AlphaTON Capital Corp, a digital asset treasury and AI infrastructure company perhaps best known for development of the Telegram ecosystem, has completed a US$ 30 million lease agreement with NVIDIA, to deploy 504 B200 GPUs in a Canadian data center, expanding its AI compute capacity.
According to a press release, the GPUs are being delivered this week, and the deployment is expected to generate a minimum of US$ 1.2 million in revenue per month starting in March 2026. Revenue is derived from hourly GPU rentals, with B200 chips generating US$ 4 to US$ 6 per hour for general AI workloads and 4.41 TON per hour for Cocoon AI workloads.
This deployment aims to enhance AlphaTON’s ability to provide AI infrastructure to developers, enterprises, and the Telegram ecosystem, while creating predictable revenue streams for the hosting operator.
Brittany Kaiser, CEO AlphaTON Capital, North America, and Canada, said, “Many AI workloads cannot be run on Big Tech infrastructure due to privacy, sovereignty, and data protection constraints.”
For this deployment, the estimated monthly revenue is approximately US$ 1.45 million, and when combined with AlphaTON’s existing and scheduled GPU deployments, totaling 1,080 GPUs by April 2026, monthly revenue for the hosted infrastructure could reach US$ 3.11 million.
This expansion directly increases the utilization and revenue potential of the Canadian data center. The facility benefits from energy-efficient design, abundant hydroelectric power, competitive energy costs, and natural cooling, all of which reduce operational costs and support environmental sustainability. By leveraging existing infrastructure, AlphaTON can scale AI capacity rapidly without heavy upfront capital expenditure.
The broader market for AI infrastructure is growing rapidly. Big Tech invested over US$ 400 billion in AI infrastructure in 2025, with projected cumulative spending of US$ 600 billion in 2026. Regulatory and privacy considerations are driving demand for alternatives to large cloud providers.
The Canadian GPU deployment represents a strategic approach to capital-efficient growth through leasing structures. For the data center operator, it translates into higher utilization, predictable revenue, and the opportunity to participate in the expanding market for confidential, privacy-focused AI infrastructure. By hosting this deployment, the facility gains a long-term role in a sector that is rapidly becoming a critical component of the digital economy.