Canadian banks have been the TSX’s version of the U.S. AI trade — the single sector carrying the index. The S&P/TSX is positioned to beat the S&P 500 for a second straight year (its first back-to-back win in fifteen years), and just as U.S. gains have been driven by chipmakers, Canada’s have been driven by banks.

The magnitude is striking: the Big Six are up +33% so far this year (almost +70% on a year-over-year basis), but the ex. bank index is not even up +6% (the TSX headline has advanced nearly +12%).

Let me go one step further.

The share of the TSX that is now dominated by the banks just crossed above a 25% market share for the TSX, which is unprecedented and represents a near 2.5 standard deviation event. Just so that you know where I am going with this.

Like the AI trade in the S&P 500, the Canadian Banks in the TSX have entered bubble territory. If you have been long and lucky, now is the time to book some profits.

Canada’s bank stocks are historically expensive. No one seems willing to bet against them

Here’s what’s actually behind the bank bubble burst of bullishness — and then the reasons to be skeptical — because this is a case where the fundamentals are genuinely good, but probably don’t justify the full move. Not nearly.

The single biggest driver is the credit-loss story shifting into reverse. Heading into 2025, the consensus fear was a mortgage-renewal cliff — roughly two million borrowers from the 2020-2021 record-low-rate vintage coming up for renewal at rates 2%-3% higher, heading into a real estate deflation cycle, with attendant predictions of a default wave. We were expecting a negative credit cycle too, but it never materialized. Rate cuts in 2024 and 2025 eased the credit stress, and the predicted surge in defaults proved way overblown.

The result showed up directly in 2026Q2 earnings as sharply falling provisions: RBC’s PCL (Provision for Credit Losses) fell 36% on a YoY basis, BMO’s fell 28%, and TD’s fell 20%. Falling loan loss provisioning flows straight to the bottom line.

That produced across-the-board earnings beats. From the fiscal 2026Q2 prints: RBC net income +25% (EPS +27%), BMO net income +34% (EPS +41%), CIBC +23%, and TD +15%. Return on equity expanded meaningfully — CIBC’s adjusted ROE hit +16.4%, and BMO’s jumped to +13.5% from +9.8% a year earlier. And the banks rewarded holders with dividend hikes (RBC +7.3%, TD +3.7%, BMO +2.4%), which supports the shares in a market that prizes Canadian dividend yields and payout ratios.

Two other tailwinds: capital-markets/trading revenue has been strong (RBC’s revenue +19%, CIBC’s +21%), and the U.S./international growth engines — TD’s East Coast franchise, Scotiabank’s pivot toward mature North American markets and its KeyCorp stake, RBC’s wealth division, and U.S. expansion — have added to the story.

Now for the skeptical read on the situation.

The valuations are the tell. The Big Six Banks have historically traded around 11x earnings; today, the average is roughly 15x on expected 2027 earnings — a level rarely seen in the past. The operating backdrop, as solid as it is, doesn’t support a re-rating of this size.

As for earnings quality, it is worth scrutinizing. Falling PCLs are a powerful EPS lever but a non-repeatable one — that kind of provision-release boost that typically is done once. Once provisions normalize, you need actual loan and revenue growth to carry earnings, and that’s precisely what has been soft. Leaning on capital-markets revenue is also fragile: trading and advisory fees are historically volatile and a low-multiple line of business, which is why the rating agencies and OSFI have discouraged banks from over-relying on them — yet investors are treating soaring capital-markets profits as the new normal.

Meanwhile, a couple of specific risks should be on our radar screen, as in the recent run-up in bond yields. This is the elephant in the room that could reawaken the mortgage-renewal problem. Fixed mortgage rates track government bond yields, which have surged as markets price potential BoC rate hikes to combat oil-driven inflation. The banks avoided the mortgage cliff in 2025, but that might not hold into late 2026 and 2027.

The momentum risk is the one that should worry a fundamentals-oriented investor most. The rally has been increasingly momentum-driven. Just like U.S. Tech and Asian semiconductors. So if you have been late to the trade like us, it doesn’t make much sense to jump in at this point and follow the herd. Patience will be virtuous.

For those investors who rode the Canadian bank wave, perhaps think about trimming your oversized holdings and keeping them at an appropriate weight in the portfolio. And let’s not confuse a series of one-offs that have boosted earnings with something we should be extrapolating into the future, which the pricing action in this last leg up has come to signify. All the more so at P/E multiples rarely seen before — not to mention a price-to-book ratio of 2.7x versus the ten-year average of 1.7x. That is a near-60% premium!

This all looks more like a late-stage momentum move on a good-but-not-great fundamental base rather than a durable re-rating. We shall wait for valuations to mean-revert and the momentum component behind the rally to subside before making a decision to move into this space.

David Rosenberg is founder of Rosenberg Research.