U.S. Bulk Distribution Market Outlook 2026 to 2030: One Vacancy Rate, Two Markets

MMCG Invest, LLC | Industrial Market Outlook | October 2026

The national vacancy rate for big-box warehouses, 9.6% at midyear 2026, describes a market that no longer exists. Underneath it sit two markets moving in opposite directions. Buildings of 750,000 square feet and larger, built since 2020 with 40-foot clear heights, are leasing faster than at any time since 2022 and have already pulled their vacancy down to about 7%. Buildings of 200,000 to 500,000 square feet, the size most developers built on speculation in 2022 and 2023, are stuck above 11% and are being carved into smaller suites to find tenants. This report sets out where each market stands, why the split opened, what it costs to build into it, and what a lender should expect to see in a feasibility study before financing a warehouse in 2027.

Sources: MMCG reconciliation of CoStar (1), CBRE (2), Cushman & Wakefield (3), JLL (4), Colliers (5), Savills as reported by Bisnow (6), Prologis (7) and CompStak (8). The method is set out in Section 2.

Four things are true at once in the American big-box market, and most commentary picks only one of them.

First, the overhang is being worked off. From the third quarter of 2022 to the end of 2025, developers delivered more large warehouse space than tenants absorbed in every quarter, and vacancy in buildings of 200,000 square feet and larger rose from a trough of about 4% to a peak of 10.3% in the third quarter of 2025 (1). In both quarters of 2026 so far, absorption has run ahead of deliveries. Every major data provider recorded the turn in the second quarter of 2026: CBRE's national vacancy rate fell for the first time since the second quarter of 2022 (2), Cushman & Wakefield's fell 10 basis points to 6.9% (3), JLL's fell 60 basis points to 6.8% (4), and Prologis reported 66 million square feet of net absorption, the most since 2022 (7).

Second, the recovery is not evenly spread. Occupiers are signing for the largest and newest buildings and leaving the older and mid-sized ones. In the first half of 2026, CBRE counted 38 leases of a million square feet or more, more than double the first half of 2025, and leasing of 700,000 square feet and above was up 125% (2). JLL puts vacancy in Class A buildings over a million square feet at 5.8% (4). At the same time, the 200,000 to 500,000 square foot cohort sits above 11% vacant (1), and 70% of the space under construction in that size band is still available, against 40% in the larger band (1).

Third, the pipeline has emptied and is now refilling from the bottom. Big-box space under construction is down more than half from its 2022 peak, and starts in the first half of 2026 were level with the first half of 2025 and lower than in any other first half since 2016 (1). The brokers see the all-size pipeline rising year over year (3, 4), which is a rebound from a trough rather than a new wave. Census data show the dollar value of private warehouse construction put in place still falling, down 8.1% year over year in August 2026 (9).

Fourth, at today's rents, speculative big-box development does not pencil in most of the country. Hard costs for a large distribution building average $78 per square foot nationally (10), the national bulk asking rent is $8.50 (1), and bulk product trades at a cap rate around 7.4% (1). Our arithmetic in Section 11 shows that a merchant developer in the inland and Sun Belt hubs needs rent of about $11 to $12 per square foot to earn a 150 basis point spread over the exit cap rate, 29% to 75% above what bulk space in those markets rents for today. This is why three-quarters of Prologis's second-quarter starts were build-to-suit (11) and why the pipeline will refill slowly.

The investment and lending implications follow from the split rather than from the average. Owners of modern 500,000 square foot and larger buildings in the inland hubs have pricing power again. Owners of mid-sized speculative buildings delivered in 2023 and 2024 face a choice between waiting for the cycle and spending $7 or so a square foot to demise them. Lenders should underwrite the two as different asset classes, and a feasibility study for a new project should prove its rent against its cost basis, its competing supply by size cohort, and its clear height against what tenants are actually signing for.

What counts as bulk distribution. We define bulk distribution as warehouse and distribution buildings of 200,000 square feet and larger, excluding flex, manufacturing and other specialised industrial types. This matches the threshold CoStar uses for its bulk distribution series (1) and the 200,000 square foot floor Colliers and NAI Hiffman use for their big-box reports, both of which add a minimum clear height of 28 feet (12, 13). The brokers' own "big-box" leasing statistics use a higher bar: JLL and Avison Young count leases of 500,000 square feet and above (4), and Cushman & Wakefield reports the 500,000-plus cohort separately (3). When we cite a big-box figure, we say which threshold it uses.

Why national vacancy rates disagree. As of the second quarter of 2026, the published national industrial vacancy rate ranges from 6.5% at CBRE (2) to 6.8% at JLL (4), 6.9% at Cushman & Wakefield (3), 7.2% at Prologis (7), 7.3% at Colliers (5), the mid-7% range at CoStar (14), 7.8% at Marcus & Millichap (15) and 8.2% at Savills (6). The 170 basis point spread is wider than any one source's quarterly change. The cause is inventory definition. Cushman & Wakefield counts competitive buildings above a minimum size set market by market and excludes most owner-occupied stock, which gives it an inventory of 18.3 billion square feet (3). CoStar's universe is broader, which is why its bulk-only pipeline of 280 million square feet exceeds CBRE's all-size pipeline of 252 million (1, 2). None of these figures is wrong. They measure different things, and the error is in mixing them in one sentence.

