By The Roofing Brief Team. Last reviewed: August 2026.
Commercial membrane market share is usually quoted from paid vendor reports that never publish their sample. This report measures it a different way: from the U.S. Energy Information Administration’s own building survey. Using the 2018 CBECS public use microdata, single-ply and synthetic sheeting membranes cover 28.9% of all U.S. commercial roof area, 33.8% of low-slope (flat plus shallow pitch) roof area, and 44.3% of flat roof area. Built-up roofing covers 17.2% of all commercial roof area, down from 23.1% in 2012.
Executive summary
- U.S. commercial buildings hold roughly 62.0 billion square feet of roof area across 5.9 million buildings and 96.4 billion square feet of floorspace (Roofing Brief calculation from EIA CBECS 2018).
- Membrane (CBECS category “plastic, rubber, or synthetic sheeting, single or multiple ply”) is the single largest commercial roof surface by area at 28.9%, narrowly ahead of metal at 28.5%.
- On flat roofs specifically, membrane share is 44.3% and built-up roofing is 28.1%. Membrane and built-up together hold 72.4% of flat commercial roof area.
- Between CBECS 2012 and CBECS 2018, membrane share of commercial roof area rose 2.7 percentage points while built-up fell 5.9 percentage points. Metal rose 5.0 points.
- In buildings constructed 2013 to 2018, membrane holds 40.9% of roof area and built-up holds just 7.8%, the clearest available signal that new low-slope construction has moved to single-ply.
- CBECS does not split membrane into TPO, EPDM and PVC. No U.S. government dataset does. Every TPO-versus-EPDM-versus-PVC share figure in circulation is a vendor estimate with undisclosed methodology, and published vendor estimates disagree by more than 15 percentage points on which product leads.
- Only two of the major single-ply manufacturers file U.S. segment financials. Carlisle Companies reported $3,721.7 million of Carlisle Construction Materials revenue in FY2025; Amrize reported $3,301 million of Building Envelope revenue in FY2025.
- Vendor demand estimates for U.S. single-ply (about 40 million squares, or 4.0 billion square feet, forecast for 2028) cannot be reconciled with the 17.9 billion square foot commercial installed base without accounting for building types CBECS excludes. That reconciliation gap is quantified below.
Key findings with sources
- Single-ply and synthetic sheeting membrane covers 17,948 million square feet of U.S. commercial roof area, 28.9% of the total (Roofing Brief calculation, EIA CBECS 2018 microdata, United States, 2018).
- Built-up roofing covers 10,684 million square feet, 17.2% of commercial roof area (same source).
- Metal surfacing covers 17,693 million square feet, 28.5% (same source).
- Asphalt and fiberglass shingles cover 9,461 million square feet of commercial roof area, 15.3%, mostly on religious worship, lodging and small service buildings (same source).
- Of the 30,764 million square feet of roof area on buildings whose roof tilt is recorded as flat, membrane holds 44.3% and built-up holds 28.1% (same source).
- Membrane share of commercial roof area was 26.3% in 2012 and 28.9% in 2018 (Roofing Brief calculation, EIA CBECS 2012 and 2018 microdata).
- Built-up share fell from 23.1% in 2012 to 17.2% in 2018, a loss of about 2.2 billion square feet of covered area in absolute terms (same source).
- 66.4% of membrane roof area is flagged as a cool roof in CBECS 2018, against 47.4% of built-up, 48.4% of metal and 21.2% of asphalt shingle roof area (same source).
- Carlisle Companies’ FY2025 Form 10-K states that re-roofing “represents approximately 70% of our commercial roofing business” (Carlisle Companies Incorporated, Form 10-K for fiscal year ended December 31, 2025, filed February 2026).
- The same filing describes Carlisle as “one of four major manufacturers in the single-ply industry” (same source).
- The same filing describes an aging non-residential building stock, “over 70% of which is more than 25 years old” (same source).
- Amrize Ltd’s FY2025 Form 10-K reports Building Envelope segment revenue of $3,301 million in 2025, $3,375 million in 2024 and $3,113 million in 2023, equal to 27.9%, 28.8% and 26.7% of total company revenue (Amrize Ltd, Form 10-K for fiscal year ended December 31, 2025, filed February 18, 2026).
- Amrize acquired Firestone Building Products (renamed Elevate) in 2021, Malarkey in 2022 and Duro-Last, a PVC roofing systems manufacturer, in 2023, per the same filing.
