Editorial note: this profile uses Granite filings through the third quarter of 2026 (September 30, 2026) and the November 4, 2026 results release and investor presentation. The price used is the November 20, 2026 close. Figures are dated where they appear. The valuation sections are a November 2026 assessment and are not updated for later events.
Granite REIT was spun out of Magna International, the Ontario auto parts giant, in 2003 to hold the real estate Magna’s plants occupied. At September 30, 2026 it owned 140 income-producing properties, plus 7 under development, across Canada, the United States, Austria, Germany, the Netherlands and the United Kingdom, with 61.9 million square feet of leasable area. Magna still accounts for about a quarter of its rent. Net leverage is 32%, and the distribution consumes roughly 60% to 67% of its adjusted cash flow.
On those facts Granite looks like one of the safer REITs in the country, and I think that reading is right. The harder questions are about price and growth. Cheap, mostly swapped debt is rolling over into a more expensive market, one tenant still matters a great deal, and the units have recovered much of their discount to the last published net asset value. My conclusion is that Granite is a high-quality industrial REIT with a well-covered distribution, but at CAD 88.50 a unit my probability-weighted ten-year return is about 7.1% a year, which does not clear the 9.5% hurdle I apply to a new purchase. The price at which it does clear is near CAD 73.40.
Four notes before the numbers. Granite reports in Canadian dollars, so every figure is CAD unless I say otherwise. Granite announced a three-for-one unit split on November 4, 2026, with a record date of November 18 and a payment date of November 23. Trading on a split-adjusted basis begins November 24, so at the time of writing every quoted price is still a pre-split price. I therefore keep every per-unit figure, the price, the distribution, FFO, AFFO and the net asset value, on the pre-split basis, and I give post-split equivalents (divide by three) alongside. I never combine a post-split price with a pre-split earnings figure. The reference price is the November 20 close of CAD 88.50, which is CAD 29.50 after the split, and it comes from a single market-data source, so refresh it before acting. This is my analysis, not a recommendation. For the broader framework on how REIT distributions, AFFO and taxes work, start with my introduction to Canadian REITs.
What Granite Owns
Granite’s portfolio at September 30, 2026 was 140 income-producing properties and 7 development properties, with investment properties carried at CAD 9.73 billion, according to its third quarter 2026 report and property statistics. The table shows the latest geographic split of fair value next to the year-end 2025 figures and the year-end capitalization rates, the most recent rates by country that I could verify.
| Country | Fair value, Sep 30, 2026 (CAD millions) | Share, Sep 30, 2026 | Fair value, Dec 31, 2025 (CAD millions) | Capitalization rate, Dec 31, 2025 |
|---|---|---|---|---|
| United States | 5,160.5 | 53.0% | 4,951.0 | 5.37% |
| Canada | 2,026.0 | 20.8% | 2,127.0 | 4.73% |
| Austria | 883.0 | 9.1% | 886.1 | 8.54% |
| Netherlands | 787.4 | 8.1% | 820.0 | 5.33% |
| Germany | 646.5 | 6.6% | 658.4 | 6.00% |
| United Kingdom | 36.9 | 0.4% | 36.0 | 8.44% |
| Development properties | 193.5 | 2.0% | included above | not applicable |
| Total | 9,733.8 | 100.0% | 9,478.4 | 5.6% overall |
Source: Granite Q3 2026 property statistics and 2025 annual report; shares are my calculation. The two dates are not like for like: the September 2026 country figures exclude development properties, which are shown separately, while the December 2025 country figures include them. Capitalization rates are for income-producing properties; the overall rate was 5.7% at September 30, 2026. Granite has since bought UK and US properties, so the current mix differs slightly.
Three things stand out. About 79% of the property value sits outside Canada, and 83% of 2025 revenue did (the United States alone was 55%), so Granite is as much a US industrial landlord as a Canadian one. The 8.54% Austrian capitalization rate is far above everything else, and I infer it reflects the Magna special-purpose plants there, which the market values more like operating facilities than generic warehouses. Granite does not state that link in the documents I read, so treat it as my inference. And Canada carries the lowest capitalization rate, which means the highest-priced real estate is the home market.
The structure has also changed. Granite spent 2013 to 2024 as a stapled-unit structure, each unit being a REIT trust unit paired with a share of a corporation. It unwound that structure on October 1, 2024 and now has a conventional REIT trust unit, which makes it simpler to compare with other Canadian REITs.
The Third Quarter and the 2026 Outlook
Granite’s third quarter was strong. Revenue rose to CAD 166.6 million from CAD 153.0 million a year earlier, and net operating income (NOI), the rent left after property costs, rose to CAD 136.0 million from CAD 127.1 million. Funds from operations (FFO) were CAD 1.60 per unit and adjusted funds from operations (AFFO) were CAD 1.45 per unit, up 8.1% and 15.1% from the prior-year quarter. The figures below come from Granite’s third quarter release, also carried by BusinessWire.
| Measure | Q3 2026 | Q3 2025 |
|---|---|---|
| Revenue (CAD millions) | 166.6 | 153.0 |
| NOI (CAD millions) | 136.0 | 127.1 |
| Same-property NOI growth, cash basis, constant currency | 6.9% | 5.2% |
| FFO per unit | 1.60 | 1.48 |
| AFFO per unit | 1.45 | 1.26 |
| AFFO payout ratio | 61% | 67% |
Source: Granite Q3 2026 release. Per-unit figures are diluted and before the unit split. Net leverage was 32% at September 30, 2026 against 35% at December 31, 2025. Same-property NOI growth was 7.8% for the first nine months.
