Life Sciences Real Estate Steadies as Lab Construction Dries Up

A recovery that starts with supply

By late summer 2026, the main U.S. brokerages broadly agreed that the life sciences real estate market had stopped getting worse. That is a narrower claim than saying it is getting better. In an August report, Cushman & Wakefield described the sector as still oversupplied but moving from correction toward stabilization. It pointed to two things: a construction pipeline that has nearly run out, and venture funding at its strongest level in four years.16 Rents and vacancy have not yet turned. Average asking rents were $64.17 per square foot in the second quarter, 5.3% lower than a year earlier, and the firm expects rents to keep slipping in heavily supplied regions for some time. It said stronger funding and tenant demand could steady them over the next 18 to 24 months.16

In this cycle, "stabilization" mostly means that new supply has stopped arriving. Tenants are not yet returning in large numbers. Owners of lab buildings, the REITs that own them and the lenders behind those properties should plan around that difference.

The pipeline has nearly emptied

All of the major brokerages agree on construction. According to Cushman & Wakefield, life sciences inventory grew from 171 million square feet in 2021 to 239 million today. Space under construction has dropped to about 2% of inventory, down from 17% at the mid-2023 peak.16 Only 1.2 million square feet was delivered in the first half of 2026, 82% less than in the same period of 2025. Most remaining projects are pre-leased or build-to-suit.16 CBRE reported a similar trend at the start of the year. It put the lab/R&D pipeline at about 6 million square feet, the lowest since early 2019 and far below the 2023 peak of more than 37 million. It expected any construction after that to be limited to fully leased build-to-suit projects.11

The change in how projects get built may outlast the cycle. In February, Cushman & Wakefield reported that 63% of the roughly 6 million square feet due by the end of 2026 was preleased, and that build-to-suit projects made up nearly a quarter of 2025 deliveries.14 Construction-industry reporting shows lenders behaving the same way. Projects for anchor tenants with long leases still get financed, while some speculative buildings started during the boom have struggled to secure permanent loans.2

Where the data disagree

The forecasts and the measured data do not fully match. CBRE predicted in January that vacancy would level off at a cyclical peak of 23.3% during 2026.11 By the end of July it was reporting a record 23.8% lab/R&D vacancy, with asking rents down 8.1% from a year earlier, even as venture funding and biotech hiring improved.20 Cushman & Wakefield's second-quarter figures also showed vacancy still rising, to 24.3%.25 CBRE's first-quarter data included more than 1.1 million square feet of negative net absorption and a fifth straight quarter of falling rents.23

JLL reads the same period more positively. Its June lab report said availability had fallen by about 2 million square feet since mid-2025, including the largest quarterly drop in a decade. JLL concluded that the market had bottomed.18 Demand across Boston, the Bay Area, San Diego and Raleigh-Durham rose 44% year over year to nearly 8 million square feet in the first quarter.18 The same report also warned that rents face pressure through the end of the decade. It estimated a supply-to-demand ratio near 6:1 and said the market would need three times the peak annual absorption of the last cycle to reach equilibrium.1819

One likely reason for the difference is what each firm measures. JLL tracks availability, which counts space being marketed. Vacancy counts space that is physically empty, and it lags. A lab can be leased before it is occupied, and sublease space can be returned before it shows up as vacant. In practice, leasing activity has bottomed while occupancy has not. JLL's finding that tenants are moving into post-2020 buildings, while pre-2000 stock adds availability, suggests that aggregate figures hide two different markets: newer buildings and older ones.18

The split is geographic as well. Vacancy in the three largest hubs is above 30%.13 Cushman & Wakefield reported 3.6% vacancy in Los Angeles–Orange County, and said Philadelphia, suburban Maryland, New Jersey and Denver all stayed below 20%.1614 JLL found that secondary markets as a group lost nearly 3 million square feet of demand over three years while adding 4.4 million square feet of availability, so the smaller markets are not a uniform success story either.19

