Dropbox Watchdog

Search issues

Search the Dropbox Watchdog archive

All issues

2024: Dropbox's 2012 credentials resurface in the 'Mother of All Breaches'

January 2024

MediumStatus: Historical ~68 million (2012 credentials, re-aggregated) affectedProduct: AccountsYear: 2024

In January 2024 a 26-billion-record compilation dubbed the 'Mother of All Breaches' surfaced online — and the 68 million credentials stolen from Dropbox in 2012 were among the datasets bundled into it.

What happened

In January 2024, security researchers reported the discovery of an enormous aggregated dataset — roughly 12 terabytes containing about 26 billion records compiled from thousands of past breaches — which they nicknamed the 'Mother of All Breaches' (MOAB). Most of it was repackaged from known incidents, but consolidating it into a single searchable trove meaningfully lowered the effort for attackers to mount credential-stuffing and targeted phishing at scale.

Among the datasets folded into MOAB were the roughly 68 million Dropbox user credentials stolen in the 2012 breach, alongside records from LinkedIn, Adobe, Twitter, and many others. For Dropbox specifically, MOAB is a vivid demonstration of breach 'long tail': credentials taken in 2012, not fully disclosed until 2016, and force-reset that year, were still circulating and being re-weaponized twelve years later in 2024.

The entry is included because it shows that a breach's consequences do not end with disclosure or a password reset. Anyone who reused their pre-2012 Dropbox password elsewhere — and never changed it — remained exposed as those credentials kept resurfacing, now bundled with billions of others for easy abuse.

Impact

MOAB underscores why the 2012 Dropbox breach still matters: stolen credentials are effectively permanent, and aggregation makes them more dangerous over time, not less. It reinforces the case for unique passwords and 2FA, and it is a concrete reminder that a provider's 'we reset the passwords' response does not undo the lasting exposure of the underlying theft.

Dropbox's Response / Official Position

Dropbox had already force-reset affected passwords in 2016 and promotes two-step verification; MOAB was an industry-wide aggregation rather than a new Dropbox incident, and Dropbox did not issue a fresh incident response specific to it. The continued circulation reflects the original 2012 theft.

Sources

Related guides

Spot an error, or have a source to add?
Report an error / suggest update

Related issues

9 sources
HighApproximately 5,000 accounts; files accessed in fewer than a third (about 1,500 by 9to5Mac's arithmetic)

The 2026 Lenovo ID sign-in flaw: ~5,000 Dropbox accounts entered without a Dropbox password

A flaw in how Lenovo verified account-holder email addresses let an attacker register a Lenovo ID on a victim's email, and Dropbox's Lenovo ID sign-in link then trusted that identity without ever asking for a Dropbox password — reaching roughly 5,000 accounts.

Security Incidents & Data BreachesCurrent / Ongoing Issues (2024–2026)
Read documentation

Across multiple years, attackers have built convincing fake Dropbox login pages — reached via PDF lures and redirect chains through trusted cloud storage — to harvest victims' real business email and Dropbox credentials.

Security Incidents & Data BreachesAccount Lockouts & Support Failures
Read documentation

ConsentFix, an OAuth-consent phishing technique first documented by Push Security in December 2025 and reported on independently through mid-2026, delivers its Microsoft 365 lures through trusted file-hosting platforms — reporting names both Dropbox and DocSend (a Dropbox company) as hosts for the password-protected files attackers use to get past mail filters.

Security Incidents & Data BreachesCurrent / Ongoing Issues (2024–2026)
Read documentation

On 10 September 2026 Dropbox announced, together with OpenAI, a native Dropbox experience inside ChatGPT. Users who connect their account can browse, search, @mention and preview Dropbox files there. Four Dropbox pages captured for this entry describe permissions and a per-user disconnect path, but none of them states whether content sent to ChatGPT is used to train models, how long OpenAI retains it, or any admin switch to block the connector.

Privacy & Encryption ConcernsDeveloper, API & Platform
Read documentation