Management’s 2026 outlook, tightened in the same release, calls for FFO of CAD 6.32 to CAD 6.37 per unit (previously CAD 6.30 to CAD 6.40) and AFFO of CAD 5.50 to CAD 5.55 per unit (previously CAD 5.45 to CAD 5.55). That implies FFO growth of about 7% on 2025’s CAD 5.91 and AFFO growth of about 6% on CAD 5.21. The release describes this as maintaining guidance, although the ranges moved, so I read it as a narrowing around the same midpoints. Same-property NOI growth is guided at 6.0% to 6.5% for the year, and the company assumes it will refinance CAD 300 million of debt maturing on December 11, 2026 with an unsecured term loan at a rate slightly below 4.0%. Over a longer view, the investor presentation shows FFO per unit growing at about 10.0% a year and AFFO per unit at about 8.5% a year from 2021 to the 2026 midpoint, which is the record the next ten years have to be measured against.
Not everything in the quarter was good. In-place occupancy fell to 96.7% at September 30 from 98.0% at year-end, committed occupancy (which counts signed leases not yet started) was 97.0% as of November 4, and the weighted average lease term shortened to 5.0 years from 5.5. Granite also reported net fair value losses on investment properties of CAD 16.4 million in the quarter, driven by wider discount and terminal capitalization rates in Europe, lower Canadian market rents and higher US vacancy, partly offset by a gain on a property held for sale. None of these is alarming. Together they say the 6% to 7% growth rate is rent catching up to market on expiring leases and not a permanent state.
Leasing spreads show where the growth is coming from. Granite signed about 1.6 million square feet of new leases and renewals in the quarter at rents 35% above expiring rents, and the investor presentation says about 70% of 2026 expiries had been renewed at roughly 22% higher rents. Spreads are lumpy, and a handful of large leases can swing a quarter, so I treat 22% to 35% as evidence that rents are still below market on average and not as a run rate.
Three items in the release do not reconcile on their face, and I report them as published and do not try to bridge them. The Q3 report shows 140 income-producing properties and 61.9 million square feet at September 30, 2026, while the release boilerplate cites 149 investment properties and about 62.4 million square feet without a date. Granite closed two UK acquisitions in October and sold the Plainfield, Indiana property on October 1, but I could not demonstrate how the two property counts connect and I do not claim to. The net fair value loss on investment properties is CAD 16.4 million for the third quarter and CAD 22.0 million for nine months in the report, while the highlights cite CAD 21.5 million of losses on investment properties, partly offset by a gain on a held-for-sale asset that is not quantified, and the release does not say how the figures relate. And the CAD 270.5 million acquisition headline is made up of two Houston purchases in August of CAD 104.1 million and CAD 33.3 million and two UK purchases of CAD 80.4 million and CAD 52.7 million, with the release narrative giving a slightly different Houston figure that I cannot explain. None of this changes my view, and where figures differ I use the reported tables.
Is the Distribution Safe?
On the numbers, yes. Granite’s AFFO payout ratio was 61% in the third quarter and 64% for the first nine months, and its FFO payout ratio was 55% and 56%. In November the board raised the targeted annual distribution 4.23% to CAD 3.70 per unit (CAD 0.3083 a month), effective with the December 2026 distribution, up from CAD 3.55. After the split that is CAD 1.2333 a unit a year. Against the midpoint of the AFFO outlook, CAD 5.525, the new distribution is a 67.0% payout, and against the FFO midpoint it is 58.3%. For 2025 the AFFO payout was 65% and the FFO payout 57%. Granite’s 2025 annual report describes the increase commencing in 2026 as the fifteenth consecutive year of distribution increases, a record I take from the issuer and did not rebuild from the full distribution history.
A payout ratio is only as good as the cash flow measure under it. Both FFO and AFFO are non-standard measures, and Granite’s AFFO deducts maintenance capital spending, tenant allowances and leasing commissions, which the company guides at CAD 40.0 million for 2026. That is about CAD 0.64 a unit. The outlook gap between FFO and AFFO is about CAD 0.82 a unit, so the rest comes from other adjustments I have not itemized. The conservative reading is that the AFFO payout is the one to watch, and at 67% it leaves about CAD 1.8 a unit, roughly a third of AFFO, to fund development, debt reduction and buybacks.
I stress-tested that coverage against the refinancing numbers later in this article. If AFFO fell 10% from the outlook midpoint and nothing else changed, the payout would be about 74%. If the full run-rate refinancing cost at 4.5% landed on nominal coupons with no offsetting growth, AFFO would fall to about CAD 5.06 and the payout would be about 73%. In the harshest version, an effective cost of 4.5% after swaps, AFFO would fall to about CAD 4.67 and the payout would be about 79%, still covered. None of these is a forecast, because rent growth and acquisitions work the other way, but they show the cushion.
Granite has also used the units themselves. In 2025 it repurchased about 2.2 million units for CAD 144.6 million at an average CAD 67.01. In 2026 it issued units through its at-the-market program, 709,975 in the second quarter at an average CAD 96.39 and 723,200 in the third quarter at an average CAD 96.82, about 1.4 million units and CAD 138.5 million of gross proceeds over nine months, and a new normal course issuer bid began on May 26, 2026 with no repurchases reported in the first half. Issuing at prices well above today’s CAD 88.50 was good timing, but about 2.4% more units dilutes growth in per-unit AFFO, which is one reason my growth assumptions sit below the company’s recent pace.
A low payout is why this looks like one of the safer distributions in Canadian real estate, and Granite has raised its targeted distribution in each year I checked, including the 4.23% increase announced in November. But a distribution is a board decision, not a promise, and the real risk to it is not coverage today. It is whether growth in cash flow keeps pace with the cost of the debt.
The Magna Question
Magna International was 26.0% of Granite’s annualized revenue, about CAD 141.7 million, and 19% of its leasable area at September 30, 2026, according to Granite’s property statistics. The November investor presentation shows 25.7% of revenue and 19.1% of area on its own date and basis, and the year-end 2025 annual report showed 26.3%. Magna is the tenant in 26 leases covering 11.9 million square feet, with a fair value of CAD 2.07 billion, or 21.3% of the portfolio, and Granite counted 27 Magna properties at year-end 2025. The top ten tenants together were 44.3% of annualized revenue at year-end 2025, and the next two largest after Magna were Amazon at 3.6% and Wayfair at 2.7%. Concentration this high is rare among REITs. Magna’s share of reported 2025 revenue was about 23%, which differs from 26% only because the two measures are built differently.
