The AI Innovation Gap: U.S. Business Booms While Canada Lags
Artificial intelligence is supposed to make everything easier, including starting a business. Chatbots can serve as advisers, while AI applications can speed up many core functions of a business. Tech start-ups can grow faster with fewer employees and less funding than in the past, too.
It stands to reason, then, that AI should enable more people to start businesses. There are early signs of a trend in the United States, even if it’s not yet definitive. But in Canada? Not so much. That does not bode well for the domestic tech ecosystem in particular, which is already struggling to retain founders and hatch lasting businesses
Some have posited that AI is already fuelling more business creation, while OpenAI chief executive officer Sam Altman has speculated that billion-dollar companies will be spawned and overseen by lone operators.
Torsten Slok, chief economist at Apollo Global Management, wrote earlier this year that business formation in the U.S. is “exploding,” pointing to an upward trend in the number of applications filed to start companies. Monthly applications jumped 21 per cent between ChatGPT’s release in November, 2022, and April, 2026. “The surge in new U.S. business formation is being fueled by AI and large language models that are dramatically reducing the cost and complexity of launching a company,” Mr. Slok wrote.
But that figure is essentially tracking paperwork, and not all of those applications will result in functioning businesses. The U.S. Bureau of Labor Statistics monitors what it calls births, meaning companies that record employment growth for the first time in a given quarter. That number has actually fallen about 8 per cent, to 323,000 in the third quarter of 2025, from 351,000 in the same period of 2022. The quarterly birth rate, meanwhile, essentially measures the portion of new companies among all establishments. It’s averaged about 3.5 per cent since ChatGPT was released, just a touch higher than it was in the preceding years.
There’s a lag between filing to start a business and hiring employees, of course. Fortunately for our purposes, the U.S. Census Bureau tries to bridge that gap by projecting the number of new businesses (defined as those with payroll tax liabilities) that will form within one year of any given month, based on the number of applications. That number is indeed growing, averaging about 17 per cent higher for the first four months of the year compared with the same period in 2022.
Still, this doesn’t tell us much about the role of AI. Emin Dinlersoz, principal economist with the U.S. Census Bureau, declined to speculate on why business applications are rising when e-mailed for comment.
There could be a few factors. Some research has shown that the expansion in unemployment benefits in the U.S. during the pandemic contributed to a surge in business formation, while the rise in remote work might be playing a role, too.
Avi Goldfarb, a professor at the Rotman School of Management at the University of Toronto, said that AI is a plausible explanation, but it’s too soon to draw conclusions. “We won’t be able to unpack that for a while,” he said, adding that it took about a decade to figure out how the internet shaped the economy. (Maybe AI is just making it easier to file paperwork, he suggested.)
The notion that AI could lead to more business creation is sound, he continued, but there is a flip side. The technology could also help big, powerful companies become even bigger and more powerful. Stanford University economist Erik Brynjolfsson and Harvard fellow Zoe Hitzig made that argument in a paper last year. “Centralization is likely to intensify as AI capabilities advance,” they conclude.
Here at home, Statistics Canada estimates the number of entrants, referring to businesses with employment in the current month that were not previously active. The figure strips out seasonal reopenings and instances where a large company forms a new entity. Between 2015 and 2020, Canada averaged about 15,687 entrants a month. Since ChatGPT, the average is not much higher at 15,730. The pace has been pretty steady, aside from pandemic-era disruptions.
“You would think we’d already be seeing new business formation. It’s been three years now, and the lower cost is there,” said Joel Blit, associate economics professor at the University of Waterloo. Business adoption of AI in Canada is still relatively low, while the country hasn’t done much to promote AI literacy, which could be factors.
“If you’re trying to create thousands of little startups, you really need to get everyone understanding the potential,” he said. The entrepreneurial drive has to be there, too; it might be unlikely for someone who has never thought about starting a business to do so simply because ChatGPT is around.
Prof. Blit knows first-hand how helpful AI can be when running a business. He and his wife, family physician Lisa Friars-Blit, run a skin cancer screening clinic, and he considers ChatGPT to be another co-founder. It’s helped to develop marketing plans, design a logo and provide legal advice about the clinic’s lease to prepare for meeting with an actual lawyer. “It has saved us tons of money when it comes to lawyer fees, marketing, all kinds of stuff,” he said.
Clinics have existed before AI, of course. The real promise of any transformative technology is that it will create entirely new markets and spawn businesses that weren’t possible before. That is becoming true with AI, from companies such as Ada Support Inc. that make customer service chatbots to fanciful outfits like Cowboy Space Corp., which aims to blast data centres into orbit.
While Mr. Dinlersoz of the U.S. Census Bureau did not want to speculate about surging business applications, he did point to a study of his that found filings for new AI-related companies began rising around 2012 and then surged in 2023, consistent with the boom kicked off by ChatGPT. He and his co-authors found these applications are more likely to transition to startups and achieve higher revenue and employment rates, though they have a lower survival rate.
In Canada, the conditions are not favourable for a sustainable AI startup boom at the moment. One study last year by the investment firm Leaders Fund pointed out that tech founders are leaving the country for the U.S. at a higher rate than in the past. San Francisco is the epicentre for AI, capital is more abundant and founders say there are fewer regulatory hurdles. RBCx, the technology banking arm of Royal Bank of Canada, found that Canadian venture capital firms raised just over $2-billion last year, the worst showing since 2016. Toronto, Vancouver and Montreal have all fallen in a ranking of global tech ecosystems in recent years, according to data tracked by advisory firm Startup Genome.
One company that embodies all of these AI startup trends is uRun, which could not exist without the technology. The company is known as an inference provider, and has devised a way to generate AI video faster to allow for interactive editing. Its five employees are heavy users of AI, too. “I don’t think there’s any way even a company of 50 people could build what we built without modern AI tools,” said co-founder Keegan McCallum.
Finally, it’s an example of the pull of the U.S. Mr. McCallum is based in Victoria, but he’s working on moving to San Francisco. “The go-to-market and support parts of the business will probably need to be there for the customers, because that’s where they all are,” he said.
But one hopes it’s not where all the AI-fuelled companies will be created, eventually.
Sign up for our
This article was first reported by The Globe and Mail
