125 Million Text Messages Sat Exposed for Four Years After a Telecom Misconfiguration
Baltic telecom provider CSCTelecom left a database containing more than 125 million SMS messages exposed, including messages from banks, airlines, and healthcare providers.

125 million SMS messages, sitting in the open
A major telecom provider in the Baltic states left a database containing more than 125 million SMS messages exposed, and the messages sat there for years. CSCTelecom, which operates in Lithuania, Latvia, and Estonia, had a misconfiguration at its Latvian branch that left the database accessible to anyone who found it. The records span from October 2022 to September 2026, four years of text messages from banks, airlines, healthcare providers, and other European organizations.
Nobody had to hack anything. The database was simply left open, and the exposure was a configuration mistake rather than an intrusion. That distinction matters, because it changes the lesson. This was not a sophisticated attack defeating a strong defense. It was a basic failure, the kind of mistake that a checklist and a routine scan would have caught, left unnoticed for years.
What was in the messages
The exposed messages came from organizations that use SMS for things that matter. Banks send one-time passcodes and fraud alerts by text. Airlines send booking confirmations. Healthcare providers send appointment reminders and test results. A database of 125 million such messages is a detailed map of who was doing what, when, and with whom, across three countries and dozens of organizations.
The exposure is especially serious because SMS is still the fallback for two-factor authentication. A one-time code sent by text is only secure if the text itself is private. When four years of those codes sit in an exposed database, the codes themselves are long expired, but the phone numbers, the patterns of which services a person uses, and the timing of their activity are all still useful to an attacker. A criminal who knows which bank you use, which airline you fly, and which clinic you visit has a much easier time writing a phishing message you will believe.
The third-party breach pattern keeps repeating
CSCTelecom is the latest in a run of breaches that all share the same shape: the victim is not the company that got breached, it is the company's customers. The Fireworks AI breach exposed data belonging to Uber, Notion, and Cursor through a shared cloud provider. The JPMorgan, Visa, and Mastercard vendor breach exposed customer contact data through third-party service providers. In each case, the company whose name is on the story did nothing wrong itself. Its vendor did.
That pattern is the real story of 2026's breach landscape. Companies have spent a decade hardening their own networks, and attackers have responded by going after the suppliers, the cloud providers, the marketing vendors, and the telecom operators that sit between a company and its customers. The weakest link is no longer the target's firewall. It is the vendor the target never audited.
What should have caught this
A database containing 125 million messages does not get exposed overnight. It gets exposed when someone changes a configuration, nobody reviews the change, and no automated scan flags the result. The tools to catch this exist and are cheap. Cloud security scanners check for publicly accessible databases as a matter of routine. An organization that stores other people's messages should be running those scans continuously, and should be able to answer, at any moment, which of its databases are reachable from the internet.
CSCTelecom has not said how the database was discovered or how long it was accessible before it was secured. What is clear is that the exposure lasted long enough to cover four years of messages, and that the organizations whose messages were in the database had no way to know. A bank cannot protect a text message after it hands the message to a telecom provider. That is the point of the exercise: the security of your data depends on the security of companies you have never heard of.
What to do about it
For individuals, the practical lesson is to stop treating SMS as a secure channel. If a service offers an authenticator app or a passkey instead of SMS codes, use it. Our guide on checking whether your phone is hacked covers the warning signs worth knowing, and the data breach email check guide shows how to find out whether your address has shown up in a breach.
For organizations, the lesson is to audit the vendors that handle your customers' data, and to ask the specific question: which of your databases are reachable from the internet, and who checked last? The CSCTelecom breach was not a failure of technology. It was a failure of process, and processes are fixable. The 125 million messages are already out there. The next database does not have to be.