It is also better than it was. Magna was 36% of revenue at the end of 2020 and 29% at the end of 2021, according to Granite’s earlier filings, and the company’s stated strategy is to reduce Magna exposure by growing with other tenants. Granite’s own words are that Magna may or may not renew leases as it optimizes its manufacturing footprint, and it lists tariffs among the risks.
The property mix matters as much as the percentage. Granite classifies seven properties as special purpose, with a fair value of about CAD 1.26 billion (12.9% of the portfolio), about 8.9 million square feet and annualized revenue of CAD 96.3 million (17.7%). They sit in Canada (two properties, CAD 402.9 million), Austria (four, CAD 775.2 million) and Germany (one, CAD 77.3 million), and six of the seven, in Canada and Austria, are occupied exclusively by Magna. Magna has invested heavily in them, which supports renewal, but they are harder to re-let and less liquid than a generic warehouse if it does not. The market has noticed: the special-purpose capitalization rate was 7.47% at June 30, 2026 against 7.34% at year-end 2025. By my arithmetic the special-purpose revenue is at most about two thirds of the CAD 141.7 million Magna total, and the German property is not a Magna one, so the true figure is lower. That leaves at least a third of Magna’s rent in conventional buildings that a new tenant could occupy. This split is my calculation, not a Granite disclosure.
The largest single block is Graz, Austria, two special-purpose properties of about 5.0 million square feet. Granite extended those leases for ten years to January 31, 2034 in an agreement announced in 2023. In the first five years of the extension, annual rent resets on February 1, 2024 by the change in the consumer price index over the previous 60 months, capped at 10%, and Granite said the Graz facilities were about 8% of annualized revenue in January 2023. That release does not name the tenant, so the link to Magna is my editorial inference and not a confirmed disclosure. It rests on Granite’s earlier releases about the Graz extensions, which named Magna Steyr as the tenant of those facilities, and on the fact that six of the seven special-purpose properties are Magna-exclusive. I did not find a current Granite document that names the tenant of the Graz leases, and the release does not describe renewal options, termination rights or the escalator for the second five years. I also have no Magna-specific lease expiries beyond the weighted average remaining term quoted below.
Beyond Graz, Granite’s Magna leases generally pass occupancy costs to the tenant, escalate in fixed steps or with inflation, and carry renewal options at market rent or an inflation-based rent. Magna has a right of first refusal on sales of its properties, and some leases bar a sale to a Magna competitor without consent. Magna’s weighted average remaining lease term was 5.4 years in the November presentation, down from 5.7 years at year-end 2025 and 5.6 at June 30, 2026, which is longer than the portfolio’s 5.0 years.
I cannot give you a Magna-only lease expiry schedule because Granite does not publish one in the documents I could read. What it publishes is a portfolio-wide schedule, and I show it here as such. It is not a Magna schedule, though it tells you when the rent base as a whole comes up for renewal.
| Expiry year | Annualized revenue expiring (CAD millions) | Share of total |
|---|---|---|
| 2026 | 14.2 | 2.6% |
| 2027 | 45.0 | 8.4% |
| 2028 | 83.4 | 15.5% |
| 2029 | 76.0 | 14.1% |
| 2030 | 75.1 | 14.0% |
| 2031 | 21.5 | 4.0% |
| 2032 and later | 222.7 | 41.4% |
| Total | 537.9 | 100.0% |
Source: Granite Q2 2026 report, as at June 30, 2026, all tenants. 169 leases; weighted average lease term 5.1 years at that date. This is the portfolio-wide schedule, not a Magna schedule.
Magna’s credit is strong but not a guarantee. It is rated A3 by Moody’s, A- by S&P and A (low) by Morningstar DBRS, and the Moody’s and S&P outlooks were negative as of the filings I read, with DBRS stable. Granite’s properties are generally leased to Magna operating subsidiaries, and Magna International Inc. is not a guarantor of those leases, so the parent’s rating describes the group and not necessarily the subsidiary that signs the lease. An A rating means a low probability of default, not zero, and a negative outlook means the rating could fall. The fair reading is balanced. Magna is Granite’s biggest strength, because it supplies long, inflation-linked rent from a strong credit, and its biggest single point of dependency, because the special-purpose plants have few other users. The near-term risk is low given Graz runs to 2034, but the risk is concentrated in a small number of buildings, and it is the reason the market applies a higher capitalization rate to them.
The Debt: Cheap Money Rolling Into a Dearer Market
Granite’s balance sheet is conservative. Net leverage was 32% at September 30, 2026, down from 35% at year-end, and its indebtedness ratio, which is based on total debt, was 6.5 times against 7.3 times. The investor presentation shows debt to adjusted EBITDA of 6.3 times against a long-term target of 30% to 35% net leverage and 6.5 to 7.0 times. Interest coverage was 5.0 times for 2025, and Morningstar DBRS rates the REIT BBB (high), having had it on a positive trend in March 2026. Net debt was CAD 3,089 million, total debt CAD 3,192 million and cash CAD 102 million at September 30. The CAD 1.0 billion credit facility was amended on January 22, 2026 to mature on January 22, 2031, with a one-year extension option, and CAD 29 million was drawn at quarter end.
