AlphaTON Capital has moved to deepen its exposure to decentralized artificial intelligence infrastructure by signing a 46 million dollar compute deal centered on NVIDIA hardware, marking a significant expansion of its digital asset treasury strategy. The agreement adds hundreds of advanced GPU chips to AlphaTON’s balance sheet as it looks to scale Cocoon AI, a decentralized compute network designed to operate natively within the Telegram ecosystem. The chips are scheduled for delivery in February and represent the firm’s first large scale deployment focused on confidential compute, a model that aims to process AI workloads while preserving user privacy. AlphaTON said the move builds on its existing asset base and reflects a deliberate effort to tie its treasury strategy to revenue generating infrastructure rather than passive token exposure. The announcement comes as demand for AI compute continues to outpace supply globally, pushing firms to secure long term access to hardware.
The Cocoon AI network allows users to rent out their own GPUs to process AI queries and receive compensation in Toncoin, positioning the platform as an alternative to centralized AI services operated by large technology firms. AlphaTON views this model as a way to combine decentralized finance principles with AI infrastructure, leveraging the scale of Telegram’s user base to drive adoption. The network is built on the TON blockchain and is designed to integrate directly into Telegram’s ecosystem, which reaches roughly one billion monthly active users. AlphaTON has previously said deploying compute capacity into Cocoon would create a new revenue stream while reinforcing the utility of the TON ecosystem. By backing decentralized compute rather than proprietary models, the firm is betting that privacy focused AI services can gain traction as scrutiny grows around data use and centralized control.
The structure of the deal reflects how capital intensive AI infrastructure has become. AlphaTON contributed a portion of the funding in cash while financing the majority through non recourse debt and staged equity payments tied to deployment milestones. This approach allows the firm to scale compute capacity without fully depleting its balance sheet while maintaining flexibility as market conditions evolve. The company has also taken steps to ensure it has access to additional capital, including filing a large shelf registration late last year to support future investments across AI and the broader TON ecosystem. Those initiatives include potential funding for mini app development and payments products tied to blockchain based services, signaling a multi pronged expansion strategy rather than a single bet on compute alone.
AlphaTON’s latest move underscores how crypto linked treasuries are increasingly intersecting with the AI infrastructure buildout. By acquiring physical compute assets and aligning with decentralized networks, firms are seeking exposure to the AI growth cycle without relying solely on traditional equity investments. The company said its approach is designed to reduce volatility risk by anchoring value to tangible assets and long term partnerships while maintaining upside tied to emerging AI demand. As competition for GPUs intensifies and decentralized alternatives gain attention, AlphaTON’s expansion into AI compute highlights a broader trend where blockchain firms are positioning themselves as infrastructure providers at the intersection of digital assets and next generation computing.