- A vendor study of the U.S. single-ply market states that “reroofing applications account for about 80% of single-ply sales in most years” and forecasts demand of 40 million squares in 2028 (US Single-Ply Roofing (Plastic and Rubber Membranes) Market Report 2024, announced October 15, 2024; vendor estimate, methodology not disclosed).
- Published vendor splits of the single-ply market conflict: one 2025 set gives TPO 37.7%, PVC 34.3%, EPDM 21.8%; another 2023 set gives PVC 40%, TPO 35%, EPDM 25%. Both are revenue shares, not area shares, and both are global rather than U.S. only.
What “commercial membrane” means in this report
Membrane here means a sheet-good roof covering installed on a low-slope commercial roof: TPO, EPDM, PVC, KEE and similar thermoplastics and thermosets, plus multi-ply modified bitumen sheets. The EIA survey category that captures this is “plastic, rubber, or synthetic sheeting (single or multiple ply)”. Built-up roofing (hot asphalt, felts and often a stone ballast) is a separate CBECS category and is reported separately throughout.
Two boundaries matter for anyone quoting these numbers. First, CBECS respondents pick one predominant exterior roof surface, so a building with a membrane roof over a small metal canopy is counted entirely as membrane. Second, CBECS surveys commercial buildings, defined by EIA as buildings “in which at least half of the floorspace is used for a purpose that is not residential, industrial, or agricultural”. Factories, warehouses attached to manufacturing plants, farm buildings and apartment blocks are outside the frame.
Commercial membrane market share by roof area, 2018
Membrane is the largest single roof surface on U.S. commercial buildings by covered area, at 28.9%. Metal is effectively tied at 28.5% because metal dominates warehouse and service buildings, which are large and almost always single storey. Built-up sits third at 17.2%. The table below is a Roofing Brief calculation: floorspace divided by floor count, aggregated with CBECS survey weights.
| Roof surface (CBECS category) | Roof area (million sq ft) | Share of commercial roof area | Share of commercial floorspace |
|---|---|---|---|
| Plastic, rubber, or synthetic sheeting (single or multiple ply) | 17,948 | 28.9% | 32.8% |
| Metal surfacing | 17,693 | 28.5% | 21.8% |
| Built-up (tar, felts, or fiberglass and a ballast) | 10,684 | 17.2% | 20.0% |
| Asphalt, fiberglass, or other shingles | 9,461 | 15.3% | 14.6% |
| Slate or tile shingles | 2,510 | 4.0% | 4.1% |
| Wood shingles, shakes or other wood | 1,093 | 1.8% | 1.6% |
| Concrete | 800 | 1.3% | 2.1% |
| Other | 1,834 | 3.0% | 3.1% |
| All commercial buildings | 62,024 | 100% | 100% |
Source: Roofing Brief calculation from U.S. Energy Information Administration, Commercial Buildings Energy Consumption Survey 2018 public use microdata (released November 2021, revised December 2022), variables RFCNS, SQFT, NFLOOR and FINALWT. Geography: United States. Reference year: 2018.
Membrane share of low-slope roof area is the number most people actually want
Membrane is only installed on low-slope roofs, so the all-buildings figure understates its position in its own market. Restricting to buildings CBECS records as flat, membrane share rises to 44.3%. Adding shallow pitch buildings brings in a large volume of metal-clad warehouses and pulls membrane share back to 33.8%. Both cuts are given below because the “low-slope” boundary is a judgement call and the two answers differ by more than ten points.
| Roof surface | Flat roofs only: area (M sq ft) | Flat roofs only: share | Flat plus shallow pitch: area (M sq ft) | Flat plus shallow pitch: share |
|---|---|---|---|---|
| Membrane (single or multiple ply sheeting) | 13,620 | 44.3% | 17,663 | 33.8% |
| Built-up | 8,650 | 28.1% | 10,327 | 19.8% |
| Asphalt or fiberglass shingles | 3,081 | 10.0% | 5,926 | 11.3% |
| Metal surfacing | 2,643 | 8.6% | 13,598 | 26.0% |
| Concrete | 618 | 2.0% | 738 | 1.4% |
| Slate or tile | 485 | 1.6% | 1,427 | 2.7% |
| Wood | 360 | 1.2% | 840 | 1.6% |
| Other | 1,307 | 4.2% | 1,712 | 3.3% |
| Total | 30,764 | 100% | 52,231 | 100% |
Source: Roofing Brief calculation from EIA CBECS 2018 microdata, variables RFTILT and RFCNS. RFTILT is an interviewer judgement made against reference images, not a measured slope, which is the main limitation on this table.
