Cocoon TON AI Launch brings decentralized encrypted AI compute to TON, enabling GPU owners to earn rewards and expanding ecosystem.
Author: Tanishq Bodh
December 1, 2025 – Cocoon, the Confidential Compute Open Network backed by Telegram co-founder Pavel Durov, has officially gone live on The Open Network (TON). The launch marks a major milestone in decentralized AI infrastructure. Announced on November 30, the Cocoon TON AI Launch allows GPU owners worldwide to contribute compute power for encrypted AI inference tasks and earn Toncoin (TON). The rollout positions TON as a potential challenger to hyperscalers like AWS and Azure, offering censorship-resistant and privacy-first compute at scale.
Durov, posting on Telegram, described the Cocoon TON AI Launch as a “fusion of AI and crypto” that restores control to users. Early adopters — from individual PC gamers to small data centers are already running workloads and receiving TON rewards. Cocoon routes tasks through a network of trusted execution environments (TEEs), ensuring private prompts remain encrypted throughout processing.
The announcement generated substantial buzz across Telegram channels and TON community forums, with users praising the Cocoon TON AI Launch for opening a new economic model for GPU owners. Tech funds and mining operators have begun allocating hardware to the network, accelerating throughput ahead of broader deployment.
TON’s ecosystem also responded positively: developers noted that Cocoon expands TON’s utility beyond payments, gaming, and DeFi, creating a compute marketplace directly accessible through Telegram mini-apps. Early price action for TON remained steady, though analysts expect increasing developer demand to strengthen long-term fundamentals as the launch evolves.
Despite the enthusiasm, several challenges still shadow the Cocoon TON AI Launch. To begin with, scaling to the level of cloud hyperscalers requires onboarding tens of thousands of GPUs, a logistical hurdle decentralized networks have historically struggled with. Moreover, regulatory scrutiny around AI-blockchain hybrids is increasing, especially when encrypted inference and cross-border compute services are involved.
Privacy researchers, while generally supportive of Cocoon’s architecture, note that TEEs such as Intel’s TDX rely on manufacturer-level trust assumptions. As a result, critics argue that decentralization must be paired with rigorous auditing. Otherwise, the ecosystem risks overreliance on hardware security modules that may not be fully transparent.
The launch distinguishes itself by combining TON’s sharded architecture capable of millions of transactions per second with privacy-preserving compute. Developers can submit workloads for models like DeepSeek and Qwen, while the network matches tasks to available GPU providers.
Telegram is set to become Cocoon’s first large-scale customer, integrating private AI features into bots and mini-apps. This native demand could accelerate adoption by giving developers immediate access to end users. For GPU owners, onboarding is simple: register at cocoon.org, connect hardware, and begin earning TON for idle compute cycles.
As decentralized AI networks gain traction, the launch positions TON at the forefront of a compute economy that prizes privacy, transparency, and user sovereignty.
Our Crypto Talk is committed to unbiased, transparent, and true reporting to the best of our knowledge. This news article aims to provide accurate information in a timely manner. However, we advise the readers to verify facts independently and consult a professional before making any decisions based on the content since our sources could be wrong too. Check our Terms and conditions for more info.
