What is AI office leasing demand? AI office leasing demand is the share of commercial office space being absorbed by artificial intelligence companies as they hire engineers, raise capital, and expand headquarters. In the first quarter of 2026, that demand reached a striking milestone: AI firms accounted for 11.5 million square feet, or 22.7%, of all office leasing across major US technology markets, according to CBRE data reported on May 29, 2026. For a sector still recovering from years of remote-work vacancy, AI tenants have become the single most important source of new demand. Yet the same technology propping up the office market may eventually undercut it. For the bigger picture, see our guide to AI commercial real estate tools and strategy.
Key Takeaways
- AI companies accounted for 22.7% of office leasing, or roughly 11.5 million square feet, across major US tech markets in Q1 2026, per CBRE.
- The San Francisco Bay Area is the epicenter, where AI tenants have absorbed large blocks of previously vacant Class A space.
- CBRE warns the trend could reverse as AI begins automating entry-level roles, reducing how much space companies need per employee.
- Office construction spending fell roughly 9% year over year to about $46 billion in March 2026, while data center construction surged past it.
- CRE investors should treat AI-tenant demand as a powerful but concentrated tailwind, not a permanent floor under office valuations.
AI Office Leasing Demand Explained
For most of the post-2020 era, the office story was about empty desks. National vacancy climbed, sublease space piled up, and lenders grew wary of office collateral. AI office leasing demand has rewritten that narrative in a handful of markets. CBRE reports that AI firms drove 22.7% of office leasing in the top technology hubs during Q1 2026, a concentration that would have been unthinkable three years ago. Companies like OpenAI and Anthropic, flush with multibillion-dollar funding rounds, are signing large long-term leases to house rapidly growing engineering teams.
The geographic concentration matters. The San Francisco Bay Area has been the clearest beneficiary, with CBRE noting that AI has been the key driver of leasing where several major tenants have absorbed large blocks of vacant space. New York City has seen a parallel, if smaller, surge. This mirrors a pattern we documented in our analysis of how AI companies are driving record office leasing in NYC and SF. The result is a two-speed office market: AI-adjacent submarkets are tightening, while secondary and tertiary office stock continues to struggle.
The Paradox: The Same Force That Fills Offices May Empty Them
Here is the tension that should keep CRE investors alert. The AI companies leasing today are building the very tools that could shrink office demand tomorrow. CBRE itself flagged the risk, observing that as AI capabilities begin to fulfill many entry-level job functions, companies may eventually need less space. If a 200-person firm can do the work of a 300-person firm, headcount growth slows, and headcount is the historical driver of square footage.
This is not a far-off hypothetical. We have already chronicled how AI is reshaping corporate footprints in pieces like Atlassian's office space reductions tied to AI and HSBC's plan for 20,000 AI-driven job cuts. The paradox is that AI demand is front-loaded (hiring booms during the build-out phase) while the displacement effect is back-loaded (efficiency gains arrive once the tools mature). Investors underwriting AI tenants on 10-year leases are effectively betting that growth outruns automation for the length of the term.
What This Means for Office Valuations and Underwriting
Office values hinge on net operating income (NOI), which is gross revenue minus operating expenses and excludes debt service and capital expenditures. AI-tenant demand supports NOI in three ways: it lifts occupancy, pushes asking rents in tight submarkets, and lengthens weighted average lease terms. Each of those compresses risk premiums and can support tighter cap rates, where cap rate equals NOI divided by purchase price. A building that re-leases a vacant floor to a well-capitalized AI tenant at a higher rent can see a meaningful jump in stabilized value.
But concentration is the catch. When a small number of specialized tenants account for an outsized share of demand, the market inherits their volatility. Consider these underwriting guardrails:
- Tenant credit and runway: Many AI tenants are venture-funded and not yet profitable. Scrutinize funding runway, burn rate, and whether the lease is backed by a parent guarantee or a sizable security deposit.
