The AI Killshot That Is Actually Saving Office REITs

Here is the trade the crowd built: artificial intelligence replaces knowledge workers, knowledge workers vacate offices, office landlords bleed out. It is a clean, logical three-step argument. The options market priced it. Short sellers funded it. The office REIT index crashed 12.8% in Q1 2026 alone, touching its lowest reading since the depth of the global financial crisis in 2009. The sector is universally despised.

The fundamental data disagrees with almost every word of that thesis.

In San Francisco, the city that was supposed to be ground zero for AI-driven office apocalypse, leasing volume reached 3.8 million square feet in Q1 2026, the strongest single quarter since 2014. By the end of Q2, year-to-date leasing hit roughly 7.0 million square feet. AI companies accounted for about 30% of all leasing activity in San Francisco since the beginning of 2023 and more than three-quarters of the market’s net absorption. Anthropic signed a 420,000-square-foot lease at 300 Howard Street in January. OpenAI’s San Francisco footprint has moved past 1 million square feet in Mission Bay. The supposed executioner of the office market is its most prolific tenant.

This is not a San Francisco anomaly. More than 44% of Bay Area TAMI office leasing has been attributed to AI-focused firms, propping up Class A absorption even as broader tech leasing stays cautious. The largest publicly traded office REIT just delivered a quarter that Wall Street’s AI-extinction thesis cannot explain. BXP’s Q2 2026 FFO per share came in at $1.78, beating both guidance and consensus. Leasing volume reached nearly 1.8 million square feet, 29% above the company’s own 10-year historical average for a second quarter. Management raised full-year 2026 FFO guidance to a range of $6.99 to $7.05 per share. Revenue for the quarter was $895.7 million. Occupancy climbed to 88.4%, up from 86.7% at year-end 2025. Total portfolio leased percentage, including signed-but-not-commenced leases, hit 91.3%.

The stock trades near $68. The 22-analyst consensus price target is $74.20. Barclays upgraded to Overweight in January with an $82 target. The question is not whether these analysts see opportunity. They clearly do. The question is why the stock price continues to lag the operating reality by so wide a margin. The answer is sentiment, which is now so heavily anchored to the wrong thesis that the disconnect has become the opportunity.

What the Market Expected vs. What Actually Happened

The consensus trade entering 2026 was straightforward: AI adoption crushes white-collar employment, white-collar employment is the demand driver for office space, therefore AI adoption crushes office demand. That concern is not wrong in principle. But markets do not trade on eventual possibilities. They trade on present cash flows and near-term fundamentals.

The present cash flows are not what the bears projected. Nationally, eight consecutive quarters of positive net absorption, totaling approximately 176 million square feet over two years, places the annual absorption pace on par with the 2010 to 2019 historical average. National office vacancy has also improved from a peak of 17.3% in 2024 to 16.1% as of Q1 2026. That is not a sector in structural collapse. That is a sector bifurcating sharply between trophy assets and commodity stock, with the former thriving while the latter legitimately dies.

CBRE’s 2026 Americas Office Occupier Sentiment Survey captured the RTO acceleration quantitatively: 89% of office-using organizations now require employees to work from the office at least three days per week, up from 78% a year earlier. Miami office foot traffic is running only 11% below pre-pandemic baseline. New York is 18.3% below. These are not numbers consistent with a dying sector.

The divergence between operating fundamentals and stock performance has a name. Piper Sandler senior REIT analyst Alexander Goldfarb called it directly in February: “Things are very good, getting better in a way that we’ve never experienced before, and yet the stocks suck. It’s a disconnect.”

That disconnect is where the contrarian case lives.

The CMBS Distress Trap Everyone Is Misreading

The other pillar of the bearish case is CMBS delinquency data, and it deserves precise treatment rather than blanket dismissal. Office CMBS delinquency hit a record 12.34% in January 2026. These numbers are real. The “extend and pretend” era is over, and CMBS workout specialist Michael Cohen of Brighton Capital Advisors put the situation plainly: properties that have lost value must now either be fixed or surrendered. That process is painful and ongoing.

