The Most Dangerous Data Type Is Also the Most Common
Across the LeakedSource/HEROIC database, plaintext passwords appear in 34,627 breaches. That is more than any other exposed data type, and it comes from a database that indexes 31,208,728,017 records across 38,972 distinct breaches. Email addresses follow closely at 32,511 breaches, with URLs at exactly the same count.
Plaintext passwords are the most dangerous data type because they need no cracking. An attacker doesn't have to break a hash or guess anything. They can paste the credential into a login form, or more likely, feed it into an automated tool that tries it on hundreds of sites.
How Breaches Are Built Now
The breach type distribution explains why plaintext is so prevalent:
- Stealer logs: 17,604 breaches
- Combolists: 15,000 breaches
- Databases: 4,645 breaches
- Database/combolist hybrids: 1,717 breaches
Stealer logs and combolists account for more than 32,000 of the roughly 39,000 tracked breaches. Neither comes from a single company's database being hacked. Stealer logs are harvested from infected devices, and combolists are compiled from many sources. Both are built for one purpose: delivering usable credentials, which means plaintext passwords.
The identical counts for email addresses and URLs (32,511 each) fit this pattern. Stealer-log records typically pair a login with the site it belongs to, so an attacker knows your email, your password, and where to use them. The data doesn't prove that structure, but the match is hard to ignore.
The Largest Breaches Are Credential Collections
Look at the ten largest breaches by record count and a theme emerges. Five of them explicitly list email addresses paired with plaintext passwords:
- XSS.IS Combolist: 2,472,611,041 records (also includes usernames)
- Misc Combolists: 1,927,529,995 records
- Ga$$Pacc Collection: 518,129,787 records
- AntiPublic: 348,250,277 records
- Pemiblanc: 344,093,910 records
Together, these five hold about 5.6 billion records. Exploit.in (503,910,874 records) and MySpace (301,914,981 records) also expose passwords, though their entries list the data more generically, and MySpace includes a hash type.
Two details stand out. First, these collections span 2008 to 2020, so credentials leaked years ago are still circulating and being recombined. Second, combolists like AntiPublic and Pemiblanc exist to be reused in credential-stuffing attacks, where leaked logins are tried against other services. If you reused a password anywhere in that window, an old leak can still compromise a current account.
Passwords Aren't the Only Problem
Not every large breach involves passwords, and the others show why identity data matters too.
Verifications.io exposed 722,745,807 records containing email addresses, phone numbers, first names, and last names. It has no passwords, but it hands attackers a ready-made list for targeted phishing and impersonation, because they can address you by name and reach you on more than one channel.
Weibo exposed 503,294,488 records, and the data type listed is phone numbers. Phone numbers are increasingly used as identifiers and recovery factors. Exposed numbers enable SMS phishing, SIM-swap attempts, and account-recovery abuse.
The broader data-type counts reflect this layering: password hashes (2,629 breaches), first names (1,567), last names (1,546), usernames (1,357), phone numbers (1,136), and IP addresses (823). Each piece is modest alone, but combined across breaches they build a detailed profile of one person.
The Exposure Keeps Growing
This isn't just a historical problem. The most recently dated entries in the database include the Russian Federal Tax Service (FNS) at 2,028,243 records, the Alibaba Suppliers Directory at 158,583 records, the Israel Population & Immigration Authority at 100,000 records, dream.co.nz at 16,562 records, and REMOTECOM at 2,032 records. They range from government agencies to commercial directories to small businesses, and no sector is immune.
What You Should Do
1. Stop reusing passwords. With billions of plaintext credentials in circulation, any reused password should be treated as compromised. Use a password manager to generate a unique, long password for every account, starting with email, banking, and cloud storage.
2. Add phishing-resistant multi-factor authentication. A stolen password alone shouldn't unlock your account. Prefer authenticator apps or hardware security keys over SMS codes, since phone numbers are widely exposed in breaches like Weibo.
3. Treat infostealer infections as a credential event. If your device was infected, changing one password isn't enough. Stealer logs capture everything saved in the browser. Run a full malware scan, then rotate every credential that device touched and sign out of active sessions.
4. Be skeptical of personalized messages. When names, emails, and phone numbers are exposed, as in Verifications.io, phishing becomes more convincing. Verify unexpected requests through a channel you trust.
Check Your Exposure
With 38,972 breaches tracked and more than 31 billion records indexed, the question isn't whether your data is out there, but what's exposed and where. Search your email address or username at LeakedSource to see which breaches include you and which data types were leaked, so you know which passwords to change first.