Membrane is taking share from built-up, not from metal
Comparing CBECS 2012 with CBECS 2018 on identical variable definitions shows where the movement is. Built-up lost 5.9 percentage points of commercial roof area share in six years. Membrane gained 2.7 points and metal gained 5.0 points. Membrane and built-up are competing for the same flat roofs; metal is expanding mostly through new warehouse and service building construction rather than by displacing membrane.
| Roof surface | Share of commercial roof area, 2012 | Share, 2018 | Change |
|---|---|---|---|
| Membrane (single or multiple ply sheeting) | 26.3% | 28.9% | plus 2.7 pts |
| Metal surfacing | 23.5% | 28.5% | plus 5.0 pts |
| Built-up | 23.1% | 17.2% | minus 5.9 pts |
| Asphalt or fiberglass shingles | 18.5% | 15.3% | minus 3.2 pts |
| Slate or tile | 4.2% | 4.0% | minus 0.2 pts |
| Wood | 1.6% | 1.8% | plus 0.2 pts |
| Concrete | 1.5% | 1.3% | minus 0.2 pts |
| Other | 1.3% | 3.0% | plus 1.6 pts |
| Total roof area | 55,832 M sq ft | 62,024 M sq ft | plus 11.1% |
Source: Roofing Brief calculation from EIA CBECS 2012 and 2018 public use microdata. CBECS 2012 used nine RFCNS codes and CBECS 2018 used eight; codes 8 (“no one major type”) and 9 (“other”) in the 2012 file were merged to match the 2018 “other” category. The 2012 to 2018 comparison carries the sampling error of two independent surveys and should be read as direction and rough magnitude, not as a precise annual rate.
New construction has already moved: 40.9% membrane, 7.8% built-up
Slicing the 2018 stock by year of construction gives the closest thing to a specification trend without a shipments dataset. On buildings completed 2013 to 2018, membrane holds 40.9% of roof area and built-up holds 7.8%. On buildings completed 1946 to 1959, built-up still holds 27.3%. Built-up has not disappeared from the stock, it has stopped being specified.
| Year constructed | Roof area (M sq ft) | Membrane share | Built-up share | Metal share |
|---|---|---|---|---|
| Before 1946 | 3,927 | 22.7% | 21.0% | 16.9% |
| 1946 to 1959 | 4,656 | 30.6% | 27.3% | 14.7% |
| 1960 to 1969 | 6,477 | 28.6% | 23.9% | 20.0% |
| 1970 to 1979 | 8,373 | 28.9% | 18.9% | 26.3% |
| 1980 to 1989 | 8,604 | 24.2% | 18.2% | 30.1% |
| 1990 to 1999 | 10,831 | 28.5% | 13.2% | 37.7% |
| 2000 to 2012 | 14,020 | 29.2% | 14.7% | 33.2% |
| 2013 to 2018 | 5,137 | 40.9% | 7.8% | 29.8% |
Source: Roofing Brief calculation from EIA CBECS 2018 microdata, variables YRCONC and RFCNS. This is the roof surface present in 2018, not the surface originally installed, so re-roofed older buildings show their current membrane rather than their original built-up system. That biases the older cohorts toward membrane and makes the built-up decline in newer cohorts a conservative reading.
Which building types actually carry membrane
Membrane share varies from 8.6% of religious worship roof area to 49.7% of standalone retail roof area. Warehouses hold the largest absolute block of commercial roof area in the country at 14,537 million square feet, but only 26.1% of it is membrane because metal dominates that segment at 44.5%.
| Principal building activity | Roof area (M sq ft) | Membrane | Built-up | Metal | Shingles |
|---|---|---|---|---|---|
| Nonrefrigerated warehouse | 14,537 | 26.1% | 13.6% | 44.5% | 7.8% |
| Education | 8,272 | 35.2% | 26.4% | 17.8% | 13.3% |
| Office | 6,889 | 32.5% | 22.6% | 15.0% | 16.3% |
| Service | 5,421 | 11.8% | 12.1% | 60.4% | 11.5% |
| Retail other than mall | 4,539 | 49.7% | 10.7% | 21.0% | 11.0% |
| Strip shopping mall | 4,470 | 41.4% | 30.1% | 8.2% | 13.4% |
| Public assembly | 4,435 | 23.6% | 12.6% | 31.0% | 18.1% |
| Religious worship | 3,517 | 8.6% | 6.7% | 17.5% | 49.5% |
| Lodging | 1,575 | 19.6% | 12.7% | 9.1% | 33.4% |
| Food service | 1,148 | 24.0% | 20.7% | 21.7% | 14.0% |
Source: Roofing Brief calculation from EIA CBECS 2018 microdata, variables PBA and RFCNS. Rows are ordered by total roof area. Cell values below roughly 1,000 million square feet carry wide relative standard errors in the underlying survey.
