutes ago

  @ProtonPrivacy I sent a customer a file via @ProtonDrive and it got flagged as a virus. They were running Google Chrome and AVG on a Motorola device. Have you guys gotten any reports of issues with these? Same file on Dropbox went through without an issue.

• 
  StackScan (@stackscans) reported 57 minutes ago

  Growing a SaaS is like debugging a system. You don’t fix everything at once. You isolate one issue, solve it, then move forward. Fix onboarding → conversions improve Fix retention → revenue stabilizes Fix distribution → growth accelerates Example: Dropbox focused first on one problem: seamless file syncing. They didn’t try to build a full ecosystem on day one. They nailed one core use case, then expanded. One problem at a time. That’s how real scale happens.

• 
  Rober (@robsoto1511) reported 58 minutes ago

  @MEGAprivacy would be nice if joplin could sync with mega or proton their options are onedrive dropbox and the joplin server

• 
  Bocamax (@Bocamax1) reported 60 minutes ago

  I want Grok to win. But it's useless about 90% of the time I try to use it as a free user. Dropbox got rich by offering limited free access. Gemini works 100% of the time. OpenAI works most of the time. Anthropic works most of the time. Cleary xAI is in fiscal trouble.

• 
  Shirochenko Dmitriy (@dmshirochenko) reported 60 minutes ago

  Computer vision deployments for enterprises used to take 6-12 months, juggling five tools (labeling, data management, training, cloud hosting, edge deployment), thousands of lines of custom Python glue code, and dedicated ML/DevOps teams. General tools like Dropbox caused "million paper cuts": poor tracking of images/annotations, high preprocessing/training costs. CV stayed trapped in research labs, out of reach for small teams. Inaction costs: manufacturers lose millions to undetected defects (one ag equipment maker saved $8M/facility post-adoption); 3,000+ hours annual unplanned downtime from preventable failures; 50%+ customer returns from quality issues; 6-12 month projects while competitors launch in days. Visual AI unlocks programming the physical world and value from passive video feeds. #AI