How MMCG's figures are built. The numbers labelled "MMCG estimate" in this report are reconciled from nine published datasets: the five national brokerage series, Prologis's operating data, MSCI's transaction data, CompStak's lease comps, and CoStar's bulk distribution series. CoStar's series is the calibration layer for the 200,000-plus cut because it is the only national dataset published at that threshold; where a broker publishes a cohort figure (Savills for 200,000 to 500,000 and 750,000-plus, Cushman & Wakefield for 500,000-plus, JLL for Class A over a million), we check CoStar's cut against it and show the range. Where we give a single number it is our central reading; where sources diverge we give the range and say why. Figures marked "derived" are our arithmetic on published inputs. MMCG does not maintain a building-level industrial database, and nothing in this report should be read as a proprietary count.

Sources: CBRE (2); JLL (4); Cushman & Wakefield (3); Prologis (7); Colliers (5); CoStar press release (14); Marcus & Millichap (15); Savills as reported by Bisnow (6); CoStar bulk distribution series (1).

The big-box cycle of the last four years can be told in five numbers. Vacancy in buildings of 200,000 square feet and larger bottomed at about 4.1% in the middle of 2022, when e-commerce tenants were still leasing ahead of demand. Developers responded with the largest supply wave in the history of the asset class: Cushman & Wakefield counts roughly 1.8 billion square feet of industrial deliveries between 2020 and 2023, more than the whole previous decade, with nearly half of the 2022 and 2023 deliveries in buildings of 500,000 square feet and larger (16, 17). Vacancy in the bulk cohort rose for thirteen consecutive quarters to a peak of 10.3% in the third quarter of 2025 and has since eased to 9.6% (1).

Source: Cushman & Wakefield year-end and quarterly MarketBeats (16, 17) (18, 19) (3); 2021 is not published as a single annual figure in the releases we used.

The turn came at the end of 2025: vacancy fell for the first time in the fourth quarter, and absorption overtook deliveries in the first quarter of 2026. Over the twelve months to June 2026, tenants absorbed about 206 million square feet of bulk space on a net basis, against net deliveries of about 175 million (1). Cushman & Wakefield's four-quarter rolling absorption for all sizes was 236 million square feet, against a trailing twelve months of deliveries we derive at roughly 253 million from its published annual and half-year totals (3, 17). The two series disagree on whether the all-size market has crossed over yet; they agree that the big-box segment has, because the demand recovery is concentrated in large buildings.

Three features of this recovery distinguish it from the 2021 boom. The first is who is signing. Third-party logistics providers remain the largest group of tenants, at 30 of the 100 largest leases in the first half of 2026 (20), but manufacturers took more than 12% of volume (2) and owner-users such as Walmart are buying buildings outright rather than leasing them (Section 10). The second is where. The Inland Empire led the top-100 leases with 12.6 million square feet, followed by Dallas with 10.5 million and Chicago with 9.4 million (20). The third is how long it takes. The median time to lease a vacant big box rose from 6.8 months in 2023 to more than 10 months in the second quarter of 2026, and median months vacant reached 11.7 (1). Space is being absorbed, but slowly, by tenants with time to negotiate.

The 9.6% headline is an average of a tight market and a loose one. Every published size cut shows the same shape: vacancy falls as building size rises above 500,000 square feet, and the trouble is concentrated in the middle.

Sources: Cushman & Wakefield (3); CoStar (1); Savills as reported by Bisnow (6); JLL (4); Colliers (12).

The leasing data explain the shape. Large tenants came back first and came back hard. JLL recorded a 58% year-over-year rise in leasing of 500,000 square feet and above in the second quarter (4). CBRE's 38 million-square-foot leases in the first half compare with fewer than 20 a year earlier (2). Prologis told investors in July that it had very limited availability above 500,000 square feet and none at all above a million (7). CompStak's renewal data show tenants in buildings of 500,000 square feet and larger renewing at an average 85% premium to their expiring rent in the first quarter of 2026, more than double the premium on smaller renewals (21).

Mid-sized buildings have no such tailwind. The number of leases signed in the 200,000 to 499,999 square foot cohort was running at roughly 2,000 a year in 2024 and 2025, and about 1,100 in the first half of 2026 (1), which is a steady pace rather than a recovery. The tenants who want 200,000 to 500,000 square feet are regional distributors, mid-sized 3PLs and manufacturers' suppliers, and many of them are still working through space taken in 2021 and 2022. Meanwhile the cohort carries the most speculative supply: 70% of the space under construction between 200,000 and 500,000 square feet is unleased, against 40% above 500,000 (1), and the biggest build-to-suits sit in the larger band.

The landlord response is to cut the buildings down. Colliers data for Philadelphia show that all five leases signed in speculative new warehouses inside the city between the first quarter of 2025 and the second quarter of 2026 were for demised space, and that 55% of the 38 speculative-building leases across the region were demised, accounting for 37% of the leased area (22). One owner converted a 759,000 square foot two-building complex to a multi-tenant plan and reached 71% occupancy with three tenants of 125,000 to 287,000 square feet (22). The economics are not trivial: a demising wall splitting a 287,000 square foot building costs about $2 million, or roughly $7 a square foot on the whole building (derived), before separate utilities, offices, restrooms and entrances (22). The trade is partial occupancy now against full vacancy later, and in markets with a deep small-tenant base it is the right trade.

Rents and pricing track the same divide, though in the opposite direction from what the vacancy numbers would suggest. The largest buildings are the cheapest per square foot: asking rents run about $8.90 in the 200,000 to 499,999 cohort, about $8.30 from 500,000 to 749,999, and about $8.00 above 750,000 (1). Sale prices follow: $128 a square foot for 200,000 to 500,000 square foot buildings against $115 above 750,000 (1). The reason is simple. A million-square-foot building has a thin tenant pool and a long re-leasing period, so investors demand a discount for the risk even when the current tenant is a credit retailer. The rent gap is the price of liquidity, not of demand.