The reported weighted average cost of debt was 2.62% at September 30, 2026 (2.72% at year-end 2025), and that figure is the point to examine. It is low because most of the debentures are denominated in Canadian dollars but swapped into euros or US dollars through cross-currency swaps that hedge Granite’s foreign assets, so Granite pays foreign-currency interest at the swap rate and not the Canadian coupon. The weighted average debt term has also shortened, from 3.4 years at year-end to 2.7 years at September 30.
| Debenture maturity | Principal (CAD millions) | Nominal coupon | Swapped foreign-currency rate (reported) |
|---|---|---|---|
| Dec 11, 2026 | 300.0 | CORRA + 0.77% (floating) | EUR 205.5 million at 1.355% |
| Jun 4, 2027 | 500.0 | 3.062% | USD 370.3 million at 2.964% |
| Aug 30, 2028 | 500.0 | 2.194% | USD 119.1 million at 2.096% and EUR 242.1 million at 0.536% |
| Apr 12, 2029 | 400.0 | 6.074% | EUR 277.7 million at 4.958% |
| Oct 4, 2029 | 250.0 | 3.999% | EUR 167.4 million at 3.494% |
| Dec 18, 2030 | 500.0 | 2.378% | EUR 319.4 million at 1.045% |
| Oct 4, 2031 | 550.0 | 4.348% | none listed |
| Total | 3,000.0 |
Source: Granite Q3 2026 report (swap notional amounts and rates at September 30, 2026) and annual report (principal and coupons). The December 2026 maturity is called debentures in the release and a term loan in parts of the Q3 report. Net swap fair value was a liability of CAD 135.2 million at September 30 (CAD 110.8 million at year-end). Principals sum to CAD 3,000 million.
Weighting the nominal coupons gives about 3.5% (assuming CORRA near 2.5% for the floating tranche, my assumption), against about 2.7% when I use the swapped rates, which is within about 0.1 percentage points of Granite’s reported 2.62%. That gap is the heart of the refinancing question, because the answer depends on which cost you measure. I tested four views: the nominal coupon change, the effective change after swaps, the full run-rate impact once every tranche has refinanced, and the phased impact through 2031 as each maturity comes due. Each tranche is assumed to be replaced at a single all-in rate of 4.0%, 4.5% or 5.0%, and management’s own assumption for the December 2026 refinancing, slightly below 4.0%, anchors the low end.
| Replacement rate | Nominal basis, full run-rate (CAD millions) | Nominal, share of 2026 AFFO | Effective basis, full run-rate (CAD millions) | Effective, share of 2026 AFFO |
|---|---|---|---|---|
| 4.0% | +13.8 | 4.0% | +38.4 | 11.2% |
| 4.5% | +28.8 | 8.4% | +53.4 | 15.5% |
| 5.0% | +43.8 | 12.7% | +68.4 | 19.9% |
Source: my calculation using 62.3 million diluted units and the AFFO midpoint of 5.525. All seven tranches are included, with the floating December 2026 tranche at an assumed CORRA of about 2.5% plus 0.77%. The effective basis uses my reconstruction of each tranche’s swapped rate. Replacement rates are assumptions, not forecasts.
None of this hits AFFO at once. The maturities are spread from December 2026 to October 2031, and the cost builds as each tranche rolls.
| Replacement rate and basis | 2027 | 2029 | 2031 |
|---|---|---|---|
| 4.0%, nominal | +4.9 | +9.9 | +15.3 |
| 4.5%, nominal | +7.8 | +18.2 | +28.2 |
| 5.0%, nominal | +10.8 | +26.4 | +41.1 |
| 4.0%, effective | +10.9 | +25.6 | +39.8 |
| 4.5%, effective | +13.9 | +33.9 | +52.8 |
| 5.0%, effective | +16.8 | +42.1 | +65.7 |
Source: my calculation. Incremental annual interest in the calendar year, in CAD millions, against the current cost, using maturity dates (the December 2026 maturity counts as a full year in 2027). These are effects on AFFO before any offsetting rent growth.
The right way to read this is as a drag against growth, not as an immediate hit, and dividing the cumulative cost by five, as I did in an earlier draft, understates how it compounds. A 4.5% replacement on nominal coupons costs about 8% of 2026 AFFO once every tranche has rolled, and on an effective after-swap basis the same 4.5% costs about 15%. To see what that does to growth I ran a simplified illustration in which underlying AFFO per unit grows 5% a year before refinancing, each year’s incremental interest from the table above is deducted per diluted unit, and the five-year compound growth to 2031 is compared with the unburdened path. A 4.5% replacement on nominal coupons lowers 2031 AFFO by about 6.4% and compound growth from 5.0% to about 3.6%, roughly 1.4 points a year. On an effective basis it lowers 2031 AFFO by about 12.0% and compound growth to about 2.4%, roughly 2.6 points a year, which I treat as a stress case and not a forecast, because Granite can swap new borrowing into euros or US dollars at whatever those rates are when it refinances, and I have not verified them. At 4.0% and 5.0% the nominal drag is about 0.7 and 2.0 points a year and the effective drag roughly 2.0 and 3.3. The drag also lands unevenly: in the 4.5% cases, annual growth ranges from about 1% to 5.4% as individual maturities fall. This is an illustration of interest cost alone. It excludes rent growth, acquisitions, dilution from new units and any hedging decision, and it is not a forecast. My scenario growth rates are net assumptions that already allow for refinancing, so I do not subtract the drag from them a second time. Read the other way, a base case of 3.5% net growth is consistent with underlying growth before refinancing of about 4.9% if replacement debt costs 4.5% on nominal coupons, and about 6.0% on the effective basis, against the 6% guided for 2026. Each 100 basis points on a CAD 500 million tranche costs CAD 5.0 million a year, about CAD 0.08 a unit, or 1.5% of AFFO.
For context, Dream Industrial REIT said in its second quarter that its weighted average face interest rate was 3.32%, and its management indicated new five-year Canadian debt near 4.40% to 4.50%, according to a secondary transcript. That is a reasonable marker for a 4.5% scenario. The refinancing explains why growth has to come from rent increases and acquisitions and not from the balance sheet, and why I model AFFO growth of 3.5% in my base case, well below the 6% guided for 2026.
What Is It Worth?
