TORONTO — If Canada wants to compete in the global AI race, it must overhaul the way it buys from and backs businesses, and focus its support on the most promising firms and applications of the technology, an AI advisory group told the federal government. 

In memos submitted late last year, several members of the AI Strategy Task Force assembled to advise on the national AI plan, recommended Ottawa launch new funding programs and tweak existing ones, make changes to the tax code, and update procurement requirements. Some memos called for the federal government to choose a small number of AI firms to champion, and to concentrate on fostering adoption of the technology in sectors where it can have the biggest impact.

AI Minister Evan Solomon appointed the 28-person task force composed of founders, investors and researchers in late September, giving them a month to make proposals to advise on the upcoming renewal of the national AI strategy.  All of the submissions made by the task force can be viewed here.

Canada has a poor track record of converting its strength in AI research into revenue-generating products and businesses, and in producing companies that become global giants, according to many submissions from the task force. Many said existing federal support systems are ineffective or insufficient at helping firms start and scale up. That’s been a long-standing preoccupation of Canadian policymakers and tech executives across governments and technology funding cycles. 

In his submission, League CEO Michael Serbinis called for Ottawa to set up new financing programs worth a combined $9 billion. That includes $2 billion for a fund-of-funds that would seed venture capital funds that back startups in pre-seed and seed rounds, followed by a $2 billion AI growth fund that would back them at the Series A and B stage. 

He also recommended a $5 billion sovereign wealth fund that would put between $50 million and $500 million into growth-stage deals. Under Serbinis’s proposed model, the federal government would put up some capital for the fund, and could offer tax credits to encourage Canadian pension funds and banks to participate. The new programs would help at least triple the number of upstart firms receiving financing, and produce more very large companies over the next decade, he wrote.

Inovia Capital partner Patrick Pichette also called for measures to raise more money for the AI sector from institutional investors. The federal government should require Canada’s eight largest pension funds to each put one per cent of their assets for the next five years into a new deep tech VC fund, he wrote. Ottawa should eventually mandate that the so-called Maple 8 allocate five per cent of their capital to “tech national priorities and next-generation industries.”

Task force members also recommended changes to existing federal supports like the popular Industrial Research Assistance Program and Scientific Research and Experimental Development (SR&ED) tax incentive, to simplify the application process and ensure companies can use them to help pay for AI innovation. 

For example, while firms can get SR&ED credits when they buy chips to train AI models, they can’t write off the cost of fine-tuning those same systems for particular sectors, according to Cardinal Policy founder Marc Etienne Ouimette. The rules should be adjusted to reward the “smaller scale experimentation and adaptation required to properly deploy an AI solution,” the former Amazon Web Services executive wrote in his memo. 

Some founders and investors on the task force proposed changes to taxes on capital gains that they claimed would help keep entrepreneurs and the firms they start in Canada, amid renewed fears of brain drain. 

Proposals ranged from CoLab CEO Adam Keating’s recommendation to eliminate the tax entirely for founders and staff at small companies, to Build Canada chair Daniel Debow’s suggestion to raise the exemption for those same people to $25 million per business. “Canadian talent is leaving because their economic upside is greater elsewhere,” Debow wrote. “This change will rebalance the economic playing field.”

Other memos called for Ottawa to use the tax code to encourage businesses to adopt AI. Businesses in “key industries that integrate AI into their core operations” should receive incentives, and be able to write off their capital costs sooner, wrote Cari Covent, a former longtime Canadian Tire executive. Serbinis called for a new tax credit for firms that buy from Canadian AI developers. During the 2025 federal election, the Liberals promised a rebate for small and medium-sized businesses worth 20 per cent of their AI adoption costs, but have yet to implement the program. 

What Canada needs is not “yet another tax incentive or funding program,” but a concerted push to overhaul procurement and technology adoption in both the public and private sectors, said Sam Ramadori, co-president of non-profit AI lab LawZero. 

Canadian business leaders need to be told they’re “at risk of completely losing to AI-driven foreign competitors” if they don’t adopt the technology themselves, Ramadori said, citing the significant challenges he faced selling to domestic customers while leading BrainBox AI. HVAC giant Trane Technologies acquired the Montreal startup last January.

Many task force members want the federal government itself to buy more from Canadian AI startups and tech firms. Instead of giving out grants, Ottawa “can become a strategic first customer to these potential billion-dollar-valuation companies,” wrote Sonia Sennik, CEO of accelerator Creative Destruction Lab (CDL). Covent recommended updating federal procurement rules to prioritize AI tools made and owned by Canadian firms, and ensure they meet requirements for responsible use and explainability. 

Olivier Blais, co-founder of consulting firm Moov AI, proposed that Ottawa set up a new AI readiness fund, which would pay for projects to adopt the technology in the public service and subsidize them in the private sector. The program could provide performance-based awards, “encouraging continuous improvement rather than one-time upgrades,” Blais wrote.

Some submissions also called for Canada to be more selective about the firms and sectors on which it’s betting in the AI race, reviving another once-popular approach to innovation policy. Serbinis proposed a program to designate a handful of firms as “Canadian global champions,” unlocking investment from the prosperity fund and guaranteed government contracts. A new council that included business leaders would set criteria to identify worthy companies. The founders of those firms should get larger capital gains tax breaks and greater access to top federal officials, including Solomon’s phone number, Serbinis said. 

Samdesk CEO James Neufeld also called for a champions strategy, suggesting Canada pick a few AI firms to be part of its “national defence-industrial complex” and help export their technology to NATO and other allies as part of military procurement deals.

Pichette called for an even narrower selection. He said Ottawa should declare Cohere—in which Inovia is a major investor—Canada’s “national LLM champion, and fuel it with large-revenue contracts.” The Toronto-based AI firm should receive deals worth over $1 billion each for applications in the public service and defence, he said—Cohere’s technology could help larger departments like the Canada Revenue Agency or Employment and Social Development Canada deliver services faster and more efficiently, and meet the country’s military and security needs. 

Other countries haven’t been shy about designating champions, or helping them win their markets. U.S. investors in particular have “overcapitalized” their portfolio firms to signal they’ll outlast the competition, then used their networks in Washington and big business to help them land key contracts, Pichette said, calling for Canada to do the same. 

Task force members also said the updated strategy should focus on sectors where AI can have the most impact, or industries in which Canada already has significant market share and advantages over other countries. University of Toronto professor Ajay Agrawal proposed five “moonshots” that would aim to achieve specific targets for AI-driven improvement to key social issues. For example, a health-care focused project would try to cut the time it takes for a patient to receive a referral to a definitive treatment by tenfold. 

“Markets alone will under-invest in experimentation” because companies are worried about regulatory issues or helping out competitors, wrote Agrawal, who is also CDL’s founder. That’s why Ottawa needs to back the moonshots, he argued.

A non-definitive list of sectors on which task force members recommended the updated AI strategy focus includes agriculture, cybersecurity, defence, education, finance, health care, life sciences, manufacturing, natural resources, robotics and transportation. The tactics they recommended to increase AI uptake in those industries included adoption playbooks, public datasets that companies can use to train models, and so-called sandboxes in which they can experiment under loosened regulations. 

Solomon has said the Liberal government will be “unapologetic about championing our champions,” and signalled that it will seek to address issues of AI adoption, commercialization and trust. Last month, he said the updated AI strategy will come out this quarter.