AlphaTON Capital and SingularityNET partner to advance Telegram’s Cocoon AI network

AlphaTON Capital Corp. (Nasdaq: ATON) has announced a strategic partnership with SingularityNET, CUDO Compute, and Vertical Data to accelerate the development and deployment of Telegram’s Cocoon AI network.

The collaboration introduces a new fleet of high-performance GPUs powered by hydroelectric energy in Sweden, marking a significant milestone in combining decentralized artificial intelligence, privacy protection, and environmental sustainability.

Sustainable infrastructure for ethical AI

The partnership aims to address one of AI’s most pressing challenges, the need for decentralized, privacy-preserving computing power that does not compromise environmental responsibility.

Cocoon AI, Telegram’s privacy-focused AI initiative, enables users to maintain control over their personal data while using advanced AI features.

This vision aligns closely with AlphaTON’s mission to bridge traditional finance with blockchain-based innovation.

“This partnership represents the future of AI infrastructure — one where privacy, sustainability, and decentralization aren’t competing priorities, but foundational principles,” said Brittany Kaiser, CEO of AlphaTON Capital Corp.

“At AlphaTON, we’ve always believed that technology should empower individuals, not exploit them.”

Kaiser emphasized that the initiative merges SingularityNET’s leadership in decentralized AI, CUDO Compute’s and Vertical Data’s computing expertise, and AlphaTON’s commitment to sustainable infrastructure.

The goal is to create an ecosystem that respects both human rights and planetary boundaries.

Hydroelectric-powered GPU fleet and decentralized architecture

Under the partnership, a fleet of high-performance GPUs—structured and financed through Vertical Data’s GPUFinancing.com subsidiary—will be deployed in hydroelectric-powered data centers.

This renewable energy approach is designed to minimize the carbon footprint of AI training and operations while enabling privacy-preserving machine learning across the Cocoon AI network.

“Our hydroelectric-powered GPU infrastructure demonstrates that we can scale AI responsibly,” Kaiser added, highlighting how blockchain transparency and renewable energy verification contribute to accountable and sustainable AI growth.

SingularityNET COO Janet Adams echoed these sentiments, stating that decentralized AI requires decentralized infrastructure.

“This partnership with AlphaTON Capital demonstrates how we can scale AI capabilities while maintaining the ethical foundations that make AI trustworthy. Privacy and sustainability aren’t obstacles — they’re competitive advantages,” Adams said.

Vertical Data CEO Deven Soni added that the convergence of decentralized AI and distributed GPU financing marks “a new chapter in computing.”

According to Soni, innovative financing solutions will make cutting-edge compute power accessible to ethical and sustainable AI projects like Cocoon AI, allowing them to grow without reliance on centralized capital or infrastructure.

Positioning for growth within Telegram’s ecosystem

The initiative reinforces AlphaTON’s strategic role within the TON (The Open Network) and Telegram ecosystems, where the company has become one of the largest corporate holders of TON tokens.

With over 1 billion monthly active users, Telegram represents a vast potential market for privacy-preserving AI applications, including messaging assistants, decentralized content moderation, and translation tools.

The collaboration also reflects a growing consensus in the technology sector that future AI infrastructure must prioritize user sovereignty, environmental responsibility, and decentralized governance.

By combining AlphaTON’s financial and sustainability expertise with SingularityNET’s decentralized AI protocols and the technical strengths of CUDO Compute and Vertical Data, the partnership seeks to establish new standards for ethical AI deployment.

The initial GPU deployment is expected to begin in Q4 2025, with expansion planned through 2026 and beyond, based on demand from Cocoon AI network participants.