Granite’s year-end 2025 net asset value (NAV) was CAD 103.43 per unit, which it reported in its 2025 annual report, up about 1% from a year earlier. NAV is a non-IFRS measure, and I could not find Granite’s definition of it or a reconciliation to its financial statements in the documents I read. The interim reports do not update it, so it is now almost eleven months old. IFRS unitholders’ equity per unit, which I calculated from reported equity and unit counts, was CAD 90.91 at December 31, 2025, CAD 94.13 at June 30, 2026 and CAD 94.56 at September 30, 2026. The two measures differ by CAD 12.52 at year-end 2025 and I could not reconcile them. One candidate for part of the gap is deferred income tax, because IFRS equity is struck after net deferred tax liabilities of about CAD 600 million, or about CAD 9.91 a unit, but that is smaller than the gap and I have not established that Granite’s NAV adjusts for it. I therefore treat the two as separate benchmarks and do not present an adjusted or current NAV of my own.
| Measure (pre-split) | Per unit | Post-split equivalent | Price of CAD 88.50 as a share of it | Discount |
|---|---|---|---|---|
| Issuer-reported NAV, Dec 31, 2025 (stale) | 103.43 | 34.48 | 85.6% | 14.4% |
| IFRS equity per unit, Dec 31, 2025 | 90.91 | 30.30 | 97.4% | 2.6% |
| IFRS equity per unit, Sep 30, 2026 | 94.56 | 31.52 | 93.6% | 6.4% |
Source: Granite 2025 annual report and Q3 2026 report; equity per unit is my calculation from unitholders’ equity of CAD 5,509.0 million on 60.601 million units (2025) and CAD 5,870.6 million on 62.084 million units (September 2026). The reported NAV exceeds December 2025 IFRS equity per unit by CAD 12.52, or 13.8%, and I could not establish why.
The two discounts are distinct benchmarks and not a range. Against the stale issuer NAV of CAD 103.43 the price is 14.4% lower, and against September IFRS equity per unit it is 6.4% lower. I cannot say which is the better measure of Granite’s current asset value, because the issuer figure has no definition I could find and is almost eleven months old, while the IFRS figure is a book measure. A discount to NAV is not a guaranteed margin of safety. NAV is an appraisal, appraisals move, and Granite recorded net fair value gains of CAD 7.3 million in the first quarter of 2026 and losses of CAD 12.9 million in the second and CAD 16.4 million in the third, CAD 22.0 million net over nine months or about CAD 0.35 a unit. Granite’s overall capitalization rate rose from 5.6% at year-end 2025 to 5.7% in September 2026. A 25 basis point rise from 5.7% would cut investment property values by about CAD 409 million, or CAD 6.59 a unit, and a 5% fall in value would remove about CAD 7.84 a unit, which is 7.6% of the stale NAV and 8.3% of September equity per unit.
| Measure at CAD 88.50 (pre-split) | Value | Post-split equivalent |
|---|---|---|
| Market capitalization (62.08 million units) | CAD 5.49 billion | same |
| Targeted distribution yield (CAD 3.70) | 4.18% | CAD 1.2333 per unit |
| P/AFFO on 2026 outlook midpoint (5.525) | 16.0x | CAD 1.84 AFFO per unit |
| P/FFO on 2026 outlook midpoint (6.345) | 13.9x | CAD 2.12 FFO per unit |
| AFFO yield | 6.24% | |
| NOI yield on enterprise value (Q3 NOI x 4 / EV) | about 6.3% | |
| Reference price | CAD 88.50 | CAD 29.50 |
Source: Granite releases; price is the November 20, 2026 close from a single source; my calculations. Enterprise value is market capitalization plus net debt of CAD 3,089 million, about CAD 8.58 billion, and is approximate. Units outstanding were 62,083,989 at November 3, 2026 (186,251,967 after the split).
The last line is worth a moment. Granite’s own overall capitalization rate on its properties is 5.7%, while the market values the whole enterprise at roughly a 6.3% NOI yield. Capitalization rate and NOI yield on enterprise value are not identical, since enterprise value includes development land and other items, so I take the gap as a rough signal that the market is paying less for these buildings than Granite’s appraisals do, and not as a precise measure. For reference, the same CAD 88.50 is about 8.3% above the December 31, 2025 close of CAD 81.71, and the multiple of 16.0 times 2026 AFFO is above the 14.8 times that the year-end price implied on the same outlook midpoint.
How Granite Compares With Dream Industrial
Dream Industrial REIT is the closest Canadian listed comparison, with a larger and more Canadian and European portfolio and no US exposure, and a comparison helps calibrate the numbers above. The two do not report on identical bases, so the table is directional, and the dates differ: Granite is at September 30 and Dream’s official figures are at June 30, 2026.
| Measure | Granite REIT | Dream Industrial REIT |
|---|---|---|
| Geography | US 53%, Canada 21%, Europe 24% (fair value) | Canada about 57%, Europe about 42%, no US (NOI) |
| Portfolio | 140 income-producing properties, 61.9 million sq ft | 565 buildings, 75.7 million sq ft |
| Largest tenant | Magna, 26% of annualized revenue | not disclosed in the release |
| In-place occupancy | 96.7% | 94.2% (Canada 96.0%, Europe 92.5%) |
| Payout | AFFO 61% (Q3); FFO 55% | FFO 63.1% (Q2); AFFO payout not reported |
| Leverage | net debt 32% of assets; 6.3x to 6.5x | net debt 35.8% of total assets; 6.6x |
| Organic growth | same-property NOI 6.9% (Q3), cash basis | comparative-property NOI 10.3% (Q2) |
| Valuation | 14.4% below the stale issuer NAV and 6.4% below IFRS equity per unit (distinct measures) | unit price about 83% of NAV (CAD 13.99 against 16.76) |
| Cost of debt | 2.62% weighted average, after swaps | 3.32% weighted average face rate |
| Credit rating | BBB (high) | BBB (high) |
Source: Granite Q3 2026 report and release; Dream Industrial Q2 2026 release (as at June 30, 2026). Dream’s Canadian and European shares are of wholly-owned comparative-property NOI in the quarter (Ontario about 29%, Quebec 16%, Western Canada 12%, Europe 42%). Dream’s net debt multiple is to normalized adjusted EBITDA at fair value; a secondary third-quarter transcript cited 7.3 times on a different basis. Valuation uses the unit price and NAV in the release.