Age matters as much as size, and the two overlap. The clearest finding in the 2026 data is that nearly all net demand is landing in buildings built since 2020. Cushman & Wakefield reports 137 million square feet of year-to-date net absorption in post-2020 buildings through June, of which 62.4 million was in facilities of 500,000 square feet and larger (3). Set against total year-to-date absorption of 113.6 million square feet, that implies pre-2020 stock gave back roughly 23 million square feet in the first half of 2026 (derived). CBRE counted more than 100 million square feet of negative absorption in pre-2020 buildings during 2025 (23). Our aggregation of the bulk series by year built shows buildings delivered since 2020 absorbing about 836 million square feet net from 2023 through 2025, while every older vintage lost occupancy in each of those years, about 286 million square feet in total (1).

Source: MMCG reading of CoStar's vintage series (1); absorption pattern from Cushman & Wakefield (3) and CBRE (23). Figures are rounded from charted series.

The vintage table explains a paradox that confuses many first-time readers of the data. The newest buildings have the highest vacancy, about 16%, and the oldest have moderate vacancy, about 8.5%. That is because the new stock is where the speculative wave landed and is still leasing up, while the old stock is mostly small, infill and already occupied by tenants who have nowhere cheaper to go. Obsolescence risk is not in old buildings as a class. It is in old bulk buildings: 32-foot clear, limited power, built for a tenant who has since moved to something taller. The new-building vacancy rate has fallen eight points from its peak in two years and will keep falling; the old-bulk vacancy rate will not.

Clear height is the specification that decides which side of the line a building falls on. Cushman & Wakefield found that 54% of large-format deals in the first quarter of 2026 were in buildings with 40-foot clear heights, and 61% of leases of 100,000 square feet and above were in buildings delivered this decade (24). In 2021 the firm estimated that more than 70% of warehouse inventory was built before 2000 and that 53% had a clear height of 27 feet or less (25). CBRE's occupier survey ranked clear height as the single most important building feature, with occupiers now seeking 36 to 40 feet against 18 to 32 feet in older stock (26). The mechanism is volumetric: a tenant pays rent on floor area but stores goods in cubic feet, so a 40-foot building offers 25% more storage envelope than a 32-foot building at the same rent per square foot (derived). Automated storage and retrieval systems are built to use that envelope, and Yardi Matrix makes the point that automation increases rather than reduces the need for new buildings, because the flat floors, column spacing and power that robotics require are rarely found in older stock (27).

The rent premium for height is now measurable. CompStak's regression of Southern California leases signed from 2023 to 2026 finds each additional foot of clear height worth about $0.21 a square foot in rent, up from about $0.08 for 2020 to 2022 leases, after controlling for size, location and age (28). On that coefficient a 40-foot building commands about $2.50 a square foot more than a 28-foot building in that market (derived). In New Jersey the coefficient is about $0.09 (28). Among newly delivered buildings nationally, the share with 40 feet or more of clear height has risen from close to zero before 2020 to roughly a quarter to 30% in 2025 and 2026, with another half at 35 to 39 feet (1).

Power is the second specification, and the one most likely to decide site selection by 2028. Cushman & Wakefield reports occupiers prioritising buildings with greater power capacity to support automation and AI systems (3). The grid is not keeping up. A July 2026 review for Advanced Energy United found no strong evidence that interconnection speed had improved across the seven ISOs studied, and ERCOT alone carried 1,965 active generation requests totalling 453 gigawatts (29). California's own warehouse law acknowledges the constraint by relaxing its plug-in requirement where power is unavailable (30). No source publishes utility service lead times by market for logistics sites, and a feasibility study should obtain a will-serve letter rather than assume one.

Demand for bulk space in 2026 has four sources, and they do not all point the same way.

E-commerce is growing at double digits again. Census data put e-commerce at 17.1% of retail sales in the second quarter of 2026, with online sales up 12.2% year over year against 6.7% for retail as a whole (31). Online sales had been growing at roughly 5% a year in 2024 and 2025; the share has added 0.8 points in the last four quarters (31). Prologis's long-standing rule of thumb is that online sales need about three times the logistics space of store sales (32), and the renewed acceleration is consistent with the return of million-square-foot fulfilment leases. Amazon, which spent 2022 and 2023 subleasing space, leased about 1.1 million square feet in Phoenix in January 2026 (33) and took delivery of build-to-suits in El Paso (34) and Otay Mesa in the first half.

Inventories are lean. The total business inventory-to-sales ratio was 1.30 in July 2026, down from 1.37 a year earlier and 1.40 two years earlier, and the wholesale ratio was 1.20 (35). These are cycle lows. Part of the decline is price inflation in the sales denominator, since the series is nominal, but the direction is clear: distributors ran down stock through the tariff disruptions of 2025 and have room to rebuild it. The Logistics Managers' Index confirms that inventory costs have been expanding faster than inventory levels every month since April 2025, by an average of 21 points against 12 points before, which means holding stock has become structurally more expensive and tenants will favour throughput buildings over storage buildings (36).

Imports are high in level and flat in growth. The ports of Los Angeles and Long Beach handled their busiest summer on record, with Los Angeles moving more than 2.9 million TEU from June to August 2026 (37) and Long Beach posting its busiest August (38). The National Retail Federation's Global Port Tracker forecasts 25.7 million TEU of imports for 2026, up 1% on 2025, with volumes easing in October and November after a September peak (39). The strength is timing, not growth. Importers front-loaded ahead of the July 2026 tariff changes, as they had in early and mid-2025 (40), and the Supreme Court's February 2026 ruling that IEEPA did not authorise tariffs, followed by the Section 122 and Section 301 replacements at 10% to 12.5%, has kept trade policy in flux (41, 42). For warehouse demand, the relevant effects are a pull-forward of storage need into the summer, a West Coast tilt in volumes that supports Inland Empire absorption, and a sustained premium for foreign-trade-zone and bonded space. The FTZ program handled nearly $964 billion of merchandise in 2024, of which $374 billion went through warehouse and distribution operations (43). No broker publishes a vacancy or rent series for activated zone space, so the premium is observed in operator commentary rather than measured.

