2026 Office Market Outlook: Data on Recovery and Risk

Where Does the Office Market Actually Stand in 2026?

The office recovery is real but shallow. Effective rents nationally only exceeded 2019 levels in the past five quarters and are up just 7.8% since, according to CompStak’s 2025 Biannual Office Market Report. Beneath that modest headline number sits a market splitting three ways: by geography, by lease-expiration timing, and by asset class.

For owners and lenders underwriting 2026 decisions, the aggregate recovery number is nearly useless on its own. The more useful read comes from an uneven market-by-market rebound, a wall of lease expirations concentrated in the next three years, and a widening performance gap between Class A and Class B/C assets that doesn’t move in the direction most people assume.

How Uneven Is the Office Recovery Across Markets?

Very. CompStak’s market-level Columbia CompStak Rent Index (CCRI) shows Dallas-Fort Worth and Boston posted the shallowest downturns and now sit 16.2% and 13.3% above pre-COVID levels, respectively. San Francisco and the Bay Area/San Jose, by contrast, remain 17.1% and 15.3% below their Q4 2019 baselines.

The CCRI, developed in partnership with Columbia Business School, fell 4.3% from Q4 2019 to the COVID trough nationally and stood just 7.8% above the Q4 2019 level by year-end 2025. That national figure masks a wide range of outcomes. Manhattan took the deepest initial hit among gateway markets, down 12.7%, but has since rebounded 28.5%. That makes it one of the strongest recoveries in the dataset despite starting from the steepest hole.

The takeaway for allocators: market selection matters more now than at any point since 2020. A national office rent narrative, bullish or bearish, tells you almost nothing about what an asset in San Francisco or Dallas is actually doing.

Source: CompStak’s 2025 Biannual Office Market Report — Parts One, Two, and Three. Data as of Q4 2025.

What Is the Office Expiration Wall, and Why Does It Matter Now?

More than 32% of all office leases are set to expire between Q2 2026 and year-end 2028, per CompStak. That concentration means rollover risk and repricing opportunity are both compressed into a three-year window, rather than spread evenly across a typical lease cycle.

For asset managers, the expiration wall cuts two ways. Leases signed at pre-pandemic or trough-era rents are coming up for renewal precisely when many submarkets have posted double-digit rent recovery. That’s the upside case. The downside case is tenant retention risk in markets where flight to quality has pulled demand toward newer, better-amenitized stock, leaving legacy Class B/C buildings exposed at renewal.

CompStak’s embedded rent-growth data adds a wrinkle to the standard Class A narrative. Across expiring space, 54.2% of Class B leases carry positive embedded rent-growth potential, compared with 52.6% for Class A. The largest upside pockets are concentrated in specific market-class combinations: Phoenix Class A leads at +20.5%, while New York City Class B posts +20.1%. Neither Class A nor Class B has a clean structural advantage heading into the expiration wall. It depends on where the asset sits and who signed the original lease.

Sublease supply compounds the picture. Nearly 30% of subleases roll through 2028 at discounts of roughly 25% below directly leased space, per CompStak’s Part Four report. That’s a meaningful shadow supply overhang for owners marketing space in the same submarkets. However, 63% of expiring subleases are tied to sublessors currently paying above market rent, which means a chunk of that sublease inventory will simply disappear as leases burn off rather than get re-marketed at a discount.

Source: CompStak’s 2025 Biannual Office Market Report — Parts Three and Four. Data as of Q4 2025.

Is the Class A/B Divide Getting Wider or Narrower?

Wider, but not in a single direction. Class B/C lease terms remain 8.5% below the Q4 2019 baseline, a gap that has held for 24 consecutive quarters. Non-Prime Class A has cleared the 2019 mark for six straight quarters. Prime Class A, meanwhile, has exceeded its 2019 lease-term baseline in only one of the past 24 quarters.

That last figure is counterintuitive. Flight to quality has been the dominant narrative in office leasing since 2022, yet Prime Class A tenants are signing shorter terms than they were before the pandemic almost every quarter. One read: tenants chasing trophy space are trading term length for optionality, willing to pay up for quality but unwilling to commit to the 10-to-15-year terms that used to define premium leasing. Owners underwriting Prime Class A on legacy WALT assumptions should treat this as a structural shift, not noise.

Concessions offer a partial counterweight. Concession ratios, while still elevated against historical norms, declined over the past two quarters for both Prime Class A and non-Prime Class A space, per CompStak. That’s a tentative signal that landlords are regaining some pricing leverage in the higher-quality tiers, even as term length stays compressed.

A related, less discussed force in the Class A story: AI-driven TAMI demand. More than 44% of TAMI office leasing is now attributed to AI-focused firms. That concentration is propping up Class A absorption in markets like San Francisco and Manhattan even as broader tech leasing stays cautious.

