CMHC's 4.69 Million Homes Is Now a Ceiling, and That Changes the Math
Two words now sit in front of CMHC's 2036 housing projection: "up to." As of September 10, 2026, CMHC puts Canada's need at up to 4.69 million new homes by 2036, and in the same breath it flagged that construction could slow. That qualifier turns the number from a target into an upper bound, and most of the people quoting it haven't adjusted.
The floor everyone has been selling
The common belief goes like this: Canada is millions of homes short, the gap can't close for a decade, so prices and rents have a permanent backstop. You hear it at presale launches, in investor seminars, and from homeowners in Kelowna and the GTA explaining why they'll never sell. The national shortage figure gets treated as a floor under every property in the country.
It was a reasonable story while the number read like a fixed debt. Back in 2023, CMHC estimated Canada needed 3.5 million additional units by 2030, and that figure got repeated as if it were a bill coming due. Now the horizon is 2036, the number is 4.69 million, and the wording says the real figure could be lower. A floor can't come with "up to" attached.
Where the shortage logic breaks
Run the arithmetic. Spreading 4.69 million homes across roughly ten years means something like 469,000 a year. CMHC counted 259,028 housing starts across Canada in 2025, with Ontario contributing 62,561 of them. At the top of the range, the country is building a little over half of what the projection calls for.
So the bulls win? Not quite. A need estimate describes how many homes it would take to reach a certain outcome. It says nothing about who can carry a mortgage at the asking price, or whether a renter in a particular building can absorb another increase. A national gap of millions can coexist with a glut of small investor condos in one Toronto submarket and a tight family-home market in the Okanagan. Need is measured in households. Prices are set by buyers with approvals.
And the bound itself can move. "Up to" means CMHC is publishing the high end of a range, which tells you the inputs (population growth, household formation, what counts as enough supply) are soft enough that the agency won't commit to a single figure. If the ceiling can drop from one report to the next, any investment thesis resting on the top number inherits that wobble.
The strongest objection, answered
Here's the best case against me. Even if the true need lands well below 4.69 million, starts at 259,028 a year leave a gap so large that the direction is obvious, and CMHC itself warned construction may slow, which makes the gap worse.
That's fair, and I'd concede the national direction. The trouble is that nobody buys the national direction. You buy one unit, in one city, financed at one rate, and 2026 has stacked up dated events that affect that rate and that builder's costs: the CUSMA joint review on July 1, 2026, and the Bank of Canada's current monetary policy framework expiring December 31, 2026. A slowdown in construction can just as easily come from buyers stepping back, and that same pullback can soften prices on new product well before it tightens anything.
What actually changes in your math
Treat the 4.69 million as the most CMHC thinks the country needs. Then underwrite as if your property gets none of it.
For investors: price a purchase on local rent you can document today and carrying costs at a rate above the current one. If the deal only works because "Canada is short," it doesn't work.
For homeowners: your equity rests on buyers in your neighbourhood who can qualify, and the CMHC insurance cap of $1.5 million (effective since December 2024) still defines who can enter with a small down payment.
For everyone: a backstop that shrinks or holds depending on the next report gives you no protection to lean on.
The shortage is real. It just stopped being a guarantee once CMHC itself added "up to."
Two words now sit in front of CMHC's 2036 housing projection: "up to." As of September 10, 2026, CMHC puts Canada's need at up to 4.69 million new homes by 2036, and in the same breath it flagged that construction could slow. That qualifier turns the number from a target into an upper bound, and most of the people quoting it haven't adjusted.
The floor everyone has been selling
The common belief goes like this: Canada is millions of homes short, the gap can't close for a decade, so prices and rents have a permanent backstop. You hear it at presale launches, in investor seminars, and from homeowners in Kelowna and the GTA explaining why they'll never sell. The national shortage figure gets treated as a floor under every property in the country.
It was a reasonable story while the number read like a fixed debt. Back in 2023, CMHC estimated Canada needed 3.5 million additional units by 2030, and that figure got repeated as if it were a bill coming due. Now the horizon is 2036, the number is 4.69 million, and the wording says the real figure could be lower. A floor can't come with "up to" attached.
Where the shortage logic breaks
Run the arithmetic. Spreading 4.69 million homes across roughly ten years means something like 469,000 a year. CMHC counted 259,028 housing starts across Canada in 2025, with Ontario contributing 62,561 of them. At the top of the range, the country is building a little over half of what the projection calls for.
So the bulls win? Not quite. A need estimate describes how many homes it would take to reach a certain outcome. It says nothing about who can carry a mortgage at the asking price, or whether a renter in a particular building can absorb another increase. A national gap of millions can coexist with a glut of small investor condos in one Toronto submarket and a tight family-home market in the Okanagan. Need is measured in households. Prices are set by buyers with approvals.
And the bound itself can move. "Up to" means CMHC is publishing the high end of a range, which tells you the inputs (population growth, household formation, what counts as enough supply) are soft enough that the agency won't commit to a single figure. If the ceiling can drop from one report to the next, any investment thesis resting on the top number inherits that wobble.
The strongest objection, answered
Here's the best case against me. Even if the true need lands well below 4.69 million, starts at 259,028 a year leave a gap so large that the direction is obvious, and CMHC itself warned construction may slow, which makes the gap worse.
That's fair, and I'd concede the national direction. The trouble is that nobody buys the national direction. You buy one unit, in one city, financed at one rate, and 2026 has stacked up dated events that affect that rate and that builder's costs: the CUSMA joint review on July 1, 2026, and the Bank of Canada's current monetary policy framework expiring December 31, 2026. A slowdown in construction can just as easily come from buyers stepping back, and that same pullback can soften prices on new product well before it tightens anything.
What actually changes in your math
Treat the 4.69 million as the most CMHC thinks the country needs. Then underwrite as if your property gets none of it.
The shortage is real. It just stopped being a guarantee once CMHC itself added "up to."
Sources
Read Next
Restructuring Your Mortgage Is Not a Rescue Mission
What a Mortgage Rate Discount Is Designed to Keep You From Noticing
Bank of Canada Cuts Won't Save Your Fixed Rate From U.S. Inflation
7 Debt Ratio Checks That Set Your Mortgage Ceiling Before Your Credit Score Matters