Bell Canada to Convert Former Winnipeg Food Processing Plant into AI Data Center
April 29, 2026
Bell Canada to Convert Former Winnipeg Food Processing Plant into AI Data Center
Bell Canada has begun converting a former protein-processing plant in Winnipeg, Manitoba, into an AI-focused data center, part of a broader push to expand its national AI infrastructure under the company’s “AI Fabric” network. The project underscores the growing trend of repurposing industrial facilities to meet the surging demand for compute capacity driven by artificial intelligence workloads.
Located in the CentrePort industrial development, the facility was originally built in 2021 as a $109.55 million plant for Merit Functional Foods, which processed peas and canola into protein. However, the company entered receivership just two years later, and in 2025, former co-CEOs Ryan Bracken and Barry Tomiski repurchased the site without disclosing future plans. The 94,000-square-foot (8,732 sqm) building is now being retrofitted for data center use, with modifications including wall adjustments and the construction of concrete pads to support generators, chillers, and transformers.
Bell Canada is investing approximately $23 million in the conversion, with the facility expected to deliver 5.5 MW of capacity upon completion. The site will be integrated into Bell’s AI Fabric, a network of data centers the company is establishing across Canada to support AI and high-performance computing workloads. The project is part of a larger wave of infrastructure investments by Bell, which last year announced plans for a series of new AI data centers nationwide.
Beyond Winnipeg, Bell is developing a facility in Kamloops, British Columbia, in partnership with Thompson Rivers University, and is building a 300 MW data center campus in Saskatchewan. The company is also collaborating with Hive Digital Technologies’ Buzz HPC to deploy AI infrastructure in Merritt, British Columbia, and in Manitoba. These initiatives reflect a broader industry shift toward distributed, energy-intensive data center networks designed to handle the computational demands of generative AI and large-scale machine learning models.
The repurposing of the former food processing plant highlights the adaptive reuse of industrial real estate as data center developers face constraints on land, power, and construction timelines. By retrofitting existing structures, Bell can accelerate deployment while reducing the environmental and logistical challenges associated with greenfield builds. As AI adoption accelerates across sectors, such conversions are likely to become more common, particularly in regions with available industrial space and access to power infrastructure.
Source: datacenterdynamics