Membrane roofs are the most reflective roofs in the commercial stock
CBECS carries a derived cool roof flag. Membrane leads every other surface: 66.4% of membrane roof area is flagged as a cool roof, against 48.4% for metal, 47.4% for built-up and 21.2% for asphalt shingles. This is the mechanism behind the TPO growth story, since white thermoplastic membrane is the default way to meet a reflectance requirement on a low-slope roof.
| Roof surface | Cool roof share of that surface’s roof area | Cool roof area (M sq ft) |
|---|---|---|
| Membrane (single or multiple ply sheeting) | 66.4% | 11,915 |
| Metal surfacing | 48.4% | 8,564 |
| Built-up | 47.4% | 5,060 |
| Concrete | 39.8% | 319 |
| Wood | 32.2% | 352 |
| Asphalt or fiberglass shingles | 21.2% | 2,008 |
| Slate or tile | 21.0% | 526 |
Source: Roofing Brief calculation from EIA CBECS 2018 microdata, variables RFCOOL and RFCNS. RFCOOL is a derived EIA variable, not a measured solar reflectance index value, and it does not distinguish a factory-white membrane from a field-applied reflective coating.
TPO versus EPDM versus PVC: what is genuinely known
No U.S. government dataset splits the membrane category into TPO, EPDM and PVC. CBECS stops at “plastic, rubber, or synthetic sheeting”. The Census Bureau does not publish product-level shipments at that resolution. Every published TPO-versus-EPDM-versus-PVC share number is therefore a private estimate, and the published estimates disagree on which product is largest.
| Source type | Reference year | TPO | PVC | EPDM | Basis |
|---|---|---|---|---|---|
| Vendor market report A (methodology not disclosed) | 2025 | 37.7% | 34.3% | 21.8% | Revenue, global |
| Vendor market report B (methodology not disclosed) | 2023 | 35% | 40% | 25% | Revenue, global |
| Vendor market report C (methodology not disclosed) | 2025 | not stated | 36.39% | not stated | Revenue, global |
| U.S. federal statistics | 2018 | Not collected. CBECS reports one combined membrane category. | Roof area, U.S. | ||
Three points a careful reader should take from that table. These are revenue shares, so a product with a higher installed price per square shows a larger share than its area would justify, and PVC and KEE systems price above TPO. They are global, so they include European markets where PVC has historically held a much stronger position than in North America. And they disagree by roughly 5 percentage points on PVC and by 3 points on TPO between two reports published a year apart, which is larger than any real one-year shift in a stock this size.
The one directional claim that multiple independent sources support is that TPO is gaining and bituminous systems are losing. A vendor study of the U.S. market states that TPO membranes “continue to gain market share in commercial applications, largely at the expense of bituminous roofing products”, citing labour scarcity favouring self-adhered sheets and TPO’s light weight allowing recover over an existing roof. The CBECS 2012 to 2018 movement above, which is federal survey data rather than a vendor estimate, points the same way: built-up down 5.9 points, membrane up 2.7 points. For a product level comparison of the systems themselves, see our TPO versus PVC roofing comparison and EPDM versus TPO analysis.
Who actually manufactures commercial membrane
Carlisle’s FY2025 Form 10-K describes the company as “one of four major manufacturers in the single-ply industry” without naming the other three. Of the large single-ply producers, only Carlisle and Amrize publish U.S. segment revenue. Johns Manville is a Berkshire Hathaway subsidiary and is not separately reported in Berkshire’s segment disclosure. GAF, owned by Standard Industries, is private and files no financial statements. This is the structural reason there is no credible public brand-share table for U.S. membrane.