Two readings follow. Granite has the stronger operating numbers on occupancy, same-property growth and leverage, and a much lower reported cost of debt, while Dream is the more diversified landlord by tenant and has no single tenant of Magna’s weight. And the definitions prevent a clean comparison: Dream’s FFO per unit, EBITDA and comparative-property NOI are built differently from Granite’s, and the discount to NAV for each rests on a different appraisal base. I would not pay a premium for Granite’s numbers without noticing that its US weighting and its Magna exposure are the things that differ most.
What Return Should You Expect?
I modelled ten years from CAD 88.50 with three scenarios. In each, AFFO per unit grows from the 2026 midpoint at a steady rate, the distribution grows at the same rate from CAD 3.70, and the units sell at the end of year ten at a stated multiple of that year’s AFFO. Cash flows are treated as arriving at the end of each year, with the first year including one month at the old CAD 3.55 rate. That treatment follows the issuer’s confirmed schedule: Granite’s November 17, 2026 announcement declared the November distribution at CAD 0.2958 a unit before the split (CAD 0.0986 after it), payable on December 15, 2026 to unitholders of record on November 30, 2026, so a buyer on November 20, 2026 is entitled to it at the old rate, and the increase to CAD 0.3083 a month starts with the December 2026 distribution. The first twelve months therefore hold one old-rate and eleven new-rate payments, CAD 3.6875 in total, and using twelve new-rate months instead would change the return by less than one basis point. Timing the cash flows monthly instead of at year end lifts the expected return to about 7.2% and the 9.5% hurdle price to about CAD 74.5, which does not change the conclusion. The expected return is the internal rate of return of the probability-weighted cash flows, not an average of the three scenario returns, and it is a nominal pre-tax Canadian-dollar figure. I also keep a five-year version as a supplement. The growth rates, multiples and the 25/50/25 weights are my assumptions and judgement.
| Scenario | Weight | AFFO growth | Exit P/AFFO | Year-10 AFFO | Exit price | Cumulative distributions | Annual return |
|---|---|---|---|---|---|---|---|
| Bear | 25% | 1.0% | 12.5x | 6.10 | 76.29 | 38.67 | 3.2% |
| Base | 50% | 3.5% | 15.0x | 7.79 | 116.90 | 43.28 | 7.1% |
| Bull | 25% | 5.5% | 16.5x | 9.44 | 155.72 | 47.44 | 9.9% |
| Probability-weighted | 116.45 | 43.17 | 7.1% |
Source: my model, per unit before the split (divide by three for post-split equivalents: for example the expected terminal value is CAD 38.82 and expected cumulative distributions CAD 14.39). Returns include distributions and are annualized. The weighted return of 7.06% is the IRR of the probability-weighted cash flows; the simple average of the three scenario returns is 6.8%. Today’s multiple is 16.0x.
I changed some assumptions from my first draft and should explain why. Ten years of 7% growth, which I had in the bull case, would have AFFO doubling and is hard to justify even after Granite’s 8.5% a year since 2021, a period helped by acquisitions, rent catching up to market and low-cost debt that cannot repeat. I cut the bull case to 5.5% a year and a 16.5 times exit, a multiple only slightly above today’s 16.0 times. The base case is 3.5% a year, not 4%, because the refinancing drag, which my illustration puts at roughly 1.4 to 2.6 points a year of compound growth through 2031 at a 4.5% replacement cost taken on its own, the dilution from at-the-market issuance and fading leasing spreads all work against rent growth in the first half of the decade. The bear case has 1% growth, not zero, because contractual inflation-linked escalators make a decade of exactly zero growth in nominal rent an unlikely outcome, and it keeps the 12.5 times exit. Under the earlier, more generous inputs (0%, 4% and 7% growth with 12.5, 15 and 17 times exits) the expected ten-year return would be about 7.7%, so the conclusion does not hinge on my changes. Shifting the weights to 35/45/20 gives 6.6%, and to 20/45/35 gives 7.5%.
The entry price drives the answer more than any scenario input.
| Entry price (pre-split) | Post-split | Weighted annual return | Bear / Base / Bull |
|---|---|---|---|
| 65.00 | 21.67 | 11.1% | 7.2% / 11.2% / 14.1% |
| 70.00 | 23.33 | 10.1% | 6.2% / 10.2% / 13.1% |
| 75.00 | 25.00 | 9.2% | 5.3% / 9.3% / 12.1% |
| 80.00 | 26.67 | 8.4% | 4.4% / 8.4% / 11.3% |
| 84.00 | 28.00 | 7.7% | 3.8% / 7.8% / 10.6% |
| 88.50 | 29.50 | 7.1% | 3.2% / 7.1% / 9.9% |
| 92.00 | 30.67 | 6.6% | 2.7% / 6.6% / 9.4% |
| 95.00 | 31.67 | 6.2% | 2.3% / 6.2% / 9.0% |
| 100.00 | 33.33 | 5.5% | 1.7% / 5.6% / 8.4% |
Source: my model. Bands are not trading signals and should be recomputed when AFFO guidance or the price changes.
