Canada commits $1.4B to sovereign compute infrastructure as it joins the AI arms race

Canada is one of the latest nations to catch the sovereign AI bug and plans to invest $2 billion CAD ($1.42 billion USD) to bolster the nation's domestic compute capacity by funding the development of new datacenters and computing infrastructure north of the 49th parallel.

The so-called Canadian Sovereign AI Compute Strategy comes after months of planning and public comment and aims to ensure local businesses, entrepreneurs, and researchers have access to the compute capacity they need to develop domestic AI products.

"The strategy we're announcing today is a major step toward securing Canada's place as a global AI leader," François-Philippe Champagne, Minister of Innovation, Science and Industry, said in a canned statement. "This is about Canada showing its ambition, supporting Canadian innovation, and investing in the economy of the future."

According to the office, Canada is already home to 10 percent of the world's leading AI researchers, and more than 140,000 professionals actively engaged in AI development in 2022-2023, a number it says has grown 29 percent compared to the previous year.

With 2 billion loonies to work with, the Land of the Midnight Sun aims to ensure its AI talent has the compute resources necessary to continue their work, while also safeguarding Canadian sovereignty by ensuring data is processed and stored within its borders.

The funding itself will be split three ways, with up to $700 million CAD ($498 million USD) allocated through the "AI Compute Challenge" to develop new or expand existing datacenters in the region.

These funds are intended to serve as a monetary incentive for datacenter operators to build out large AI compute clusters in Canada, likely similar to those being developed by CoreWeave and Lambda across the US.

This funding is open to all manner of commercial entities, industry consortia, and public-private partnerships involving AI compute infrastructure projects. However, it is worth noting that this funding doesn't appear to be intended for those pursuing experimental or research ventures.

"Preference will be given to proposals leading to fully integrated AI datacenter solutions ready for commercial deployment," the AI Compute Challenge documentation explains.

Another $1 billion CAD ($711 million USD), meanwhile, will be used to build "transformational public computing infrastructure." A big chunk of this will go toward the construction of an AI supercomputing facility. The Canadian government plans to begin soliciting proposals for the project early next year with priority given to those companies willing to front some of the cash. In the nearer term, $200 million CAD will be set aside to augment existing public infrastructure.

Finally, up to $300 million CAD ($213 million USD) will be set aside to provide affordable access to compute infrastructure for small and mid-sized businesses through the AI Compute Access Fund.

Canada is only the latest in a string of nations pushing for sovereign AI investments, which emphasize home-grown models trained in domestic datacenters. We've seen similar strategies pop up in India, Japan, and Europe.

But while these efforts are expected to contribute meaningfully to the local economy as new products reach the market, in the short term, US chipmakers and OEMs stand to make a killing on the deal, which may be why Nvidia CEO Jensen Huang, arguably the biggest beneficiary of the AI boom, has been pushing the concept so hard over the past year or so. ®