- Lease structure: Favor longer terms, healthy tenant improvement amortization, and renewal economics that survive a slowdown. Model a downside where the AI tenant does not renew.
- Submarket diversification: A portfolio overweight to one or two AI submarkets carries correlated risk. Spreading exposure protects debt service coverage ratio (DSCR), which is NOI divided by annual debt service and is typically expressed as a ratio such as 1.25x.
- Re-leasing assumptions: Stress-test downtime and concessions if an AI tenant vacates, since backfilling large blocks of specialized space can be slow.
For investors who want help building these scenarios into a model, The AI Consulting Network specializes in exactly this kind of AI-aware underwriting workflow.
The Capital Is Flowing to Data Centers, Not Offices
Even as AI fills select office towers, the larger capital story has moved elsewhere. US data center construction spending has surpassed office construction for the first time, a shift we covered in data centers surpassing office construction. Office building construction spending fell roughly 9% year over year to about $46 billion in March 2026, its lowest level since 2015, while data centers now exceed office construction outright. The signal for CRE investors is clear: AI is a demand driver for offices, but it is a capital magnet for digital infrastructure.
This reframes the office opportunity. The smart play is not to assume AI will single-handedly rescue the entire office sector. It is to target the specific Class A, amenity-rich, transit-connected buildings in AI-heavy submarkets that these tenants actually want, while underwriting conservatively for the day automation bends the demand curve. Broader context helps here: the AI in real estate market is projected to reach $1.3 trillion by 2030 at a 33.9% compound annual growth rate, and 92% of corporate occupiers have initiated AI programs, though only about 5% report achieving most of their AI goals.
Real-World CRE Applications
Owners and investors are already acting on the AI office leasing signal in practical ways. Landlords in San Francisco and Manhattan are repositioning vacant floors with the specifications AI firms prefer: high power density, robust connectivity, flexible floor plates, and collaboration-heavy layouts. Brokers are using ChatGPT, Claude, and Gemini to scan funding announcements and identify which newly capitalized AI startups are likely to outgrow their current space within 12 to 18 months. Asset managers are layering automation-risk scenarios into hold-versus-sell decisions.
The investors who win will be the ones who can hold two ideas at once: AI demand is real and material today, and AI-driven efficiency is a genuine medium-term threat to office absorption. CRE investors looking for hands-on AI implementation support can reach out to Avi Hacker, J.D. at The AI Consulting Network to translate these market signals into portfolio decisions.
Frequently Asked Questions
Q: How much office space are AI companies actually leasing in 2026?
A: According to CBRE data reported in late May 2026, AI companies accounted for about 11.5 million square feet, or 22.7%, of all office leasing across major US technology markets in the first quarter of 2026. The San Francisco Bay Area was the largest contributor to that demand.
Q: Will AI ultimately increase or decrease office demand?
A: Both, but on different timelines. In the near term, AI companies are major net absorbers of office space as they hire aggressively. Over the longer term, CBRE and others warn that AI automation of entry-level roles could reduce how much space firms need per unit of output, softening demand. Prudent investors underwrite for both phases.
Q: Should I underwrite a venture-backed AI tenant differently than a traditional corporate tenant?
A: Yes. Many AI tenants are not yet profitable and rely on venture funding, so analyze funding runway, burn rate, lease guarantees, and security deposits. Favor longer lease terms and model a downside scenario where the tenant does not renew, since re-leasing large blocks of specialized space can take time.
Q: Which office markets benefit most from AI leasing demand?
A: The San Francisco Bay Area leads by a wide margin, followed by New York City. Demand is concentrated in Class A, well-located, high-power-density buildings that suit engineering-heavy teams, rather than spread evenly across all office stock.
Q: Is office still a good investment if data centers are attracting more capital?
A: Office and data centers serve different strategies. Data centers are capturing the bulk of new AI-driven construction capital, but well-located office assets in AI submarkets can still deliver strong risk-adjusted returns if underwritten conservatively. The key is targeting the specific buildings AI tenants want rather than betting on a broad office recovery.