Architecture·December 17, 2025·12 min read
Why YouTube Live Chat Scales Without WebSockets
How YouTube serves millions of concurrent viewers with HTTP-based polling instead of WebSockets, saving tens of millions in infrastructure costs while maintaining real-time performance.
Introduction
The Core Problem: Live Chat at Internet Scale
Common Assumption: WebSockets = Real-Time Chat
Why WebSockets Do Not Scale Well for YouTube
1. Persistent Connections Are Expensive
2. WebSockets Break CDN Economics
3. Load Balancing Becomes Hard
What YouTube Uses Instead
Why This Works
Cost Comparison (Back-of-the-Envelope)
WebSockets Scenario
HTTP Polling Scenario
Estimated Savings
Why Twitch Made a Different Choice
A Broader Industry Pattern
Lessons for System Designers
Conclusion
Filed under fieldnotesDecember 17, 2025