| Company | Membrane brands | Reported segment | FY2025 segment revenue | Disclosure quality |
|---|---|---|---|---|
| Carlisle Companies | Carlisle SynTec, Versico, WeatherBond, Hertalan, Resitrix | Carlisle Construction Materials | $3,721.7 million | SEC 10-K, audited, segment level |
| Amrize Ltd | Elevate (formerly Firestone Building Products), Duro-Last | Building Envelope | $3,301 million | SEC 10-K, audited, segment level, includes shingles and wall products |
| RPM International | Tremco, POWERply, TremPly | Construction Products Group | $3.1 billion (FY ended May 31, 2026) | SEC 10-K, but roofing is not separated from sealants, coatings and services |
| Johns Manville | JM TPO, JM EPDM, JM PVC | Berkshire Hathaway “Manufacturing” | Not disclosed | No segment breakout |
| Standard Industries | GAF EverGuard, GAF TPO | Private | Not disclosed | No public filings |
Sources: Carlisle Companies Incorporated Form 10-K for FY2025 (filed February 2026); Amrize Ltd Form 10-K for FY2025 (filed February 18, 2026); RPM International Inc Form 10-K for fiscal year ended May 31, 2026. Brand lists are taken from the filings themselves where present. Segment figures are not like-for-like: Carlisle Construction Materials includes polyiso insulation and architectural metal, Amrize Building Envelope includes residential shingles and wall systems, and RPM’s Construction Products Group is mostly not roofing.
Two demand facts from these filings are worth more than any vendor forecast because a 10-K carries legal liability. Carlisle states that re-roofing “represents approximately 70% of our commercial roofing business”. Amrize attributes its 2024 Building Envelope growth to “strong demand from re-roofing activities” and reports that softer 2025 volumes were “partially offset by strong commercial roofing repair and refurbishment activity”. Commercial membrane is a replacement market, not a construction market. Our US commercial and multifamily roofing market report sizes the surrounding market, and the roof coating market size report covers the restoration alternative to full membrane replacement.
Original synthesis: four Roofing Brief calculations
Calculation 1: U.S. commercial roof area, 62.0 billion square feet
Roof footprint is not published anywhere. It has to be derived, because floorspace counts every storey while a roof covers only one. The method is floorspace divided by floor count, applied record by record with survey weights.
Using the published CBECS 2018 Table B7 groupings and their midpoints: 43,576 divided by 1 equals 43,576; 23,300 divided by 2 equals 11,650; 10,516 divided by 3 equals 3,505; 13,031 divided by 6.5 equals 2,005; 6,000 divided by 12 equals 500. Total: 61,236 million square feet.
Running the same logic on individual microdata records, where exact floor counts are available for buildings of one to nine storeys, gives 62,024 million square feet. The two methods agree within 1.3%. Varying the assumed midpoints for the top-coded floor categories (10 to 14 and 15 or more) across a wide range moves the microdata total only between 61,932 and 62,091 million square feet, because tall buildings hold a small share of total floorspace. We use 62,024 million square feet throughout and treat 61 to 62 billion as the defensible range. Inputs: EIA CBECS 2018 Table B7 and 2018 public use microdata. Limitation: mezzanines, basements and parking levels are counted as floors in NFLOOR, which slightly understates roof area.
Calculation 2: the Roofing Brief Membrane Displacement Index
Membrane share of commercial roof area rose 2.7 points from 2012 to 2018 while built-up fell 5.9 points, a displacement ratio of roughly 0.46 points gained per point lost. The remaining built-up loss went to metal and to the “other” category rather than to membrane. In absolute terms membrane covered area rose by about 3,289 million square feet while built-up covered area fell by about 2,221 million square feet, against a total stock that grew 11.1%.
The vintage cut sharpens it. Comparing the 1990 to 1999 cohort with the 2013 to 2018 cohort, membrane share moved from 28.5% to 40.9% (plus 12.4 points) while built-up moved from 13.2% to 7.8% (minus 5.4 points). Inputs: EIA CBECS 2012 and 2018 microdata, variables RFCNS, YRCONC, SQFT, NFLOOR, FINALWT. Limitation: two independent survey samples, so the index carries compounded sampling error and should not be read to one decimal place.
Calculation 3: apportioning the installed base against vendor share claims
Applying the two published vendor splits to the 17,948 million square foot membrane base shows how much the answer moves depending on which vendor you believe.
| Apportionment basis | TPO (M sq ft) | PVC (M sq ft) | EPDM (M sq ft) |
|---|---|---|---|
| Vendor split A (2025 revenue shares, renormalised) | 7,214 | 6,563 | 4,171 |
| Vendor split B (2023 revenue shares, renormalised) | 6,282 | 7,179 | 4,487 |
| Spread between the two | 932 | 616 | 316 |
The spread on TPO alone is 932 million square feet, roughly the roof area of every food service building in the country. That is the honest measure of how much is not known. These apportionments are also almost certainly wrong in a specific direction: they apply revenue shares to an area base, and PVC installs at a higher price per square than TPO, so PVC’s true area share is lower than 34% to 40% and TPO’s is higher. We publish the apportionment to show the uncertainty, not as an estimate to quote. Inputs: Roofing Brief membrane base from CBECS 2018; two published vendor share sets. Limitation: mixing a U.S. area base with global revenue shares is not methodologically sound and is presented only as a bounding exercise.