Alexandria as a test case

Alexandria Real Estate Equities, the largest lab REIT, shows how the market's problems reach a balance sheet. It reported $1.7 billion of impairments in the fourth quarter of 2025, mostly tied to land and projects in the Bay Area, and announced plans to sell about $3 billion of assets in 2026. First-quarter operating occupancy fell to 87.7% from 91.7% a year earlier. Chairman Joel Marcus called it "a very tough operating environment," and the company lowered its occupancy guidance by 150 basis points.23 Same-property cash NOI fell 11.7%, and cash rents on renewals and re-leasing dropped 15.8%.22

The second quarter was more mixed. Leasing rose 60% from the prior quarter to just over 1 million square feet, and the company kept its $6.40 FFO-per-share midpoint while narrowing the range.24 It also booked $222.5 million of impairments, mostly on land and lab-conversion projects. One example was a northern San Diego parcel now under contract to a residential developer.24 Alexandria has 1.4 million square feet of key lease expirations in 2027, carrying $100.5 million of rent, and expects that space to sit empty for 12 to 24 months.24

Alexandria is also changing who it leases to. The company is reconsidering five projects totaling 1.6 million square feet and is looking at "advanced technology" tenants, which Marcus describes as manufacturing or research users rather than ordinary AI office tenants.23 Executives named some research divisions of Amazon and defense contractor Leidos as examples. These tenants need heavy floor loads, extra HVAC capacity and secure space.27 They pay less than lab users but require less capital and start paying sooner.2123 JLL has seen the same trend across the market, with AI, robotics and "tough tech" companies taking lab buildings.18 This is a reasonable adjustment, but it gives up the rent premium that justified building labs in the first place.

The lending side

Lab loans are a small share of a commercial mortgage market whose problems are concentrated in offices. CRED iQ put the overall CMBS distress rate at 11.86% in May, with office at 17.11%, the highest of any major property type.5 Trepp reported a record office delinquency rate of 12.34% in January.3 More than $100 billion of CMBS loans mature in 2026.3 Atrium counts $3.7 billion of office CMBS maturities in Boston and $3.3 billion in San Francisco next year. Those are also two of the most oversupplied lab markets.7

Lab loans are already appearing in the distress data. The $598.6 million BMR Pool loan, secured by lab and office properties in Cambridge, the Bay Area, San Diego and Seattle, reached hard maturity in March. Occupancy had fallen to 73% and debt service coverage to 0.50x, and the loan was in special servicing.6 The loan is still current, and the borrower and special servicer agreed on a framework for a multi-year extension.6 CREFC's February update listed among newly delinquent loans one backed by a 17-property, 2.4 million-square-foot life science portfolio sponsored by BioMed Realty. The borrower had already paid down about 70% of the original balance through sales and refinancings.1

The pattern here is extensions combined with new equity. Lenders are rarely forcing sales. That fits a broader shift in which lenders want real concessions from borrowers before extending loans.3 Cushman & Wakefield said lab underwriting now depends on tenant credit, lease length and existing occupancy rather than enthusiasm for the sector.16 Lenders will finance leased buildings with strong tenants. Speculative space will mostly be handled through workouts.

Overlap with the office market

The connection between labs and offices now runs in both directions. JLL said more than 6.2 million square feet of lab space has already been reclassified into other property types, and it expects more distress.18 It had earlier projected that nearly 19 million square feet could be converted by 2030.26 In January, Alexandria dropped a planned office-to-lab conversion at 401 Park in Boston's Fenway, citing strong office demand and lab vacancy near 30%.29 Meanwhile, tenants' floor plans are moving back toward a roughly even split between lab and office, after lab space had exceeded 60% to 70% at the peak.13

Outlook

The data support calling the market stable, but only in a narrow sense. Supply has topped out and leasing activity has improved. Investment sales also grew, reaching more than $9.3 billion over the four quarters through June across 292 deals.16 Rents and occupancy are still declining, landlords are giving tenants the most leverage they have ever had, and leases for relocations have shrunk to about 62 months.18 For REIT investors, Alexandria's late-2026 lease deliveries and 2027 expirations will show more than any brokerage forecast. For lenders, the remaining risk is concentrated in older lab buildings in Boston, the Bay Area and San Diego.