Federal government outlines $2 billion in AI compute spending commitment

The federal government has launched the Canadian Sovereign AI Compute Strategy, outlining how it intends to deploy the $2 billion CAD it promised for artificial intelligence (AI) computing power as part of Budget 2024.

This funding has been allocated towards increasing the computing power available to Canada’s AI researchers, startups, and scaleups. The Liberals claim this strategy “will meet the short-term, medium-term, and long-term compute needs” of domestic players as they conduct AI research and develop made-in-Canada AI products.

$300 million will go to affordable AI compute for SMBs.

AI compute, which is typically provided by chips and data centres, refers to the energy-intensive computational resources required for AI systems to perform tasks like processing data, running algorithms, and training machine learning models.

According to Innovation, Science, and Economic Development Canada (ISED), $1 billion will fund “public” supercomputing infrastructure. This includes the establishment of a large supercomputing facility for researchers and companies and a smaller computing facility led by Shared Services Canada, the agency that provides information tech to federal departments, as well as the National Research Council of Canada (NRC). That facility will be designed for government and industry research and development, including for national security purposes.

The feds will begin seeking partners to build, manage, and run this large sovereign supercomputing facility in Spring 2025, with preference for proposals that leverage private capital to increase its scale and co-locate commercial computing infrastructure. 

As building both of these facilities will take time, $200 million of this $1 billion will be used to augment existing public computing infrastructure managed by NRC, Canada’s three AI research institutes, and the Digital Research Alliance of Canada so as to address the country’s immediate needs.

RELATED: CIFAR leader expects CAISI to help inform AI policy in Canada and abroad

To support these efforts, another $700 million has been dedicated to financing the build out or expansion of commercial AI data centres in Canada, with the expectation that they will provide substantial capacity and flexible and affordable computing offerings to Canadian companies.

These funds will be distributed by the Strategic Innovation Fund via the AI Compute Challenge to companies, consortiums, and academic-industry partnerships. Priority will be given to Canadian projects that can demonstrate sustainability (including the use of environmentally sustainable energy sources and Canadian-made AI compute hardware or software) and a high rate of return on public investment.

The remaining $300 million will go towards providing affordable access to computing power to small and medium-sized businesses (SMBs) through the AI Compute Access Fund. This fund will support the purchase of AI compute resources by Canadian businesses to meet their near-term needs as the feds seek to build more long-term domestic compute capacity. More details on the AI Compute Access Fund will be shared when it is launched in Spring 2025.

In a statement, Canada’s Minister of Innovation, Science, and Industry François-Philippe Champagne described this strategy as “a major step toward securing Canada’s place as a global AI leader.”

RELATED: Federal government commits $2.4 billion to AI compute, startups, and safety through Budget 2024

“We’re proud to be a driving force in the worldwide ecosystem, and by increasing access to domestic and secure compute capacity, we will help businesses, innovators and researchers boost the Canadian economy and stand out on the world stage,” Champagne said.

These plans follow public consultations held during the summer that saw input from more than 1,000 stakeholders. According to a Government of Canada report summarizing its findings, respondents argued that there is an urgent need to invest in AI computing infrastructure to lay the groundwork for the country’s AI ecosystem and secure its supply chain. 

The feds said industry stakeholders cited high cost as the biggest barrier to accessing AI compute, while researchers singled out Canada’s lack of publicly available high-performance computing resources as a problem.

Stakeholders: buying local, going green will be key

Stakeholders BetaKit spoke with reacted favourably to the federal government’s AI infrastructure plans. “Expanding compute capacity is crucial to unlocking the vast opportunities ahead as AI continues to drive the digital transformation of the economy,” Ben Davies, chief information officer at Toronto’s Vector Institute for AI, told BetaKit over email.

Yuri Navarro, advisor at Markham, Ont. deep tech incubator VentureLab, called it “a crucial step toward strengthening Canada’s global competitiveness in [AI].” Navarro told BetaKit via email that “Initiatives like the $300-million fund to make computing power affordable for small and medium-sized businesses and the $700 million for commercial AI data centres will drive innovation and collaboration across industry, academia, and the non-profit sector.”

Bob Beachler, vice-president (VP) of product at Toronto semiconductor company Untether AI, expressed excitement that investments will favour projects and companies developing innovative and environmentally sustainable Canadian components. “The reality is the widespread deployment of AI will be a drain on energy resources,” Beachler told BetaKit over email. 

RELATED: Untether AI’s new CEO is here for global scale

The International Energy Agency (IEA) recently estimated that data centres, together with crypto and AI, consumed nearly two percent of global electricity demand in 2022. The IEA projects that this group’s energy consumption could double by 2026.

“Favouring operators that use new and energy-efficient hardware in their facilities is a non-negotiable for our energy resources,” Beachler added. Untether AI hopes to help mitigate the environmental impact of data centres with its energy-efficient AI chips, and Beachler argued that this must remain a focus as the country builds out these AI data centres.

Dana O’Born, VP of strategy and advocacy at tech scaleup advocacy organization The Council of Canadian Innovators (CCI), described this funding as “a positive step toward bolstering Canada’s AI capabilities,” but noted that, “The success of this strategy will depend on meaningful engagement with Canadian innovators and experts.”

“It is essential to ensure that the allocated funds are deployed effectively and efficiently to maximize their impact on Canadian companies and not simply flow to foreign AI companies or large multinational infrastructure providers,” O’Born told BetaKit via email.

AI a federal priority despite stalled regulation

The feds have made AI a priority. This $2 billion in federal spending on AI computing comes in addition to $405 million committed by the Liberal government in April to help AI startups bring new tech to market, boost SMB AI adoption, and fund a recently launched Canadian AI Safety Institute (CAISI), among other priorities.

The Canadian Sovereign AI Compute Strategy is the latest in a series of AI initiatives rolled out by Canada’s Liberal government, joining the Pan-Canadian AI Strategy, the Voluntary Code of Conduct on the Responsible Development and Management of Advanced Generative AI Systems, and the proposed Artificial Intelligence and Data Act (AIDA).

AIDA, or Bill C-27—which passed its second reading in the House of Commons in April 2023—has been at consideration in committee since September 2023. Canadian AI leader Yoshua Bengio has lamented Bill C-27’s lack of progress, while former minister of justice and attorney general David Lametti is “not optimistic” it will receive royal assent.

With files from Aaron Anandji.

Feature image courtesy Justin Trudeau via Flickr.