What the Nareit Office Portfolio Comparison Shows

CompStak’s portfolio comparison of FTSE Nareit Office index owners against all other office owners puts a number on the Class A advantage. Nareit-listed landlords score 54.4 out of 100 on CompStak’s weighted scorecard, versus 50.0 for all other owners. It’s a narrow but consistent edge. The gap traces largely to asset mix: Nareit landlords hold 53.8% of their leased square footage in Prime Class A space, compared with 38.7% for everyone else.

That concentration shows up across the metrics that matter to lenders and buyers. Nareit tenants pay $62.02/SF in place, a 36.9% premium over the $45.31/SF average across all other landlords. Nareit portfolios also carry a wider mark-to-market spread (9.6% versus 4.6%), longer WALT (69.9 months versus 63.2 months), and higher rent growth since 2019 (32.5% versus 24.5%). Concession ratios run slightly lower for Nareit owners, 14.7% versus 15.6%. Tenant quality scores, notably, are identical at 0.39 for both groups, suggesting the outperformance is about asset selection and rent capture, not tenant credit.

Source: CompStak’s 2025 Biannual Office Market Report — Part Three, Part Four, and CompStak’s portfolio comparison of FTSE Nareit Office index owners vs. all other office owners. Data as of Q4 2025 and February 2026.

What Does This Mean for Office Investment Strategy in 2026?

Hold periods are lengthening and pricing hasn’t fully cleared. Median office hold periods have extended by 3.3 years since 2018, reaching a post-COVID high of 7.4 years in 2025. That extension has coincided with median annualized sale-price declines of 5% to 6% in both 2024 and 2025, per CompStak.

Read together, the hold-period and pricing data suggest owners are waiting out a market rather than transacting through it. That’s consistent with the expiration wall dynamic: many owners are betting that renewing leases at current market rents, particularly in recovering markets like Dallas-Fort Worth and Manhattan, will do more for valuation than a sale at today’s cap rates. Whether that bet pays off depends heavily on asset class and submarket, which is exactly where the Class A/B divide and the expiration wall intersect.

Key Takeaway: The 2026 office market isn’t one recovery story. It’s three simultaneous ones: a geographic split (Dallas-Fort Worth and Manhattan well above pre-COVID rents, San Francisco still below), a timing risk concentrated in the 32%-plus of leases expiring through 2028, and a Class A/B divide where Prime Class A commands premium rent but shorter terms, while Class B/C carries more embedded rent-growth upside than the flight-to-quality narrative suggests.

FAQ

Has the office market fully recovered from the pandemic? Nationally, no. Effective office rents only exceeded 2019 levels in the past five quarters and are up just 7.8% since, per CompStak. Recovery varies sharply by market, from well above pre-COVID levels in Dallas-Fort Worth and Manhattan to still below in San Francisco and the Bay Area/San Jose.

What is the office lease expiration wall? It’s the concentration of lease expirations in a short window. More than 32% of all office leases are set to expire between Q2 2026 and year-end 2028, per CompStak, creating a compressed period of both rollover risk and repricing opportunity.

Is Class A or Class B office space performing better right now? It depends on the metric. Class B/C lease terms sit 8.5% below the Q4 2019 baseline, while non-Prime Class A terms have exceeded 2019 for six straight quarters. On embedded rent-growth potential, Class B edges out Class A, 54.2% positive versus 52.6%.

How big is the sublease discount in the office market? Roughly 25%. Nearly 30% of subleases roll through 2028 at that discount to directly leased space, per CompStak. However, 63% of expiring subleases are tied to sublessors currently paying above market rent.

How long are investors holding office assets in 2026? Median hold periods have extended by 3.3 years since 2018, reaching a post-COVID high of 7.4 years in 2025, alongside median annualized sale-price declines of 5% to 6% in 2024 and 2025.

Do REIT-owned office portfolios outperform other owners? Modestly. Per CompStak’s Nareit Office portfolio comparison, Nareit landlords scored 54.4 versus 50.0 for all other owners, driven mainly by a larger share of Prime Class A assets (53.8% versus 38.7%).

Which office markets have recovered the most since 2019? Manhattan posted the strongest rebound among gateway markets, up 28.5% off a 12.7% initial decline. Dallas-Fort Worth (+16.2%) and Boston (+13.3%) had shallower downturns and now sit well above pre-COVID rent levels.

Why is Prime Class A lease term length shrinking despite flight to quality? Prime Class A terms have exceeded the 2019 baseline in only one of the past 24 quarters, per CompStak, suggesting tenants are paying up for quality space but negotiating shorter, more flexible terms rather than long-term commitments.

Institutional owners and asset managers underwriting 2026 leasing and disposition decisions need lease-level visibility into market rent, term length, and concession trends by submarket and asset class, not national averages. Request a CompStak demo to see how CompStak’s analyst-reviewed comp data and CCRI benchmarks apply to your portfolio.