A former TD capital markets analyst is talking about what ultimately caused her to leave on a Tuesday afternoon in a coffee shop on King Street West, which is close enough to view the Bay Street skyscrapers from the window but far enough away to feel like a different planet. She claims that it wasn’t the hours. Actually, it wasn’t the money. She spent three hours arguing to a risk committee why a specific data strategy was sound, even tho she knew the decision would be no before she entered the room. Three months later, she joined an AI business in Toronto that specialized in financial fraud detection. Compared to her yearly bonus, the equity package was less. Nevertheless, she accepted it.
On Bay Street, this is not an isolated incident at the moment. The Canadian banking industry, which is stable, profitable, well-known throughout the world, and heavily regulated, is witnessing a subtle but significant exodus of highly skilled analysts to the city’s developing AI ecosystem. The disillusioned junior analysts who constantly stray are not the ones departing. They are mid-career experts with real competence in finance who are leaving because they believe their knowledge is more valuable when applied somewhere where it can proceed more quickly.
Individual actions are driven by rather obvious financial calculations. In a successful year, a managing director at one of the Big Six makes between $500,000 and $1.2 million. That is actual cash. However, a senior hire who takes a 1.5 percent equity position in a business that develops from $100 million to $500 million in valuation has made more than several successful years on Bay Street condensed into a single liquidity event at an early-stage AI startup with a solid product and an addressable market. Silicon Valley has had access to that computation for many years. It is currently accessible in Toronto, where a number of venture-backed fintech firms, including Cohere and Ada, have grown to the point where significant stock grants are truly valuable. Five years ago, the exit math was different.
The majority of public conversations on this dynamic undervalue the regulatory aspect. The Office of the Superintendent of Financial Institutions (OSFI) oversees Canadian banks under a framework that puts systemic stability ahead of operational flexibility. In all honesty, it is what you want from the organizations that are funding your mortgage and storing your deposits. If you’re a technically ambitious professional, it’s much less compatible with the kind of iterative, experiment-fast culture that makes working on AI products exciting. Before a single line of code even comes close to a real client, a banker wishing to develop a machine learning model for credit underwriting in a major Canadian bank will have to spend months navigating explainability standards, model risk management committees, and regulatory disclosure frameworks. A comparable product may be in pilot at a startup in a matter of weeks. Neither strategy is incorrect. They draw a variety of individuals.
In this talent competition, Toronto has significant advantages that are frequently overlooked. Few places outside of San Francisco and London can match the skill density of the city’s truly deep AI research ecosystem, which includes the Vector Institute, the University of Toronto’s machine learning community, and the commercial research divisions of firms like Google and NVIDIA. One of the biggest concentrations of AI and technology talent on the continent may be found within the Waterloo-Toronto region. In contrast to a Bay Street departure to San Francisco, bankers who are relocating to AI startups in Toronto are not abandoning their professional networks. They are moving across a few streets and joining businesses where a number of their former coworkers are already employed.
Large, regulated, legacy-system-dependent companies have been grappling with a similar issue for the past ten years. This is what the banks are navigating. Talent with the technical know-how and willingness to take risks necessary to create competitors is becoming the talent they most want to keep. TD Bank purchased Layer 6, an AI startup founded by technologists working in the banking industry, in 2018 because the bank was unable to quickly build that capability internally. Time was bought by the acquisition. The fundamental cultural divide between how TD functions at scale and how machine learning system developers wish to work was not resolved.
