Stop overpaying - start transferring money with Ogvio. Join the waitlist & grab early Rewards NOW! 🎁
TON’s Cocoon Rewards GPU Power for Secure AI Computing
Key Takeaways
- Cocoon, a decentralized AI platform on TON, rewards GPU providers with TON tokens for powering secure, private AI computations;
- The system encrypts all data, ensures user privacy even from GPU providers, and offers an alternative to centralized AI services;
- Telegram will integrate Cocoon to enhance AI features in Mini Apps and bots.
The launch of Cocoon, a decentralized artificial intelligence (AI) platform developed on The Open Network (TON), was announced on November 30.
Cocoon is engineered to prioritize user privacy by encrypting all data during computation. This ensures that even those providing the GPU hardware for computation cannot see any user data.
The system is built so that those with extra GPU resources can supply their computing power to the network. In return, they receive TON
Did you know?
Subscribe - We publish new crypto explainer videos every week!
What is a Crypto Bull Run? (Animated Explainer + Prediction)
Developers seeking AI computing capacity can access an alternative to large, centralized infrastructure providers.
Pavel Durov, co‑founder of Telegram, has stated that the current model of dominating compute companies such as Amazon and Microsoft involves high costs and less secure data practices.
Cocoon offers a route around these issues, which allows users to access computation without third-party control or privacy drawbacks.
The official introduction of Cocoon took place at the Blockchain Life 2025 event in Dubai in October. During this event, the vision for confidential, decentralized AI infrastructure was outlined, and the need for privacy-focused solutions in the AI industry was discussed.
After its public debut, Cocoon began handling real AI tasks and dispersing TON tokens to GPU providers.
Telegram will be the first organization to integrate Cocoon into its offering. It will use the platform's AI capabilities in Mini Apps and bots, and expand AI-powered features for its audience.
Jay Neo, who previously worked on short video content at MrBeast, recently developed an AI tool called Palo. How does it work? Read the full story.