Calculation 4: reconciling top-down demand with the bottom-up installed base
A vendor forecast puts U.S. single-ply demand at 40 million squares in 2028, which is 4,000 million square feet a year. Divided by the 17,948 million square foot commercial installed base, that implies 22.3% of all commercial membrane roof area is replaced every year, or a 4.5 year replacement cycle. Single-ply systems are commonly warranted for 20 to 30 years, so the two figures cannot both describe the same universe.
Working the other way: at a 20 year average service life, 4,000 million square feet of annual demand implies a steady-state installed base near 80,000 million square feet, which is 4.5 times the 17,948 million square feet CBECS measures. The gap is not an error in either number. It is accounted for by four things, in likely order of size: manufacturing and industrial buildings, which EIA excludes from CBECS by definition and which are heavily membrane roofed; multifamily residential, which CBECS also excludes; membrane shipped in excess of roof area covered, because of seam laps, waste, walls, below-grade waterproofing and recover systems installed over an existing roof; and genuine growth in the stock since 2018. Anyone quoting a national membrane market size should state which of these four they included. Inputs: vendor demand forecast (2024 publication); Roofing Brief membrane base from CBECS 2018; EIA CBECS scope definition. Limitation: the 20 year service life is an assumption, not a measured figure, and the four reconciling factors are not individually quantified here.
Recommended charts
- Commercial roof area by surface, 2018. Data: the eight CBECS RFCNS categories in million square feet. Source: Roofing Brief calculation from CBECS 2018. Insight: membrane and metal are effectively tied for the largest share. Citation-worthy because no published chart shows roof area rather than floorspace.
- Membrane versus built-up share, 2012 and 2018. Data: two paired bars per surface. Source: CBECS 2012 and 2018 microdata. Insight: the displacement is one-directional.
- Membrane share by construction vintage. Data: eight vintage cohorts, membrane and built-up lines. Source: CBECS 2018, YRCONC. Insight: 40.9% membrane in the newest cohort against 7.8% built-up.
- Vendor disagreement on TPO, PVC and EPDM share. Data: two published vendor splits plotted as ranges. Insight: the error bars are wider than most reported trends.
- Cool roof penetration by roof surface. Data: RFCOOL share within each RFCNS category. Insight: membrane at 66.4% is the reflectance workhorse of the commercial stock.
Methodology
Source selection. Federal survey microdata was preferred over every other source. Where a figure could be computed from EIA CBECS public use files, it was computed rather than cited from a secondary summary. SEC filings were used for company and demand-structure facts because a Form 10-K carries legal liability for accuracy. Private market research was used only where no primary source exists, and is labelled as a vendor estimate every time it appears.
Inclusion and exclusion. Any statistic that could not be traced to a named source with a year and a public URL was excluded. Brand-level U.S. market share figures were excluded entirely because no verifiable source publishes them. Figures from unsourced trade blog roundups were excluded.
Handling conflicts. Where vendor estimates disagreed, both were published with the spread stated rather than one being selected. Where a Roofing Brief calculation disagreed with a vendor figure, both were published and the reconciliation was worked through explicitly rather than resolved by preference.
How the derived figures were calculated. Roof area equals building floorspace divided by number of floors, computed per survey record and aggregated using the CBECS final weight (FINALWT). Top-coded floor counts (994 for 10 to 14 floors, 995 for 15 or more) were assigned midpoints of 12 and 17, with a sensitivity test across 10 to 25 that moved the national total by less than 0.3%. Shares are computed on roof area unless a table says otherwise.
Known data limitations. CBECS 2018 is the most recent completed cycle; the 2018 reference year is eight years old as of this report. CBECS excludes industrial, manufacturing, agricultural and residential buildings. RFCNS records one predominant surface per building. RFTILT is a visual judgement, not a measured slope. Our microdata tabulation returns 96,527 million square feet of total floorspace against the 96,423 million published in Table B7, a 0.1% difference attributable to weight rounding, and no result in this report turns on that gap. All survey estimates carry relative standard errors published by EIA alongside each table.