But the CMBS distress data is a lagging indicator of pre-2022 vintage loans originated at peak valuations, many secured against secondary-quality buildings in markets that are genuinely impaired. It is not a leading indicator of demand at the Class A end of the market. The distress is concentrated in exactly the buildings the market was already abandoning: aging commodity space that a “flight to quality” tenant base has permanently vacated.

There are approximately $875 billion in commercial mortgage maturities scheduled for 2026. For owners of obsolete buildings, it is a crisis. For buyers of distressed assets with low basis costs, it is an opportunity. This pricing reset has lowered the ownership basis for new buyers, allowing them to compete aggressively on rental rates while still generating acceptable returns. The distress wave is not destroying the office sector. It is recapitalizing it at prices that make the economics work again.

Sector Implications: The Two Office Markets

Understanding the office sector in 2026 requires accepting that “office” is no longer a single investable category. The listed universe has sorted into at least four distinct camps with genuinely different geographies, asset qualities, and balance sheets. Secondary-market REITs, including Paramount, City Office, and Office Properties Income, have largely exited the public market. The names that remain are not the names that bear the most distress.

SL Green, Manhattan’s largest office landlord, reported Q2 2026 revenue of $278.9 million. Its FFO for the period was $109.6 million, or $1.45 per share. SLG’s Manhattan same-store office occupancy was 93.0% as of year-end 2025, inclusive of leases signed but not yet commenced. These numbers are far healthier than the national vacancy data suggests, because the national data includes hundreds of millions of square feet of commodity B and C product that is genuinely in secular decline. Conflating the distress of Worldwide Plaza with SL Green’s trophy portfolio at One Vanderbilt is exactly the analytical error the market is currently making.

The supply picture reinforces the premium-asset case. Office construction has fallen to very low levels in 2025 and is expected to remain constrained in 2026. New supply risk for Class A assets is essentially nonexistent in most gateway markets. CBRE expects scarcity of available prime space to intensify by year-end 2026, with spillover demand pushing into the next tier of quality. BXP management noted on the Q2 2026 call that AI is acting as a significant tailwind by driving space expansions and upgrades among tech and professional services firms.

Options Market Analysis

The options market on key office REITs reflects the same bifurcated reality playing out in the underlying assets. Implied volatility on names like BXP and SLG remains elevated relative to the broader REIT universe, which historically creates a selling opportunity for premium sellers when the fear is detached from deteriorating fundamentals. In BXP’s case, the stock delivered an $0.08 FFO beat against consensus, raised full-year guidance, and reported its strongest leasing quarter relative to history in years. The stock has not meaningfully re-rated to reflect those outcomes, leaving IV elevated against a fundamentally improving backdrop.

Short interest data from July 2026 confirms that mortgage and office REITs carry the highest short interest among small-cap real estate stocks, with bearish positioning concentrated in a handful of names. When short positioning is this crowded in a sector that is simultaneously delivering operational beats, the conditions for a compression trade are present. Any material positive catalyst, whether a Fed pivot, a large AI firm signing a flagship lease, or simply a string of in-line-to-better quarters, carries the potential to accelerate a short-covering move in the more liquid names like BXP and SLG.

The expected move framework for a position in BXP or SLG heading into Q3 earnings must account for the asymmetry between what the stock is pricing and what the operating metrics are delivering. At BXP’s current price around $68 and with the analyst consensus target at $74.20 and the more aggressive upside targets cited above, the options market is not fully pricing the probability of an operating beat cascading into a re-rating event.

Bull Case

For traders expecting continued AI-driven leasing momentum and a gradual Fed rate normalization, a defined-risk structure on BXP targeting the mid-$70s over a six-to-nine month horizon captures the analyst consensus move while limiting downside to the cost of the structure. A bull call spread buying the $70 call and selling the $77.50 call in the January 2027 expiry allows for defined risk while positioning for the re-rating event that three consecutive quarters of operational beats have not yet delivered.

Bear Case

If you believe the AI labor displacement thesis accelerates materially in H2 2026, reducing white-collar headcount at major tenants, the bear case concentrates in names with higher leverage and more commodity exposure, not in BXP’s Class A portfolio. A defined-risk structure using put spreads on names with significant B-quality exposure or near-term debt maturities, rather than the trophy-asset operators, gives the bear more fundamental support. The risk is that the short is already crowded and a squeeze remains possible even on bad news, if the bad news is less catastrophic than priced.

