24 billion credential records exposed: what to do about it
Key takeaways
- Cybernews researchers found an exposed Elasticsearch cluster on 12 June holding 24 billion records and more than 8.3 terabytes of data.
- They believe most records were infostealer logs: usernames, email addresses and plaintext passwords saved with the URL each one unlocks.
- The researchers cannot say how many records are duplicates, how old most of the data is, or who owns the database.
- Credential stuffing is the main risk: attackers try leaked email and password pairs on other accounts.
On 12 June, Cybernews researchers found an exposed Elasticsearch cluster holding 24 billion records and more than 8.3 terabytes of data. Security Affairs, reporting on their findings, says the researchers triple-checked the numbers. The headline figure is enormous, but the useful part of this story is what kind of data it was, because that decides what a person or a company should do about it.
What was in it
The researchers believe the vast majority of records were infostealer logs: usernames, email addresses and plaintext passwords, each saved alongside the URL it was meant to unlock. The data came from 36 distinct sources. More than 30 of them were Telegram channels, and over 1.7 billion records traced back to those channels.
The biggest share, 22.6 billion records, came from a source the database owner labelled "collections". The researchers could not look further into what that means, because the database was taken out of public view soon after it was found.
Mixed in with the credentials were around 17,000 records containing CVE vulnerability IDs with GitHub links, and thousands of logs of news articles and social media posts about data breaches. Cybernews reads this as the owner actively monitoring the security landscape, likely to keep the collection updated with new breaches.
What nobody knows
The researchers are open about the gaps. They cannot say how many records are duplicates, how old most of the data is, or who owns the database, and they cannot confirm how many people are affected. So 24 billion is a count of records, not of people or of working passwords.
That matters for how you react. A number this size is not a reason to assume every account is compromised. It is a reminder of two specific weaknesses that this kind of data exploits.
Weakness one: password reuse
Have I Been Pwned's Pwned Passwords page describes the general risk of reused credentials. When credentials are exposed, attackers can use known email and password combinations to get into other accounts, automating login attempts with leaked credentials in what is called credential stuffing. Even when people change their passwords slightly between sites, attackers can predict common patterns.
Weakness two: infected devices
Infostealer logs are different from a website breach. They are stolen usernames, passwords and the services those credentials were meant to unlock.
What to do
- Check addresses and passwords against known breaches. Our guide to Have I Been Pwned covers the free checks.
- Kill password reuse. Every account gets its own password, held in a password manager.
- Turn on multi-factor sign-in on your accounts.
Frequently asked questions
Was my password in the 24 billion leak?
There is no way to search this database: it was taken out of public view soon after it was found. Checking your passwords against known breaches is the practical step.
Does a big number mean billions of people were affected?
No one knows. The researchers say they cannot confirm how many people are affected or how many records are duplicates.