Nearshoring is real and narrow. U.S. goods trade with Mexico set a monthly record of $94.8 billion in July 2026, up about 28% year over year, and Laredo was the country's largest trade gateway at $35.9 billion for the month (44, 45). Mexico was the largest U.S. trading partner in 2025 at $872.8 billion of goods trade (46). El Paso was among CBRE's top markets for absorption growth in the first half of 2026 (2). Against that, manufacturing construction spending fell to $168 billion at an annual rate in August, down about 20% from a year earlier, as the CHIPS and IRA megaprojects completed (9). Future reshoring demand will come from suppliers leasing and building to suit, not from new fabs, and the border markets have delivered speculative pipelines faster than they could absorb them (Section 9).

Sources: Census Bureau (31, 35) (9); NRF and Hackett Associates (39); Port of Los Angeles (37); Census and WorldCity via FreightWaves (44); Logistics Managers' Index (36); CBRE (2, 20).

The LMI deserves a closer look because it is the one monthly survey that asks logistics managers about warehouse space directly. In August 2026 warehousing capacity read 53.5, the loosest since December 2025, while warehousing prices read 75.0, up 2.8 points on the year and 11.2 points on two years (36). Capacity is easing and prices are still rising. That mismatch is exactly what a market with a tight modern segment and a loose older segment should produce: the space that is available is not the space tenants want, so they pay more for the space they do want.

The most argued-over question in the 2026 industrial market is whether the pipeline is shrinking or growing. The answer is both, depending on the baseline.

CoStar's bulk distribution series shows about 280 million square feet under construction at midyear, down more than half from the 2022 peak of roughly 715 million, with first-half starts of 99.6 million square feet, level with the first half of 2025 and lower than in any other first half since 2016 (1). Cushman & Wakefield's all-size pipeline rose for a fourth consecutive quarter to 305 million square feet, up 18% year over year (3). JLL's rose 9.2% to 276 million (4). CBRE's increased slightly to 252 million, with second-quarter starts of 58 million the highest in three years (2). The CoStar comparison is to the peak; the brokers' comparisons are to the 2025 trough. On our reading of the same CoStar series, big-box space under construction bottomed at 242 million square feet in the first quarter of 2025 and is up 14% year over year, the same direction the brokers report. Both are right, and the honest summary is that the pipeline bottomed in 2025 at roughly a third of its peak and is now rebuilding at a pace that will not threaten the recovery before 2028.

Sources: CoStar (1); CBRE (2); JLL (4); Cushman & Wakefield (3); Colliers (5); Yardi Matrix (47); Prologis (11); Census Bureau (9).

Three things about the rebuild matter for anyone underwriting a project that will deliver in 2028.

It is build-to-suit heavy. More than a third of the national pipeline is build-to-suit and two-thirds of that is in buildings of 500,000 square feet and larger (3). Three-quarters of Prologis's second-quarter starts had a tenant signed before the ground was broken (11). Speculative starts are returning, up 11% quarter over quarter at Cushman & Wakefield (3), and they are concentrated where big-box vacancy has already fallen: Chicago's pipeline is now 55% speculative, up from 36% a year earlier (48), and 42% of Houston's second-quarter deliveries arrived preleased (49). Yardi's count of 49 million-square-foot starts in the first eight months of 2026 reflects owner-users and preleased logistics campuses more than merchant developers (47).

Costs have stopped falling. Cushman & Wakefield's 2026 construction cost guide puts the core-and-shell cost of a large distribution building at $78 a square foot nationally, up 2.3% on 2025, with medium projects at $87 and small projects at $144 (10). The increase is in materials rather than labour: copper up 39% and steel up 17% under the metals tariffs, concrete up 4.4%, lumber up 3.7%, and construction wages up 4.3% (10). West Coast and Mountain markets run 15% to 35% above the national figure; the 2025 guide listed Seattle at $108 and Denver at $106 for large projects (10). Land is the bigger variable. Entitled industrial land in the Inland Empire has traded above $2 million an acre, roughly $50 a square foot of land, while comparable land in Atlanta or the Lehigh Valley trades for a tenth of that or less (Section 11).

Construction credit is available on tight terms. The Federal Reserve's July 2026 survey of loan officers found standards for construction and land development loans basically unchanged, at the tight end of their post-2005 range, with demand weaker (50). Debt funds hold record dry powder and are competing for the right borrowers (51). Published broker guidance puts speculative industrial construction loans at 60% to 65% of cost over SOFR plus 300 to 400 basis points, with lenders wanting 25% or more preleased, and build-to-suit at 70% to 80% of cost over SOFR plus 225 to 325 basis points; these are indicative figures from lender marketing rather than a survey, and no primary source publishes industrial-specific construction terms. With one-month SOFR near 4.5%, a speculative loan costs about 8% all-in, which is above both the exit cap rate and the development yield, so merchant spec is negatively leveraged during construction and depends on value creation at stabilisation.

Four constraints will cap the rebuild regardless of credit.

