As federal budget doubles down on AI funding, industry leaders hope for greater push to commercialize

With a new federal budget doubling down on existing artificial intelligence programs and emphasizing intellectual property, some industry leaders hope the investments will help Canada maintain its early lead and create AI champions in the country.

Leaders of AI companies and research organizations told Research Money that federal funding for boosting intellectual property education and increasing computational power could help to build and grow Canadian AI startups in Canada. However, it's not yet known whether funding will tackle the long-standing challenge of commercializing AI technology from home.

Some of the budget’s notable AI investments include $443.8 million over 10 years for the Pan-Canadian Artificial Intelligence Strategy and $60 million over two years to the Innovation Superclusters Initiative, which includes an AI supercluster. The Pan-Canadian Artificial Intelligence Strategy, first announced in 2017, was the first national AI strategy in the world and led to the creation of three AI institutes (MILA, Vector Institute and AMII).

While not AI-specific, the government is investing $90 million to create ElevateIP, a program to help accelerators and incubators provide startups with access to expert intellectual property services. The Strategic Innovation Fund, which provides funding to large-scale projects in streams like research and development, will also receive $7.2 billion over seven years starting this fiscal year.

“We're seeing these small steps that these governments are taking that are extremely encouraging,” Yvan Couture, CEO of Waterloo-based AI company Primal, told Research Money in an interview.

While it’s still too early to know how the funding will be executed, industry leaders were cautiously optimistic and offered feedback on how the government can best support researchers and companies.

From funding to commercialization

When the Pan-Canadian AI strategy was created in 2017, it was more “aspirational,” Garth Gibson, CEO of Vector Institute, told Research Money in an interview. At the Vector Institute, for example, “there was no organization, there was no building, there were a couple people that were on staff," he said.

Now, the Vector Institute’s network includes more than 1,000 active researchers and more than 1,000 AI master’s students graduating each year in Vector-recognized AI programs and study paths, and the institute has published over 270 papers in global conferences and journals. With more AI talent in the country, Gibson says the federal government’s investment towards dedicated computing capacity for researchers at MILA, AMII and the Vector Institute, totalling $40 million over five years, is important for supporting researchers since AI applications need intensive computational power.

“The assumption that the computing needs of AI could be absorbed into the existing research computing needs is no longer sufficient,” says Gibson. “The way that AI uses computers is different from the way climate change researchers use computers, so they would be optimized differently."

Working on IP could keep companies in Canada

With the creation of ElevateIP and $75 million over three years to the National Research Council’s Industrial Research Assistance Program (NRC-IRAP) to provide high-growth companies with access to expert intellectual property services, the hope is that more companies — including in AI — can grow and keep their capabilities in the country.

Yvan Couture at Primal, whose company holds one of Canada’s leading AI portfolios with 145 patents, says he hopes that educating students, companies and even venture capitalists can help the country see more public companies in Canada.

He says Canada has long had a mindset of a “branch plant economy,” where the government is mostly focused on jobs created. “If I'm a researcher at Google in Canada, the proceeds of my work accrue entirely to Google,” he said.

And for companies with venture capital funding, there can be pressure to sell to the highest bidder. “We've resisted selling to American companies, and I've been fortunate enough that our shareholder base has been patient enough to try to help us find a stay-in-Canada solution," Couture added.

Companies selling to foreign partners has been a common refrain in the Canadian tech ecosystem. One recent high-profile example is Element AI, which sold to ServiceNow for US $230 million last year after garnering hype with AI pioneer Yoshua Bengio among its founders and millions in public and private investment — including from the federal government.

By educating startups to protect their intellectual property assets, the hope is that companies will grow within the country rather than selling early. However, it may not be enough if companies aren’t commercializing here.

“A lot of the technology comes from Canada, but it's not commercialized here. It's taken down to the US,” Stephany Lapierre, CEO of Tealbook, which uses AI to help procurement teams get more visibility into their data, told Research Money in an interview. “Moderna is a technology that's Canadian, but they’re not a Canadian company, [for example].”

More funding for superclusters, SIF

Lapierre at Tealbook says that the existing programs that received renewed funding are difficult for startups like Tealbook to access. She once put forward a proposal for Scale AI supercluster funding, but says the way the funding was structured was complex.

“I'd love to see more accountability on the types of KPIs provided, and what it has done for the Canadian AI ecosystem,” says Lapierre.

In October 2020, the Superclusters faced criticism from the Parliamentary Budget Officer (PBO)  for being $74 million behind schedule and “unlikely” to create 10,000 jobs. The heads of the superclusters have since announced more projects in the wake of COVID-19, with the NGen supercluster nearing its $230 million funding goal as of January 2021.

The Strategic Innovation Fund, one of the government’s most significant programs with $3.4 billion in spending to date, has also failed to improve Canada’s innovation performance, innovation and policy experts told Research Money. According to Research Money's analysis, $1.36 billion, or 40 percent, of all SIF funding has gone to companies and joint ventures in Canada whose parent company is based in a foreign country.

Meanwhile, Tealbook is finding customers and funders outside of Canada. As COVID-19 disrupted supply chains in 2020, the UK government used Tealbook's services to help them find qualified PPE suppliers and the Brooks Brothers, a US clothing company, sought Tealbook's help in sourcing materials for N95 masks. But Lapierre says they haven't been able to work with the Canadian government.

The majority of Tealbook’s institutional investors are U.S.-based, with the exceptions of StandUp Ventures and Business Development Bank of Canada, Lapierre says.

Following Canada’s AI strategy announcement, many U.S. tech giants moved to build AI research labs in the country, including Uber, Google and Microsoft. Often, Couture says we’ve celebrated companies selling — which he notes is not a bad thing — but companies are becoming more aware that there are opportunities in going public.

“We raise these big massive amounts of money, but it's all coming in from outside,” says Couture. “To me, we're building this extreme value, and we're letting that suck out of our country.”

There is $185 million over five years to help “support the commercialization of artificial intelligence innovations and research in Canada,” per the government budget, but it’s unclear how that will be executed. To help AI startups commercialize, Lapierre says the government could help the Canadian private sector become early adopters in emerging technology by de-risking investment in early-stage companies.

The government itself could be a partner to startups by digitizing the procurement process, which Lapierre calls "inefficient," and making it easier for the government to buy home-grown technologies to solve problems.

“I think it’s setting goals and accountability, and making it a lot easier for the government to buy innovation and become a customer,” says Lapierre. “If we have early customers who are Canadian customers, we can help them become a market leader, but also it helps us build better solutions, which turns into better companies, and turns into better chances of success.”

Gibson says he has not yet had a briefing call with Innovation, Science and Economic Development, but that the renewed funding shows that the government recognizes the work the community has done so far.

“That tells me that there's an appetite to continue that real world stuff," says Gibson. "When you pair that with the signal around computing infrastructure, that's a great trajectory."

R$