There's a moment in every poker game when someone pushes all their chips to the center of the table. Last week, RBC and TD did exactly that; except instead of chips, it was billion-dollar AI commitments with their names attached.
Here's a fun fact that should terrify anyone clinging to legacy systems: global corporations threw $252.3 billion at AI in 2024. An MIT study dropped a bomb in July, revealing that 95% of those organizations get zero dollars back. Not "less than expected." Not "still calculating." Zero.
So when RBC steps up and says "we're delivering $700 million to $1 billion by 2027" and TD echoes with "$1 billion annually," they're not just making predictions. They're drawing a battle line between institutions that know what they're doing and everyone else who's still figuring out which end of the AI stick to hold.
Why Most Banks Are Still Playing Dress-Up
Only eight banks globally have disclosed their expected AI returns….Eight…Out of 50 major international lenders. The rest still hide behind corporate speak about "exploring opportunities" and "strategic initiatives." Translation: we have no idea if this is working.
The Evident AI Banking Index doesn't sugarcoat it: disclosing ROI "reflects genuine operational maturity rather than optics." In other words, if you're not putting numbers on your AI strategy, you're not actually deploying AI; you're cosplaying innovation.
RBC didn't just wake up one morning and decide to commit to a billion-dollar target. They've been grinding since 2016, when they launched Borealis AI, their research institute. That's almost a decade of groundwork before they made their public bet. Meanwhile, TD's been racing to catch up after sliding from 9th to 13th in global AI maturity rankings. Their billion-dollar annual forecast? That's a comeback story with receipts attached.
The Winners Are Pulling Away—Fast
Canada held onto its crown as the region with the highest scoring for banking AI maturity outside the U.S. But here's the uncomfortable truth lurking beneath that victory lap: the gap between winners and losers turns into a canyon.
Scotiabank nosedived nine positions to 29th place. Canada's Big Five can no longer claim they're all in the top 25. BMO climbed five spots to 19th, but only after finally disclosing what they're actually doing with AI. CIBC stayed steady at 22nd, which, in a rapidly evolving race, means they're effectively falling behind.
The report doesn't mince words: "The biggest banks have the resources to build research labs, patent portfolios and talent pipelines—creating a sharp divide between leaders and mid-tier banks that cannot keep up." This isn't a participation trophy economy anymore. It's Darwinian.
The Bubble Question Everyone's Dancing Around
Fair question: are we watching the dot-com bubble 2.0, electric boogaloo edition?
The parallels are uncomfortable. In the late '90s, telecom companies laid 80 million miles of fiber optic cable across the U.S. after overestimating demand. Most of it sat dark and unused for years while companies went bankrupt. Today, tech giants are building AI data centers that could cover Manhattan. Meta, OpenAI, Oracle, everyone's betting hundreds of billions on infrastructure for demand that might not materialize for years.
Here's the twist: Canadian banks are betting they're different. They're not building speculative infrastructure but hope customers will appear. They're deploying AI into existing workflows with measurable outcomes. RBC has documented use cases spanning personal finance to capital markets. TD's automation targets are tied to specific cost reductions and revenue gains.
That's the difference between a bubble and a breakout—execution with accountability.
What This Means for Everyone Else
You've already lost if you run a financial institution and still treat AI like a research project. The leaders aren't two years ahead anymore; they're two paradigms ahead.
RBC spends $5 billion annually on technology. That's not a budget line—it's a war chest. They've built an ecosystem where AI research moves from academic papers to production deployment in under three years. That's lightning speed in banking.
For the laggards? The message is brutal: your legacy systems aren't just outdated. They're anchors dragging you to the bottom while your competitors are already sailing.
The Strategic Question You Should Be Asking
The following 18 months will be definitive. RBC's target date is 2027. TD's annual forecast starts now. Every quarterly earnings call will be a referendum on whether these bets pay off.
But here's the real question: when the dust settles and we see who hit their numbers and who missed, will your institution be on the side that executed or the side that hesitated?
Because in a game where eight out of 50 banks have the confidence to put numbers on their AI strategy, being in the other 42 isn't neutral. It's a red flag that screams "we don't actually know what we're doing."
The poker game is still playing out. But some players just showed their hands. And they're holding a lot more than everyone thought.