California's AB 98. The state's warehouse siting law took effect for new and expanded logistics projects on January 1, 2026, as amended by SB 415. It requires projects of 250,000 square feet or more near sensitive receptors, and any project needing a rezoning, to meet enhanced standards: separate truck entrances, loading bays oriented away from homes, planted buffers, idling limits and current energy codes (30). Jurisdictions in the Warehouse Concentration Region, which covers the unincorporated areas of Riverside and San Bernardino counties plus twelve cities, had to adopt designated truck routes by January 1, 2026 (52). The practical effect is to push new big-box development toward large industrially zoned sites away from housing and to raise the scarcity value of entitled and grandfathered land. No source yet quantifies the change in Inland Empire permit volume attributable to the law, and its effect cannot be separated from the cyclical collapse in starts.

Local resistance in the Northeast. New Jersey has no statewide moratorium, but a 2026 bill would require a state certificate of need for warehouses over 100,000 square feet (53), and townships are routinely imposing truck curfews and denying plans. In the Lehigh Valley, Pennsylvania's planning code prevents outright bans, but townships have removed warehousing as a by-right use and the Hanover Township amendment that would allow Majestic Realty's 2.1 million square foot airport project caps individual buildings at 400,000 square feet with a 400-foot buffer (54). The constraint is shrinking by-right acreage and longer approvals rather than prohibition.

Power. Transformer lead times of more than two years and interconnection queues measured in hundreds of gigawatts (29) mean that a site with secured power is worth more than a site without one, and increasingly worth more as something other than a warehouse.

Data centers. Colliers' March 2026 study found powered land trading at 1.63 to 2.5 times greenfield industrial land on national average over the prior two years, and more than three times in primary markets (55). Prologis started about $2 billion of data centers in the first half of 2026 and reports a power pipeline of 5.8 gigawatts (11). The largest logistics developer in the country is allocating land to data centers, and those sites leave the future warehouse supply. The effect is two-sided: data centers also generate warehouse demand. Link Logistics estimates about two million square feet of logistics demand per gigawatt of data center construction (56), and Prologis puts it at 30 to 40 million square feet per trillion dollars of data center capital spending (7). In Memphis, Link attributes about four million square feet of absorption to data center supply chains (56). Census data make the shift in capital plain: private data center construction put in place was running at $85 billion a year in August 2026, up 73%, against $53 billion for warehouses, down 8% (9).

National big-box asking rents are $8.50 a square foot triple net, up 3.5% year over year, against $11.41 and 2.0% for buildings under 200,000 square feet (1). The all-size national figures are $10.32 at Cushman & Wakefield (+2.9%, 3), $10.45 at JLL (+1.7%, 4) and $9.74 at Savills (+1.8%, 6). Big-box rent growth peaked near 14% in late 2022, fell below 3% at the start of 2025, and has since recovered to the mid-3s; small-bay growth has decelerated from 9% to 2% over the same period (1). For the first time since early 2024, the largest buildings are posting faster rent growth than the smallest, which is the leasing recovery of Section 4 showing up in price.

Headline rent is only part of the occupier's bill, and the rest of the bill has been moving against tenants.

Sources: CoStar (1); CompStak (8, 21); Prologis (11, 7); EastGroup (57, 58); Rexford (59); Savills (60); Newmark (61); Bureau of Labor Statistics (62, 63).

Three points stand out.

Concessions have peaked. CompStak's series for leases of 200,000 square feet and larger shows free rent rising from 1.7% of term in late 2022 to 4.9% in late 2025, then edging down to 4.8% in the first quarter of 2026, the first decline in eleven quarters (8). On a 62-month lease that is about three months free (derived). Prologis's own free-rent ratio rose from 2.9% to 4.4% of lease value over the year to June (11), so the giveback is real, but the direction has turned. Sublease space, which gave tenants a cheap alternative through 2024 and 2025, began shrinking in early 2026 (60) and Cushman & Wakefield names its decline as one reason vacancy fell in the second quarter (3).

Mark-to-market is still positive but compressing, and it varies by market more than by landlord. Prologis is still rolling leases up 37% on a net effective basis, EastGroup 34%, and both report cash spreads of roughly half that because escalators and free rent front-load the difference (11, 57). Rexford, concentrated in infill Southern California where rents peaked in 2022, is now rolling leases down, with cash spreads of −11.3% in the second quarter and guidance of −10% to −15% for the year (59). A lender applying a portfolio-average spread to a single building in a market that peaked early will overstate its income. The 2021 and 2022 vintage leases in Southern California and New Jersey are the ones to stress.

Rent is the smallest line in the tenant's budget. Prologis Research's estimate, which no later study has superseded, is that rent is about 5% of supply chain cost, or 25 to 50 basis points of a retailer's revenue, while transportation is roughly half and warehouse labour the largest cost inside the building (32). The Establish Davis database puts transportation at 4.62% of sales against 2.16% for warehousing including labour (64). Total U.S. logistics cost was $2.4 trillion in 2025, 7.8% of GDP (65). Warehouse wages have risen faster than rent: average hourly earnings in warehousing and storage were $26.84 in June 2026 (62), and the national median for forklift operators rose 12.6% in the three years to May 2025, about 4% a year compounded (63), against bulk escalators of 3.4%. The consequence is that a building which cuts labour hours or drayage miles can charge more than the market and still lower the tenant's total cost, which is why the 40-foot building wins the lease at a $2 premium and the 28-foot building loses it at a $2 discount.

