In This Article:
Telus Communications Inc. has announced plans to build a pair of data centres powered by leading-edge Nvidia Corp. chips as the Vancouver-based telecom seeks to establish itself as a key player in Canada’s artificial intelligence infrastructure buildout.
This summer, Telus will launch an AI-focused data centre in Rimouski, Que., that will include a 10,000-square-foot server module equipped with up to 500 Hopper graphics processing units (GPUs) from Nvidia. Telus plans to deploy the latest Blackwell Nvidia GPUs — which are used for advanced AI workloads and compute-intensive tasks — sometime next year. The Rimouski facility — which it is calling a “Sovereign AI factory” — has the capacity for six additional server modules.
The company will open another facility in Kamloops, B.C., once the compute capacity in Rimouski has been utilized; it does not have a fixed launch date for its B.C. data centre. Company spokesperson Athyu Eleti said that the Kamloops factory is an existing facility that will require “minimal retrofitting” to accommodate the AI operations, allowing them to quickly bring it up to speed when needed.
Telus, the first Canadian telecoms player to take a crack at Canada’s AI compute market, declined to disclose how much it has invested in its AI data centres. Nvidia chief executive Jensen Huang told CNBC last year that Blackwell GPUs could cost between US$30,000 to US$40,000 per unit, while Hopper chips are priced around US$25,000 to US$30,000, according to analyst estimates.
Telus also said it is the first North American service provider to become an official Nvidia “cloud partner” — a deal that was announced last month and allows the telecom to integrate Nvidia GPUs and AI software to enable powerful and advanced AI model training and inference capabilities, Eleti said. Telus’ AI data centre clients include Canadian businesses, startups and researchers looking to develop smarter AI products and streamline operations.
The company’s AI data centre announcements come as technology groups and business leaders urge the next government to increase investment in and adoption of AI across the country, and as the AI race heats up worldwide.
The federal government last year launched a $2 billion Sovereign AI Compute Strategy that aims to boost domestic computing power for homegrown companies, startups and researchers. U.S. President Donald Trump has called on White House officials to develop a new AI Action Plan by July to ensure and maintain U.S. AI dominance. Meanwhile, U.S. tech giants are locked in a race to build the industry’s biggest AI data centres, from OpenAI Inc. and Microsoft Corp.’s US$100 billion Stargate project to Amazon Web Services’ Project Rainier that is also expected to hit a US$100 billion budget.