The sensitivity to growth and exit multiple, at CAD 88.50 and with the distribution growing with AFFO, shows what has to go right.
| AFFO growth | 12x exit | 13x | 14x | 15x | 16x | 17x |
|---|---|---|---|---|---|---|
| 1.0% | 2.8% | 3.5% | 4.1% | 4.6% | 5.2% | 5.7% |
| 2.0% | 3.8% | 4.4% | 5.1% | 5.6% | 6.2% | 6.7% |
| 3.5% | 5.3% | 5.9% | 6.5% | 7.1% | 7.7% | 8.2% |
| 5.0% | 6.7% | 7.4% | 8.0% | 8.6% | 9.2% | 9.7% |
| 6.0% | 7.7% | 8.4% | 9.0% | 9.6% | 10.2% | 10.7% |
| 7.0% | 8.7% | 9.3% | 10.0% | 10.6% | 11.1% | 11.7% |
Source: my model. Ten-year annualized return including distributions.
To earn 9.5% from CAD 88.50 with a 15 times exit, AFFO has to grow about 5.9% a year for ten years, and with a 16 times exit about 5.4%. That is roughly what management guides for 2026, sustained for a decade, and it is above what my bull case assumes. As a stress, a decade of AFFO declining 1% a year with a frozen CAD 3.70 distribution and an 11 times exit would return about 0.5% a year, with the payout reaching 74% by the end. Applied to CAD 10,000 with no reinvestment, 113 units would produce about CAD 12,990 in the bear case, CAD 18,100 in the base case and CAD 22,955 in the bull case, counting distributions and the final sale.
At about 7% a year, Granite is a fair price for a good business and not a bargain. The distribution alone supplies about 4.2%, so the rest has to come from growth, and a decade of 3.5% to 5.5% growth is a respectable outcome that does not reach 9.5% from today’s price.
What Price Makes Granite Attractive?
I use a fixed hurdle: 9.5% a year, nominal, pre-tax, in Canadian dollars. I do not lower it for quality, because quality is already in the cash flows and the scenario weights. The hurdle price is the price at which the probability-weighted cash flows return exactly 9.5%, which is the present value of those cash flows at 9.5%, and it comes to CAD 73.42, or CAD 24.47 after the split. The other band edges come from the same calculation at 11%, 8% and 7%.
| Band | Expected return at that price | Price range (pre-split) | Post-split |
|---|---|---|---|
| Compelling | 11% or more | at or below 65.70 | at or below 21.90 |
| Attractive entry | 9.5% to 11% | 65.70 to 73.40 | 21.90 to 24.47 |
| Fair value | 8% to 9.5% | 73.40 to 82.30 | 24.47 to 27.43 |
| Full | 7% to 8% | 82.30 to 88.95 | 27.43 to 29.65 |
| Expensive | below 7% | above 88.95 | above 29.65 |
Source: my model, edges rounded to CAD 0.05. At CAD 88.50 Granite sits in the Full band, just inside its upper edge. For a 12% return the entry price is about CAD 61.15 (CAD 20.38 after the split). The base case alone returns 9.5% at about CAD 73.65.
CAD 88.50 is about 20.5% above the hurdle price of CAD 73.42. Put differently, Granite would have to fall by about 17.0% for a new purchase to meet my hurdle, or the cash flows would have to be better than my base case. Even at the Full band edge of CAD 88.95, the expected return is 7%, and at the bottom of the Fair value band it is 8%.
How Granite’s Distribution Is Taxed
Granite is a Canadian REIT, and its distributions are not mainly dividends. Granite publishes how each year’s distribution was characterized on its income tax information page, and the mix is unusual. The per-unit amounts below are before the split, and divide by three for a post-split equivalent.
| Year | Distribution per unit (CAD) | Other income | Foreign non-business income | Actual dividends | Return of capital |
|---|---|---|---|---|---|
| 2025 | 3.412 | 16.17% | 78.75% | 0.07% | 5.01% |
| 2024 | 3.308 | 20.22% | 74.68% | 0.07% | 5.03% |
| 2023 | 3.209 | 30.81% | 64.08% | 0.10% | 5.01% |
| 2022 | 3.108 | 33.93% | 60.88% | 0.06% | 5.13% |
Source: Granite tax information page. No capital gains were allocated in these years. The page gives the characterization but not the foreign tax paid per unit; the yearly tax distribution notices may.
Most of the distribution is foreign non-business income, which fits a portfolio that is mostly outside Canada, and the share has risen from about 61% in 2022 to about 79% in 2025 as the other-income share has fallen. It is taxed at your full marginal rate, and none of it is a capital gain. The tiny dividend component, 0.05% to 0.10%, is described by Granite as actual dividends. Granite’s tax page does not say whether any of it is designated as eligible for the dividend tax credit, so I make no assumption, and at 0.07% it is about CAD 0.70 in every CAD 1,000 distributed, so the credit would not change the result either way. For practical purposes Granite’s distributions carry no meaningful dividend tax credit.
On the T3 slip, the form a trust uses to report income allocations to unitholders, foreign non-business income is reported in box 25, foreign non-business income tax paid in box 34 and return of capital in box 42, according to CRA’s T3 slip guidance. In a non-registered account the box 25 amount is included in your income, and the box 34 amount goes onto Form T2209 to calculate a federal foreign tax credit, which for each country is generally limited to the lesser of the foreign tax you paid and the Canadian tax on that income. That is different from foreign tax that Granite or its subsidiaries pay inside the REIT structure before cash is distributed, which is not itself a credit to you. Granite’s tax page gives the characterization but not a foreign tax amount per unit, so I cannot say whether box 34 will show anything on your slip. Check your own slip. Take a hypothetical CAD 1,000 of 2025-style distributions held in a non-registered account by an investor at an illustrative 40% marginal rate, ignoring the dividend slice and any foreign tax credit: CAD 787.50 would be foreign non-business income, CAD 161.70 other income and CAD 50.10 return of capital, and the tax on the first two would be about CAD 380, or 38% of the cash. The return of capital is not taxed when received but lowers your adjusted cost base, so it raises the gain when you sell. Any credit available to you would reduce that CAD 380 by an amount I cannot give you.
