Telegram Unveils Blockchain-Based AI Network
Telegram CEO Pavel Durov announced the launch of Cocoon, a decentralized AI inference network built on The Open Network (TON), during the Blockchain Life Conference in Dubai on October 29. The new platform aims to process AI queries privately using blockchain infrastructure, allowing developers to pay for GPU access in Toncoin (TON) tokens.
Telegram will serve as Cocoon’s first large-scale user, routing parts of its AI workloads through the network and introducing the service to its global base of over one billion users.
How Cocoon Connects AI and Blockchain
Cocoon functions as a decentralized marketplace connecting developers who need computational power with GPU owners willing to contribute resources. GPU providers earn Toncoin for completing inference tasks, creating a transparent and incentive-driven ecosystem for AI computation.
Developers can now register through Cocoon’s official channel by submitting details such as their model type (e.g., DeepSeek, Qwen), daily query volume, and input-output token size. GPU contributors must specify hardware specs, including GPU count, model, VRAM capacity, and uptime expectations.
Durov said the initiative marks a new step toward confidential, distributed AI systems, combining Telegram’s infrastructure with TON’s blockchain scalability.
AlphaTON Capital Provides Infrastructure Support
Shortly after the announcement, AlphaTON Capital confirmed its investment in the project, pledging to deploy high-performance GPU clusters across global data centers to power Cocoon’s network.
AlphaTON CEO Brittany Kaiser described the launch as a turning point for privacy-focused AI systems, emphasizing that Cocoon’s design aligns with growing concerns around secure data processing. Executive Chairman Enzo Villani added that privacy and transparency will define the next phase of AI adoption, calling Cocoon a “critical infrastructure layer” for decentralized computation.
With Telegram’s reach and TON’s blockchain framework, Cocoon represents one of the first large-scale integrations of AI inference and decentralized finance infrastructure.