Operating expense pass-throughs are the quiet third cost. Newmark's study of 18 major industrial markets found operating expenses at about a quarter of total occupancy cost, with property taxes 60% of that, common area maintenance 33% and insurance 7%; insurance was the fastest-growing component at 45% over three years (61). At a $10.32 national asking rent, a quarter of occupancy cost is about $3.44 a square foot, or $2.06 of taxes, $1.14 of CAM and $0.24 of insurance (derived). The study is from 2022 and no national update exists, so a feasibility study should build the expense line from the subject's own tax bill and insurance quote. Prologis's rental expenses rose 8.8% year over year in the second quarter against a 7.5% rise in rental revenue (11), which is the direction to assume.

National figures hide a 20-point spread in vacancy and a four-fold spread in rent. Exhibit 8 sets out MMCG's reconciled readings for the markets that matter to bulk distribution, grouped by their role in the network. The all-size vacancy is from the broker report named; the big-box reading is our rounding of CoStar's 200,000-plus cut, checked against broker big-box cuts where they exist.

Sources: CBRE market figures (66, 49) (67, 68) (69, 70); Cushman & Wakefield MarketBeat market tables (3, 34); JLL (71, 72); Colliers (73, 74) and via Lehigh Valley Business (75); Kidder Mathews (76); Avison Young (77); NAI Hiffman (13); Savills (78); Texas A&M International University compilation (79); CoStar bulk distribution series for the big-box readings and the rents marked CoStar (1). Where two brokers disagree we show both; the differences reflect inventory definitions, not errors.

A few patterns in the table are worth drawing out.

The inland hubs are leading the recovery. Dallas-Fort Worth absorbed about 24 million square feet of bulk space in the year to June, Phoenix 16 million, Indianapolis 15 million, Houston 14 million and Columbus 10 million (1). These five markets together account for close to 40% of national big-box absorption. They share cheap land, no siting legislation of the California kind, and rents between $6 and $9 that leave room for a tenant to trade up from older stock. Chicago is the standout for tightness: metro vacancy of 4.5% at JLL (72), big-box vacancy of 6.5% at Avison Young with rents up 6.6% (77), and only two available buildings of about a million square feet in NAI Hiffman's count of 870 big boxes (13).

The coasts are repricing. Los Angeles, the Inland Empire, Northern New Jersey, Philadelphia and Miami all show rents flat to down in the bulk cohort, and Rexford's negative spreads are the same story at the lease level (59). The Inland Empire deserves its own sentence because the brokers disagree about it more than about any other market: CBRE, Colliers and Lee & Associates record falling vacancy and positive absorption in the second quarter, while Cushman & Wakefield records rising vacancy and 3.8 million square feet of negative absorption year to date (66, 73) (3). All of them record record leasing. Our reading is that the West, closer to the ports, has tightened to about 6% while the East, where the 2022 and 2023 speculative wave landed, remains near 9.5% and is where AB 98 bites hardest.

The border and the Southeast built ahead of demand. El Paso's vacancy has fallen from 17.9% to 10.3% in a year on strong absorption (34), Laredo carries 11 million square feet under construction on 61 million of inventory (79), Austin sits above 16% on any measure, and Savannah, Jacksonville and Charlotte all show double-digit bulk vacancy. These markets will recover, because their demand drivers are real, but a new project delivering into them in 2028 will compete with 2024 product still in lease-up.

Bulk distribution buildings changed hands at a pace of about $34 billion in the year to June 2026, with more than $8 billion in the second quarter alone, at an average $122 a square foot, up 4.6%, and a cap rate of 7.4%, down 4 basis points (1). The all-industrial figures from MSCI tell the same story at larger scale: $32.5 billion of industrial sales in the second quarter, up 27% year over year, following $114.3 billion for 2025 (80, 81). Volume is recovering. Price is not: MSCI's industrial price index fell 0.4% year over year in the second quarter, reversing the sector's run as the pricing leader, before an annualised 4.5% rise from June to August (80).

Sources: Prologis (7); STAG (84); Terreno (82); Colliers (85); CoStar (1); MSCI via Colliers (80); Plymouth closing (86).

The spread from 4.1% to 7.5% is 340 basis points on the same asset class in the same quarter. It is not noise. It is the gap between coastal infill bought by a REIT with a 3.3% cost of debt (11) and secondary-market bulk bought by private capital with a construction-loan coupon near 8%. For a feasibility study the relevant number is the one for the subject's market tier and building size, and for most new big-box projects outside the gateways that is 7% to 7.5%. CBRE's midyear survey found investors expecting no further compression: 63% expect cap rates to hold over the next six months, 20% expect compression, down from 42%, and 17% now expect expansion (87).

Users are the net buyers. The most distinctive feature of 2025 and 2026 is that occupiers are buying their buildings. In the first half of 2026, users were the largest net buyers of bulk distribution property while private owners were net sellers (1). Newmark records owner-user purchase volume at a record $12.6 billion annualised (88). Walmart is the clearest case. It bought at least six distribution buildings totalling about 6.5 million square feet between May 2025 and May 2026, from a vacant 1.3 million square foot building in Kings Mountain, North Carolina, at about $93 a square foot, to a 507,000 square foot cold storage building in Riverside, California, at about $440 (89, 90). The largest single trade of 2026 was its $212 million purchase of a 1.2 million square foot East Hartford, Connecticut, building at $175 a square foot (1).

Sources: public records and local press as cited (89, 90); East Hartford per CoStar (1). Per-square-foot figures are derived from reported price and area. No cap rates were disclosed.

Why a retailer with 192 U.S. distribution facilities (91) would start buying the real estate is a question about automation rather than property. Walmart is retrofitting its regional distribution centers with automated systems, has more than half of its e-commerce volume running through automated facilities, and expects its supply chain capital spending to peak in 2026 and 2027 (92). At Kings Mountain the building cost $121.5 million and the equipment going into it about $90 million, with a further $58 million of upfit (89). When the machinery is worth more than the shell, owning the shell protects the payback on the machinery. Each such purchase removes a modern million-square-foot box from the pool investors can own, and sets the specification that the next speculative building will have to match.

