183 AI Bots on Telegram with 44M Users: What the Data Shows

183 AI Bots on Telegram with 44M Users: What the Data Shows

I wrote a deep-dive research piece on AI agents in Telegram and what it means for TON. Here's the summary.

Telegram is the default interface for AI products

I found 183 AI bots with a combined 44.3 million monthly active users. These aren't dead demos — they're live products with real users. Telegram has become the default AI interface: no app download, no login screen, you just message a bot and start working.

For context, 44.3M MAU is larger than most standalone AI applications. And this only counts bots I could find and measure. The real number is likely higher.

Why Telegram specifically? Three reasons. First, 900M+ global users — massive distribution out of the box. Second, the Bot API is genuinely good for AI use cases — streaming responses, inline keyboards, file handling. Third, the cost of acquiring a user is near zero when they're already in the messenger every day.

BotFather doubled to 7.3M MAU

BotFather — the bot you use to create bots — doubled its MAU to 7.3M in just two months. That means more people are building bots than ever, and a significant chunk of them are building AI bots. The platform is growing on the supply side, not just demand.

The Cocoon ecosystem is already working and growing. The client is no longer just Telegram itself — the infrastructure extends beyond the messenger.

TON connects the agent economy

TON is plugging into x402, EIP-8004, A2A protocols, discovery, and reputation layers. The infrastructure for agent economy is assembling into a coherent stack. Agents need to pay each other, verify each other, and find each other. TON + Telegram is a natural combination for all three.

If blockchain in a messenger seemed like a strange idea before, AI agents make it click: an agent needs a wallet, an agent needs an interface, an agent needs payment infrastructure. Telegram + TON covers all three.

Full report: Telegram as the Default AI Interface


Follow me on Twitter/X @danokhlopkov for more on AI agents and automation.