Last updated: August 2026.
Source quality ranking
Tier 1 (primary, used for all derived figures): U.S. Energy Information Administration, Commercial Buildings Energy Consumption Survey 2018 public use microdata and detailed tables B7 and B10; EIA CBECS 2012 public use microdata; EIA CBECS scope and definitions documentation.
Tier 2 (audited corporate disclosure): Carlisle Companies Incorporated Form 10-K, FY2025; Amrize Ltd Form 10-K, FY2025; RPM International Inc Form 10-K, FY ended May 31, 2026.
Tier 3 (vendor estimates, labelled throughout): published single-ply membrane market reports covering 2023 to 2026, used only for the TPO, PVC and EPDM split and for U.S. demand in squares, always with the methodology caveat attached.
Excluded: brand-share claims without a stated source; contractor blog roundups; any figure that appeared only in AI-generated summaries; market-size numbers whose geography or basis (revenue against volume) was not stated.
Most quotable statistics
- Single-ply and synthetic membrane covers 28.9% of U.S. commercial roof area and 44.3% of flat commercial roof area (Roofing Brief analysis of EIA CBECS 2018).
- Built-up roofing share of commercial roof area fell from 23.1% to 17.2% between 2012 and 2018 (Roofing Brief analysis of EIA CBECS).
- On U.S. commercial buildings completed 2013 to 2018, membrane holds 40.9% of roof area and built-up holds 7.8% (Roofing Brief analysis of EIA CBECS 2018).
- Re-roofing is “approximately 70% of our commercial roofing business”, per Carlisle Companies’ FY2025 Form 10-K.
- 66.4% of commercial membrane roof area is flagged as a cool roof, the highest of any roof surface (EIA CBECS 2018).
- No U.S. federal dataset splits commercial membrane into TPO, EPDM and PVC. Published vendor estimates disagree by 5 percentage points on PVC’s share within a single year.
Data limitations
- The membrane category in CBECS combines TPO, EPDM, PVC, KEE and multi-ply modified bitumen. It cannot be split further from public data.
- CBECS 2018 has a 2018 reference year. A newer cycle has not been released as of August 2026, so the shares here describe the 2018 stock, not the 2026 stock.
- CBECS excludes industrial, manufacturing, agricultural and residential buildings, so these figures are not a national roof-area census.
- Roof area is derived, not surveyed. Complex roof geometry, parapets, mezzanines counted as floors, and rooftop equipment are not modelled.
- Vendor market share figures are revenue-based and global, and are not comparable with the U.S. area shares in this report.
- The 2012 to 2018 comparison rests on two independent survey samples and on the assumption that the RFCNS question was interpreted consistently across cycles.
Recommended fields for a downloadable dataset
- survey_year, geography, building_activity, year_constructed_band, floors_band
- roof_surface_category (eight CBECS RFCNS values), roof_tilt (flat, shallow, steep)
- buildings_thousands, floorspace_million_sqft, derived_roof_area_million_sqft
- share_of_roof_area, share_of_floorspace, cool_roof_share
- relative_standard_error, source_table, source_url, method_note
Press summary
New analysis of U.S. Energy Information Administration microdata finds that single-ply and synthetic membranes now cover 28.9% of all U.S. commercial roof area and 44.3% of flat commercial roof area, making membrane the largest commercial roof surface in the country by covered area. The analysis, by industry publication The Roofing Brief, derives roof footprint from the Commercial Buildings Energy Consumption Survey by dividing floorspace by floor count, producing a national commercial roof area of about 62 billion square feet. Between the 2012 and 2018 survey cycles, built-up roofing lost 5.9 percentage points of share while membrane gained 2.7 points and metal gained 5.0 points. On buildings completed between 2013 and 2018, membrane holds 40.9% of roof area against 7.8% for built-up. The report also documents that no federal dataset splits membrane into TPO, EPDM and PVC, and that published private estimates of that split disagree with each other by up to 5 percentage points.
Suggested headlines
- Membrane is now the largest commercial roof surface in America, federal data shows
- Built-up roofing lost a quarter of its market share in six years
- 40.9% of new commercial roof area is membrane. Built-up is down to 7.8%
- Nobody actually knows the TPO versus PVC split, and the vendors disagree by five points
- 62 billion square feet: what U.S. commercial roofs are really made of
Frequently asked questions
What is the commercial membrane market share in the United States?