Private capital is taking the public landlords private. Makarora and Ares closed the $2.1 billion acquisition of Plymouth Industrial REIT in January 2026 at an implied cap rate near 7.5% (86). Brookfield and CPP Investments agreed in July to buy LXP Industrial Trust for about $5.2 billion, a 20% premium to the 90-day average price, for 53 million square feet of Sunbelt and Midwest Class A warehouses (93). Rexford, meanwhile, raised its 2026 disposition target to $1.5 to $2.0 billion (59). The public market is pricing modern bulk distribution below what private buyers with long horizons will pay for it, and the gap is being closed by takeouts rather than by REIT share prices.

Credit is sound, with a watch item. Industrial CMBS delinquency was 1.14% in August 2026, against 12.0% for office and 7.85% overall (94). It nearly doubled from 0.65% in March, but the move came from one or two portfolio loans, including the $145 million National Warehouse & Distribution Portfolio that went to special servicing for imminent default in March (95). The watch item is the 2021 and 2022 vintage: loans underwritten at sub-4% cap rates now refinancing into transaction cap rates above 6% (81). Industrial loan originations rose 6% year over year in the second quarter (96), and CBRE's lending index shows permanent-loan underwriting at a 1.43x debt service coverage ratio, a 10.2% debt yield and 59.6% loan-to-value, all more conservative than a year earlier (97).

A warehouse feasibility study exists to answer one question for the lender: will the rent this building can achieve, in this market, against this competing supply, cover the cost of building it and the debt on it? In 2026 the arithmetic is tight enough that the answer is often no, and the study's job is to show exactly where it turns.

The development spread. The only audited, current disclosure of logistics development yields alongside exit cap rates is Prologis's. Its second-quarter stabilisations earned a 6.3% yield on cost against a 5.4% estimated cap rate, a spread of about 90 basis points, and its starts were underwritten at 7.2% against 5.4%, about 180 basis points, inflated by data centers in the mix (11). Those are the numbers of a developer with land bought years ago and debt at 3.3%. A merchant developer buying land today and borrowing at 8% needs more, and our rule for speculative product is a yield on cost at least 150 basis points above the exit cap rate.

MMCG assumptions and derivation. The $78 hard cost is Cushman & Wakefield's 2026 national average for large projects (10); the $85 and $95 figures for the Texas, Arizona and Inland Empire cases, the land prices, soft costs, site work and exit cap rates are MMCG assumptions, the land prices informed by the transaction evidence in Section 7. Bulk asking rents are MMCG readings of the CoStar series (1). The 0.89 factor converts triple-net rent to net operating income at 92% occupancy and 3% management. Tenant improvements and leasing commissions are excluded and would widen every gap. This is an illustration of the arithmetic, not a valuation.

The exhibit says what the starts data say. In none of the six markets does a speculative big box earn a 150 basis point spread at today's rent; at market rent, every case yields less than its own exit cap rate. Phoenix comes closest, because its rents have risen faster than its land; the Inland Empire is furthest, because land there costs about $45 a square foot. Even a Dallas-Fort Worth developer who accepts a 100 basis point spread, holds land at half today's price and builds 10% cheaper needs about $9.52, 16% above the market. That is why the pipeline is build-to-suit, why asking rents on space under construction run $9.46 against $8.50 for the standing stock (1), and why half of that space is preleased (1). Speculative big boxes get built by owners with land bought years ago, by sponsors who accept a thinner spread, or for a tenant who has already signed. A study that shows a speculative project earning a 7% yield on cost at a market rent in a 7.25% cap market has made an error somewhere, and the lender should find it.

Which lender, which test. The capital stack for a warehouse depends on who will occupy it.

Two of these programs are routinely misunderstood in warehouse deals. The SBA programs cannot finance a speculative or multi-tenant investment warehouse at all; they fit an owner-user who will occupy the majority of the building, and the occupancy percentages apply from day one of new construction (98). The USDA program can finance a distribution or cold storage building in a rural area up to $25 million, but the borrower's location test is by address, the guarantee percentage steps down above $5 million, and a new business borrowing more than $1 million must commission an independent feasibility study from a consultant acceptable to the agency (101, 102). Fiscal 2026 ended on September 30, so the guarantee and fee figures above should be checked against the fiscal 2027 notice before they are relied on.

What MMCG tests in a warehouse feasibility study.

What follows is MMCG's view as of October 2026. It rests on the published data above and on two assumptions: that big-box starts stay below about 250 million square feet a year through 2027, and that the import slowdown the NRF expects in the fourth quarter is a timing effect rather than the start of a contraction.

MMCG estimates. CoStar's own forecast for the all-industrial market is for vacancy to edge higher into 2027 before declining, with net absorption exceeding new supply by late 2027 and rent growth averaging 1.9% a year in 2026 and 2027 (14). Our bulk-only path is faster because the large-building segment has already crossed over.

Vacancy keeps falling, and the cohorts converge from the top. With absorption running about 30 million square feet ahead of deliveries and the pipeline shrinking, bulk vacancy should lose 50 to 100 basis points a year. The 500,000-plus cohort reaches a level that supports speculative starts in the inland hubs during 2027. The 200,000 to 500,000 cohort takes a year longer and clears mainly through demising and through older product leaving the competitive stock. Cushman & Wakefield's view that "peak industrial vacancy is likely in the rearview mirror" (24) applies to bulk now and to the all-size market by 2027.

