The national office recovery just accelerated. That is the least useful way to read the new data.
Placer.ai’s H1 2026 office report found that visits across its nationwide index rose 6.0% year over year, up from 1.9% growth in H1 2025. Attendance is now 31.2% below H1 2019, the smallest first-half deficit since the pandemic.
Good news, certainly. But the investable signal is underneath the average: office recovery has become intensely local. Miami is within 11.5% of its 2019 baseline. Denver is still 44.6% below it. San Francisco remains deeply impaired, yet its office traffic is growing at double digits. And across the country, the employees showing up most often are increasingly the ones who live within five miles.
For owners, lenders and tenants, this changes the underwriting question. Stop asking whether “the office” is back. Ask whether this building removes enough friction to earn the trip.
The market table tells two different stories
| Market | H1 2026 vs. H1 2019 | YoY vs. H1 2025 |
|---|---|---|
| Miami | -11.5% | +7.7% |
| New York | -15.8% | +4.0% |
| Dallas | -24.9% | +5.4% |
| Atlanta | -28.9% | +7.4% |
| Houston | -35.8% | +5.9% |
| Washington, D.C. | -37.2% | +3.0% |
| Boston | -38.5% | +4.3% |
| Los Angeles | -39.2% | +13.0% |
| San Francisco | -41.4% | +10.9% |
| Chicago | -42.6% | +6.9% |
| Denver | -44.6% | +3.3% |
The first column measures recovery. The second measures momentum. Confusing the two is how investors overpay for a comeback story or miss one before the leasing market catches up.
Miami and New York are the recovery leaders. Miami sits just 11.5% below 2019 traffic and still grew 7.7% year over year. New York is 15.8% below baseline, helped by a finance sector that adopted stricter attendance policies earlier than many other industries.
Los Angeles and San Francisco are the momentum leaders. LA gained 13.0% year over year and San Francisco 10.9%. Their absolute attendance remains weak, but their direction is unmistakable. San Francisco’s AI leasing cycle is beginning to show up in pedestrian traffic, not just press releases.
Denver has neither advantage yet. It remains furthest from 2019 and grew only 3.3% year over year. In May alone, Denver was still 48.4% below its 2019 baseline on a per-working-day basis, according to Placer.ai’s monthly index.
Tuesday won. Friday did not.
The five-day workweek has not returned evenly. Tuesday office traffic was only 19.3% below 2019 in H1 2026. Wednesday was down 24.2%, Thursday 28.7%, Monday 34.7%, and Friday remained 52.9% below its pre-pandemic level.
This is no longer a temporary scheduling quirk. Midweek concentration has held for years. Buildings may be leased for five days, but many are being consumed as three-day assets.
The local differences are large enough to affect operations. Friday generated 14.9% of weekday visits in Miami and Dallas, compared with 9.9% in Chicago. New York, despite its strong overall recovery, posted a soft 11.9% Friday share. Houston came in at 11.1%, which Placer.ai connects in part to the energy industry’s use of compressed 9/80 schedules.
That should change how an owner staffs a lobby, schedules food service, prices flexible amenities and evaluates ground-floor retail. A lunch operator that looks weak on a five-day average may be excellent Tuesday through Thursday. A retailer built around Friday happy hour may be underwriting a day that no longer exists in that submarket.
The five-mile ring may matter more than the mandate
The strongest finding in the report is not about corporate policy. It is about distance.
Nationwide, office visits originating within five miles of the workplace were 15.7% below H1 2019. Every longer-distance group remained down by more than 35%. The short-distance cohort also grew fastest in H1 2026:
- Under five miles: +7.6% YoY
- Five to 10 miles: +6.0%
- 10 to 25 miles: +4.9%
- More than 25 miles: +3.8%
Placer.ai notes that this measures visits, not unique workers. That distinction matters. The data does not necessarily prove that everyone moved downtown. It suggests that people with easier commutes come in more frequently.
San Francisco had the largest close-in share at 48.1%, up 12 percentage points from 2019. New York reached 45.3%, while Chicago hit 41.5%. Houston barely moved, from 23.6% to 24.6%.
The office recovery is increasingly a commute-friction trade.
That is a location argument, but not the old “CBD versus suburb” argument. A suburban office near workforce housing may have a better five-mile catchment than a trophy tower dependent on a 70-minute regional-rail trip. The useful measure is not distance from downtown. It is distance from the employees a tenant needs.
What this changes in underwriting
1. Build a five-mile labor-shed map before relying on market vacancy. Count relevant workers, housing units, transit access and drive times inside the short-commute ring. A metro can recover while a commute-dependent submarket keeps losing visits.
2. Separate recovery from momentum. Miami and New York support a current-income thesis. Los Angeles and San Francisco may support a repricing thesis, but only where leasing, tenant quality and close-in workforce density confirm the traffic trend.
3. Underwrite the Tuesday peak and the Friday trough. Review elevators, parking, HVAC, security, cleaning and retail sales by weekday. Monthly averages conceal the operating pattern tenants actually experience.
4. Treat nearby housing as office infrastructure. Owners cannot shorten every commute, but they can favor sites connected to housing, support last-mile transit, and program services that reduce the cost of showing up.
5. Do not mistake a mandate for durable demand. Policy can compel attendance. A convenient building creates repeat attendance. The second is more defensible because it survives a change in management.
A note on what the index measures
Placer.ai’s nationwide office index tracks visits to roughly 1,300 top-tier office buildings. It includes commercial office buildings and offices with ground-floor retail, but excludes government buildings and residential-commercial mixed-use properties. The platform estimates visits using an anonymized device panel and machine learning.
That makes the index a strong measure of visitation direction, not a substitute for lease-level occupancy, badge data or tenant interviews. Use it to identify where to investigate, then verify the thesis at the building.
The bottom line
Office traffic is rising, but the recovery is not converging. It is splitting by market, weekday and commute radius.
The winners will not simply be the cities with the strictest bosses or the buildings with the newest amenity decks. They will be the assets that sit near the right workers and make Tuesday-through-Thursday attendance easy enough to repeat.
Pull the origin data on your building. What share of visits starts within five miles, and how has that changed since 2019? That number may tell you more about the next lease than the national return-to-office headline.