Single-ply and synthetic sheeting membranes cover 28.9% of U.S. commercial roof area, 33.8% of low-slope roof area and 44.3% of flat roof area, based on a Roofing Brief analysis of the EIA Commercial Buildings Energy Consumption Survey 2018 microdata. That makes membrane the largest commercial roof surface by covered area, narrowly ahead of metal surfacing at 28.5%.
What share of commercial roofs is TPO compared with EPDM and PVC?
No U.S. federal dataset answers this. CBECS records one combined membrane category. Published private estimates disagree: one 2025 set gives TPO 37.7%, PVC 34.3% and EPDM 21.8%, while a 2023 set gives PVC 40%, TPO 35% and EPDM 25%. Both are global revenue shares, not U.S. area shares, so neither should be quoted as a U.S. installed-base figure.
How much commercial roof area is there in the United States?
About 62 billion square feet across 5.9 million commercial buildings. The figure is a Roofing Brief calculation, derived by dividing CBECS 2018 floorspace by the number of floors in each building and aggregating with survey weights. Using the published grouped tables instead of individual records gives 61.2 billion square feet, so 61 to 62 billion is the defensible range.
Is TPO taking share from EPDM or from built-up roofing?
The federal evidence points to built-up. Between CBECS 2012 and 2018, built-up roofing lost 5.9 percentage points of commercial roof area share while the combined membrane category gained 2.7 points. A vendor study of the U.S. market reaches the same conclusion, stating that TPO gains share “largely at the expense of bituminous roofing products”. CBECS cannot separate TPO from EPDM within the membrane category.
What percentage of commercial roofing work is replacement rather than new construction?
Carlisle Companies’ FY2025 Form 10-K states that re-roofing “represents approximately 70% of our commercial roofing business”. A vendor study of the U.S. single-ply market puts reroofing at about 80% of single-ply sales in most years. Both figures describe sales mix rather than roof area, and the 10-K figure covers one manufacturer rather than the whole market.
Which building types use commercial membrane most?
Standalone retail leads at 49.7% of roof area, followed by strip shopping malls at 41.4%, education at 35.2% and office at 32.5%. Warehouses hold the largest absolute block of commercial roof area at 14,537 million square feet but are only 26.1% membrane, because metal surfacing dominates that segment at 44.5%. Religious worship buildings are the lowest at 8.6%.
Why do vendor market reports give bigger numbers than this analysis?
Because they measure a different universe. CBECS covers only buildings where at least half the floorspace is non-residential, non-industrial and non-agricultural, so factories, warehouses attached to plants and apartment blocks are excluded. Vendor demand figures also count membrane shipped rather than roof area covered, which includes seam laps, waste, wall applications and recover work over existing roofs.
How old is the commercial roof stock that membrane has to replace?
Carlisle’s FY2025 Form 10-K describes an aging non-residential building stock, “over 70% of which is more than 25 years old”. CBECS 2018 shows that 23,433 million square feet of commercial roof area sits on buildings completed before 1980, and that 18.9% to 27.3% of those older cohorts still carry built-up roofing, which is the pool most exposed to membrane conversion.
Are membrane roofs more reflective than other commercial roofs?
Yes, by a wide margin in the federal data. CBECS 2018 flags 66.4% of membrane roof area as a cool roof, against 48.4% for metal surfacing, 47.4% for built-up and 21.2% for asphalt shingles. The cool roof flag is a derived EIA variable rather than a measured solar reflectance value, and it does not distinguish a factory-white membrane from a field-applied reflective coating.
What is the most recent year of data available for commercial membrane share?
2018. CBECS 2018 is the most recent completed survey cycle with published microdata as of August 2026. Every roof-area share in this report describes the 2018 building stock. Vendor reports quote more recent years but do not publish their sample, their response rate or their definition of a square, so their recency comes at the cost of verifiability.
How to cite this report
The Roofing Brief, “Commercial Roofing Membrane Market Share Report 2026”, August 2026. Derived figures are Roofing Brief calculations from U.S. Energy Information Administration, Commercial Buildings Energy Consumption Survey 2012 and 2018 public use microdata. Company figures are from SEC Form 10-K filings by Carlisle Companies Incorporated, Amrize Ltd and RPM International Inc.
Related Roofing Brief research: the all-materials roofing market share report covers residential and commercial materials together, while this report isolates the low-slope membrane segment using primary survey microdata.