A supply-side squeeze, not a demand boom, is doing the work. We pulled apart the Q2 2026 U.S. office market data, kept every number, and added the parts a typical CRE research note won't say out loud.
Seven straight quarters of improving absorption sounds like a recovery. It isn't, not the way anyone leasing space would recognize. Q2 2026 net absorption was still negative — landlords lost 360,000 more square feet of occupied space than they gained. The trailing four-quarter number looks strong only because the same period a year ago was ugly.
What's actually declining is supply. New deliveries hit a 14-year low. The construction pipeline is 60% below its long-run average. Owners are demolishing and converting obsolete buildings faster than tenants are filling the good ones. Vacancy is falling because the denominator is shrinking, not because demand is surging.
This is a scarcity story, not a demand story — and that distinction matters for how durable it is. Scarcity-driven tightening can reverse the moment someone starts building again. Demand-driven tightening compounds. Watch the construction line, not the vacancy line, if you want to know whether this is real.
Net absorption — the actual change in occupied space — has been a coin flip for five straight quarters. Two negative, three positive, no real trend line. The "7 consecutive quarters of improvement" headline is a trailing 4-quarter rolling average, which smooths this exact volatility out of view.
Rent hasn't moved in five quarters — up or down. That's the tell that landlords have zero pricing power in either direction right now. Nobody's desperate enough to cut, and nobody's confident enough to push. That's a market in a holding pattern, not a market in recovery.
This is the actual engine of the "recovery." Deliveries are down 24% YoY to a 14-year low. Space under construction sits at 19.7M sf — less than a third of the long-term norm — and inventory itself has shrunk 0.6% (-33M sf) over five quarters as obsolete buildings get demolished or converted to housing and other uses.
Nobody is building spec office right now, full stop. Only five metros even have >1M sf under construction. That's a gift to anyone holding well-located Class A — less competition is coming for years, not quarters — and a trap for anyone still holding commodity Class B/C, which this supply squeeze does nothing to save.
The national number hides a real divergence. The South posted the only meaningfully positive absorption quarter of the four regions; the Northeast — dragged down almost entirely by Midtown Manhattan giving back a chunk of its huge prior-quarter gain — posted the worst.
Class A vacancy fell 50 bps YoY nationwide — five times the improvement in the broader market — and captured +24.5M sf of trailing 4-quarter absorption, the best reading since mid-2020. The recovery, such as it is, is almost entirely a flight-to-quality story.
"Office is recovering" is doing a lot of work to describe a market where trophy towers in six metros are absorbing space and everything else is treading water. If your thesis is long office broadly, this chart is the reason to narrow it to Class A in demand-constrained submarkets — not the sector.
Pulled directly from the 92-market dataset — where vacancy is tightest and loosest, where rent is highest, and which gateway markets saw the sharpest YoY vacancy improvement.
Savannah at 3.8% vacancy next to Seattle at 32.7% is the same asset class on paper and a completely different investment case in practice. Averages in office are close to meaningless — this is a market you underwrite metro by metro, building by building, or not at all.