WhatsApp manages a daily volume of 100 billion messages through a system built on persistent TCP connections, kept open continuously between a user's phone and the company's chat servers.
Rather than opening a new connection every time a message is sent, each device stays linked to its assigned server, allowing messages to move instantly the moment they arrive.
The technology underneath
Every individual chat server runs on the Erlang programming language, paired with the FreeBSD operating system.
Erlang was originally built for telecoms equipment that had to stay running for years without failure, which is part of why WhatsApp adopted it for a system that needs the same kind of reliability at enormous scale.
Each chat server is capable of supporting approximately 2 million concurrent connections, an unusually high figure that keeps the overall number of servers WhatsApp needs relatively small.
How a message actually finds its recipient
The system relies on a routing table that maps every user to a specific server.
When a message is sent, it can hop between servers as needed until it reaches whichever server the recipient's device happens to be connected to at that moment.
No messages left behind
WhatsApp does not store messages once they have been successfully delivered to the recipient.
The platform also does not use cloud backups for messages by default, meaning any chat a user deletes is permanently removed and cannot be recovered afterwards.
That design choice reflects a broader privacy stance: rather than retaining a searchable archive on its own servers, WhatsApp is built so that messages exist only on the sender's and recipient's devices once delivery is complete, a structural decision that shapes everything from how the company handles legal requests to how little data would be exposed in the event of a server breach.