Rent growth stays modest through 2027, then reaccelerates. With a 9.6% vacancy rate and 11% of availability in sublease space, landlords of standing stock will not push asking rents much past 3% next year. Growth comes back in 2028 as the pipeline that is being started now, at half the 2022 pace, proves too small for the absorption rate. The exception is the modern 500,000-plus building in a tight inland market, where rents are already rising 6% to 7% in Chicago (77) and the Prologis portfolio is rolling leases up more than a third (11).

Starts recover slowly and selectively. Merchant speculative development returns first in Chicago's I-80 corridor, Houston, Dallas-Fort Worth, Columbus and Indianapolis, where big-box vacancy is at or near the national average and land is cheap. Even there, on the Exhibit 11 arithmetic, it returns first for developers who already own land or accept a thinner spread, so the first wave will be small. It does not return in the Inland Empire or Northern New Jersey at scale until rents recover the ground lost since 2023 or land reprices, and in California AB 98 extends that timeline. Build-to-suit for retailers, manufacturers and data center supply chains carries the pipeline in the meantime.

Capital stays expensive and selective. Cap rates hold between 7% and 7.5% for bulk product outside the gateways. The 340 basis point spread between coastal infill and secondary bulk narrows only if the ten-year Treasury falls, which we do not assume. The take-private wave continues while public REITs trade below private-market value, and users keep buying modern boxes to protect automation capital.

The risks are to timing, not direction. The National Retail Federation expects imports to fall year over year in October and November (39); a longer slowdown would delay the 2027 crossover for the 200,000 to 500,000 cohort by two to four quarters. A reversal of the Section 301 tariffs would reduce FTZ and bonded demand at the margin. The USMCA review and tariff proposals against Canada are the main threat to border-market absorption. Insurance and property tax inflation could outrun the rent growth we project in Gulf and Florida markets. And the 2021 and 2022 vintage loans in Southern California and New Jersey will produce some distress as they refinance, which will show up in CMBS delinquency before it shows up in rents.

None of these risks changes the central fact of the market. For four years the big-box story was oversupply. From here it is a supply shortage in the making for the buildings tenants actually want, and a slow repositioning of the ones they do not.

What is a bulk distribution warehouse?

A warehouse or distribution building of 200,000 square feet or more, typically with clear heights of 32 to 40 feet, cross-dock or rear-load configuration, and trailer parking, used for storing and distributing goods rather than manufacturing. CoStar's bulk distribution series and the Colliers and NAI Hiffman big-box reports all use the 200,000 square foot threshold (1, 12) (13); brokers' big-box leasing statistics often use 500,000 square feet.

What is the vacancy rate for big-box warehouses in 2026?

About 9.6% for buildings of 200,000 square feet and larger at midyear 2026, against about 6.3% for smaller buildings and between 6.5% and 8.2% for the whole industrial market depending on the source. Vacancy in buildings of 750,000 square feet and larger is about 7%, and in Class A buildings over a million square feet about 5.8% (1, 4).

Is industrial vacancy still rising?

Not for big boxes. Vacancy in the 200,000-plus cohort peaked in the middle of 2025 and has fallen since. Every major broker recorded the first decline in all-size national vacancy since 2022 in the second quarter of 2026 (2, 3) (4). CoStar's all-industrial forecast has vacancy edging higher into 2027 before declining (14).

How much does it cost to build a large warehouse in 2026?

About $78 a square foot for the core and shell of a large distribution building nationally, $87 for a medium one and $144 for a small one, before land, site work, soft costs and tenant improvements (10). On our 2026 assumptions, all-in costs run from about $111 a square foot in Indianapolis to about $239 in the Inland Empire, with land the main difference (Exhibit 11).

What are warehouse rents?

The national asking rent for buildings of 200,000 square feet and larger is $8.50 a square foot triple net, and for smaller buildings $11.41 (1). The all-size national average is about $10.32 (3). Rents range from under $5 in Memphis to more than $17 in Northern New Jersey and Miami (Exhibit 8).

Can an SBA loan finance a warehouse?

Only for an owner-user. The operating business must occupy at least 51% of an existing building, or 60% of a new one rising to 80% within ten years, and speculative or investment warehouses are ineligible as passive businesses (98). The 7(a) maximum is $5 million and the 504 debenture maximum $5 million to $5.5 million.

Can a USDA loan finance a warehouse?

Yes, through the Business and Industry guarantee, for a borrower located in a rural area of under 50,000 population, up to $25 million, with an 85% guarantee under $5 million and 80% above in fiscal 2026. A new business borrowing more than $1 million must obtain an independent feasibility study (101, 102).

Which markets are best for new big-box development?

On the 2026 numbers, the inland hubs with big-box vacancy at or near the national average and cheap land: Chicago's I-80 corridor, Houston, Dallas-Fort Worth, Columbus and Indianapolis, though even there a speculative building needs rent about 40% to 75% above today's market (Exhibit 11). The coastal gateways have the rents but not the land price, and the border and Southeast growth markets still carry 2024 speculative product in lease-up (Section 9).

MMCG Invest, LLC prepares lender-facing feasibility studies and market analyses for commercial real estate projects, including industrial, distribution and cold storage. The figures labelled MMCG estimates in this report are reconciled from the published sources cited, with the method described in Section 2 and the assumptions stated in each exhibit. CoStar data are cited with attribution under license; CoStar's charts and market tables are not reproduced.

For general information. Nothing here is investment, lending or legal advice, and it should not be relied on for a specific transaction without a project-level study. The outlook in Section 12 is an opinion dated October 2026. Figures from third parties are reproduced as published and have not been independently audited. Figures marked derived are MMCG's arithmetic on published inputs.