Neutral Case

A short-premium strategy exploiting elevated IV in the near-term options on BXP, such as a cash-secured put at the $65 strike through September expiry, captures decay against a stock that has shown operational resilience. This approach works if the stock continues to trade in a range and the AI-extinction thesis does not materially accelerate. The risk is a sharp move lower on a macro shock unrelated to office fundamentals, such as a credit event in the CMBS market that triggers indiscriminate sector selling.

Risk Analysis

The bear case has three legitimate components that traders taking the contrarian position must hold honestly. First, the CMBS maturity wall is real and the transfer of assets that workout specialist Michael Cohen described is ongoing. Any large distressed sale or forced liquidation in the office sector can create headline risk that drives indiscriminate selling across the quality spectrum, even when premium assets are unaffected. Second, the leverage profiles of office REITs are meaningful. Higher-for-longer rates extend the refinancing pressure on buildings already under stress and can spill into lender sentiment about the sector broadly. BXP is managing this through asset sales, having raised $1.2 billion in asset sale proceeds since its investor conference, and securing a $1.2 billion construction loan for 343 Madison Avenue at what management characterized as attractive terms. But the debt management story requires execution, not just intention. Third, the AI labor displacement thesis is a probability distribution, not a binary outcome. If generative AI meaningfully reduces white-collar headcount at law firms, financial services companies, and professional services firms over the next three to five years, the demand base for office space contracts in ways that even strong AI firm leasing cannot fully offset. BXP management acknowledged on the Q2 call that the long-term labor impacts of AI remain difficult to predict.

The counterargument is that the market has already discounted a severe version of that outcome by pricing office REITs at their worst levels since the 2008 to 2009 financial crisis, during a period when the operating data is materially better than it was during that crisis. That pricing discount is the opportunity.

Forward Outlook

The defining dynamic for office REITs through the remainder of 2026 is not whether demand exists. It does, and the AI-driven leasing data from San Francisco and other tech-heavy markets confirms it. The defining dynamic is whether the debt maturity wave produces a contagion event that spreads beyond genuinely distressed secondary assets into the trophy-asset portfolio. If it does not, the operating improvement at BXP, SLG, and Cousins Properties will eventually force a re-rating. The gap between BXP’s current price around $68 and the analyst target range of $74 to $82 represents that re-rating potential.

BXP’s long-term target is 91% occupancy by end of 2027, and management noted on the Q2 call that current leasing velocity creates potential upside to that target. A signed-not-occupied pipeline of roughly 1.3 million square feet gives the occupancy trajectory near-term visibility that is not reflected in the stock. The $3.6 billion development pipeline adds longer-term growth optionality against a supply backdrop where new construction has effectively slowed dramatically.

The crowd’s thesis is not irrational. It is simply misapplied. AI will displace some white-collar jobs. The companies building that AI are simultaneously signing some of the largest office leases in San Francisco. The market has priced the first effect and ignored the second. When that gap closes, the people who sold office REITs because “AI kills offices” will discover they sold them to the companies that are actually filling them.

Action Checklist

• Verify which office REIT names carry genuine trophy-asset exposure versus commodity B/C stock before establishing any position. BXP (Class A gateway markets), SLG (Manhattan premier), and CUZ (Sun Belt Class A) are not interchangeable with distressed secondary-market operators.
• Track the CMBS maturity resolution calendar. A large distressed liquidation in Q3 or Q4 can create entry points on the quality names via indiscriminate sector selling.
• Monitor AI venture capital funding velocity in San Francisco and New York. The direct line from VC funding to office leasing demand is a useful leading indicator, but it is not the only one.
• Watch BXP’s signed-not-occupied pipeline conversion through H2 2026. Pipeline conversion is the near-term catalyst for occupancy improvement.
• If constructing a defined-risk bull structure on BXP or SLG, size it against the elevated IV environment and target expirations that allow the Q3 2026 earnings event to serve as a catalyst rather than a near-term binary.
• For any bear exposure on the sector, isolate it to names with high leverage, near-term debt maturities, and genuine commodity-quality exposure. Shorting trophy-asset operators when they are already trading at crisis-era multiples, while AI firms sign large leases in their markets, is the crowd trade. It is also the wrong one.