A registered account is therefore an attractive home, but not a complete answer. In a TFSA there is no Canadian tax on the distribution, and in an RRSP the tax is deferred, so the character of the income does not matter for Canadian tax. CRA’s foreign tax credit folio says income earned in a TFSA or RRSP is not counted for the foreign tax credit, and the plan, not the individual, is the holder of the units, so the personal T3 reporting and credit described above are for non-registered holders and are not available in the same way to registered ones. Foreign tax that Granite or its subsidiaries pay abroad reduces what is available to distribute whatever account holds the units, and I could not establish how much that is. Units listed on the TSX are generally qualified investments for registered plans, but confirm with your broker. My guides to dividend tax treatment, the TFSA and RRSP versus TFSA cover the mechanics. In a non-registered account, capital gains realized on a sale are subject to the capital gains rules in force at that time, as explained in my capital gains guide. This is general information and not personal tax advice.
Practical Notes for Canadian Buyers
Granite’s units trade on the TSX in Canadian dollars under GRT.UN. The unit split needs no action. Unitholders of record on November 18, 2026 receive two additional units for each unit held on the November 23 payment date, according to Granite’s split announcement. The TSX applies a due bill from the open on November 18 through the close on November 23, so units bought in that window carry the entitlement, and trading on a split-adjusted basis starts November 24. The ticker does not change, and Granite does not expect the split to create taxable income or a gain or loss. A cost base and a price quoted before the split should be divided by three afterward, and the at-the-market and buyback programs will be adjusted for the split.
Currency matters more here than for a domestic REIT. About 79% of Granite’s property value and 83% of its 2025 revenue are outside Canada, in US dollars, euros and pounds. Granite offsets part of the exposure by borrowing in Canadian dollars and swapping into foreign currency, which is the cross-currency swap position described above. A stronger Canadian dollar reduces the Canadian-dollar value of rent earned abroad, though the foreign-currency debt service shrinks with it, and those swaps carried a net liability of CAD 135.2 million at September 30.
Granite also appears in Canadian REIT ETFs. If you would rather own the sector than one name, my comparison of ETFs, mutual funds and individual stocks covers the trade-offs, and my piece on REITs versus direct real estate covers why a REIT is the default for most investors. For a much smaller, Quebec-focused REIT with a different mix of industrial, office and retail, see my BTB REIT profile.
What Could Go Wrong
The risks fall into three groups. Cyclical risks hurt earnings without destroying capital: industrial rents stop rising, vacancy builds as new supply arrives, or refinancing costs more than the 4.5% I tested. In-place occupancy has already slipped from 98.0% to 96.7%. A fair value decline hurts NAV, though not cash flow, unless it triggers covenant or credit rating trouble, and Granite’s 32% net leverage leaves a wide cushion.
Concentration risks are specific to Granite. Magna is still a quarter of the rent, a non-renewal at a special-purpose plant would be hard to backfill, the parent does not guarantee the subsidiary leases, and the negative rating outlooks on Magna are worth tracking. Currency is the other: roughly 79% of the property value is outside Canada, and a rise in the Canadian dollar reduces reported results. US exposure of 53% also ties Granite to the US industrial market and the US dollar.
The permanent risks are the least likely: a large acquisition at the wrong price, a sustained fall in European values that forces a re-pricing of the special-purpose assets (the capitalization rate on those assets has already widened by about 0.13 percentage points this year), or equity issuance below NAV to fund growth. The AFFO payout of 67% makes a distribution cut very unlikely without a severe shock, and I found no cut in the record I checked.
My Verdict
Granite is a high-quality industrial REIT. It has a conservative balance sheet, with 32% net leverage and a BBB (high) rating, occupancy of 96.7%, same-property NOI growth near 7%, a lower cost of debt than its peer, and a tenant base that is more diversified than it was. Its main weakness is concentration, with Magna still near 26% of the rent and a small number of special-purpose plants carrying much of that, plus a refinancing that will lift its interest costs over the next five years.
The distribution looks sustainable. The new CAD 3.70 is a 67% AFFO payout on the 2026 outlook midpoint, and the stress cases, a 10% fall in AFFO or a full refinancing at 4.5% with no growth, push the payout to between about 73% and 79%, not above 100%. A cut would take a much harsher shock than I tested, although it remains a board decision and not a contract.
The valuation is the issue. At CAD 88.50, or CAD 29.50 after the split, my probability-weighted ten-year return is 7.1% a year, with scenarios of 3.2%, 7.1% and 9.9%. That does not clear the 9.5% hurdle I apply to a new purchase, and Granite does not meet the hurdle for a new purchase at this price. Only the bull case reaches it, and it needs AFFO growth of about 5.9% a year for a decade at a 15 times exit, which is roughly the pace management guides for 2026 sustained for ten years. The units become attractive at or below about CAD 73.40 (CAD 24.47 after the split), which is the hurdle price, and compelling at or below about CAD 65.70 (CAD 21.90). Between CAD 73.40 and CAD 82.30 the units are fairly priced, and between CAD 82.30 and CAD 88.95 they are full. This is a test for a new purchase and not a case for selling units already owned, because a holder’s decision depends on cost base, tax and alternatives that this article does not know.
What would change the thesis? On the positive side, fourth-quarter results that publish a new NAV well above CAD 103, same-property NOI growth holding near 6% through 2027, a 2027 refinancing at or below 4.0%, and a pullback in the units toward the low CAD 70s would each move the expected return up. On the negative side, in-place occupancy below about 95%, news of a Magna non-renewal at a special-purpose plant, a refinancing well above 5%, a further widening of the special-purpose capitalization rate, or net leverage rising above the mid-30s would each reduce it. I hold this view loosely, because the model rests on assumptions I chose, the NAV is stale, the price comes from a single source and the next annual report will update the picture.
This article is general information and my own opinion, not personalized investment, tax or legal advice. Investment models are estimates, not forecasts. Consider your own circumstances or consult a licensed adviser before acting.
