What is driving the new wave of office demand from AI tenants? AI companies have become one of the fastest growing sources of office leasing, reversing the simple story that artificial intelligence only shrinks office demand by cutting headcount. According to Colliers, AI firms accounted for more than a third of the technology sector's office demand in the first quarter of 2026, leasing roughly 670,000 square feet, a sharp jump from a 12% share in 2025. For investors weighing the office sector, this is a development worth understanding alongside the bear case on AI and offices.
Key Takeaways
- AI tenants drove more than one-third of the tech sector's office demand in Q1 2026, leasing about 670,000 square feet, up from a 12% share in 2025, per Colliers.
- JLL Research estimates the Bay Area already holds roughly 17 million square feet of AI-occupied space and projects 20 to 35 million more square feet of AI demand by 2030.
- AI tenants lease differently, prioritizing speed and certainty over price, with some deals moving from first tour to signed lease in weeks.
- The same technology fuels a two-sided market: AI drives new leasing in top buildings while accelerating obsolescence in weaker ones.
- For investors, AI demand is concentrated by geography and building quality, so underwriting must treat it as a targeted tailwind, not a broad one.
What the Data Shows: AI Tenants Are Leasing Aggressively
The data shows AI tenants moving from a niche to a primary driver of office demand in a single year. Colliers reported that AI firms made up more than a third of the technology sector's office demand in the first quarter of 2026, with roughly 670,000 square feet leased, compared with a 12% share in 2025. That is a structural shift in who is absorbing space, concentrated in a handful of innovation markets rather than spread evenly across the country.
The forward projections reinforce the trend. JLL Research estimates the Bay Area alone already contains about 17 million square feet of AI-occupied office space and could add another 20 to 35 million square feet of AI demand by 2030. The companies powering the AI buildout, from frontier labs running models like Claude, ChatGPT, and Gemini to the startups built on top of them, are signing leases at a pace that stands out in an otherwise cautious office market.
Why AI Tenants Lease Differently
AI tenants lease differently because their priority is speed and certainty, not squeezing the last dollar out of a negotiation. As Adam Henick of Current Real Estate Advisors put it, the priority for many of these firms is speed and certainty rather than penny pinching, and some deals have moved from first tour to signed lease in a matter of weeks rather than the months a typical corporate deal takes. Flush with capital and racing to scale headcount, well funded AI companies treat office space as a tool for growth rather than a cost to minimize.
That behavior reshapes leasing dynamics for landlords. AI tenants often want move in ready, high quality space with power and infrastructure to support dense technical teams, and they will commit quickly when they find it. For owners of the right buildings in the right markets, this is a rare source of fast, high credit demand. It also explains why AI leasing is so concentrated: these tenants cluster where talent, infrastructure, and other AI firms already are.
The Two-Sided Story: Demand Engine and Obsolescence Risk
The honest read is that AI is a two-sided force in the office market, creating demand in some buildings while destroying it in others. The same wave that has AI firms leasing aggressively in premier space is also pressuring weaker assets, as tenants in many industries trim headcount and consolidate into better buildings. We covered that bear scenario in our analysis of whether AI will shrink commercial real estate office demand, and the supply side risk in our look at AI office obsolescence and the flight to quality.
Both things are true at once. Office demand is bifurcating: trophy and high quality buildings in innovation markets are seeing genuine new absorption from AI and a broader flight to quality, while commodity buildings in weaker locations face accelerating obsolescence. The municipal view matters too, as detailed in the New York City Comptroller's analysis of AI and office demand. The mistake is treating AI as a single directional force on offices; it is a sorting mechanism that widens the gap between winners and losers.
What It Means for CRE Investors and Underwriting
For investors, the practical implication is that AI office demand is a targeted tailwind that must be underwritten by geography and building quality, not assumed across the board. A well located, high quality asset in a market like the Bay Area, parts of New York, or another AI hub may genuinely benefit from this demand, supporting occupancy and rent assumptions. A commodity office building in a secondary market gets little of this tailwind and faces the obsolescence headwind instead.
This argues for specificity in underwriting. When modeling absorption and rent growth, investors should ask whether a given asset is actually in the path of AI demand or merely near it, and should treat AI tenant credit and durability with appropriate caution given how young and capital dependent many of these firms are. The wider context remains sobering: Deloitte's 2026 commercial real estate outlook found only about 5% of firms report achieving most of their AI goals, and the share reporting transformative impact fell to roughly 1% from about 12% a year earlier. Demand is real but uneven, and capital follows the buildings positioned to capture it. The AI Consulting Network helps investors pressure test those market and tenant assumptions before they reach a model.
How to Position a Portfolio for AI-Driven Demand
Positioning for AI demand means concentrating on the building and market characteristics these tenants actually want, and being honest about which assets will never qualify. The features that attract AI tenants are quality, location near talent and infrastructure, ample power, and the ability to deliver move in ready space quickly. Owners of assets with those traits should be prepared to move at the speed AI tenants expect, since slowness loses deals to faster landlords.
For assets outside the path of AI demand, the strategy is different and just as important: plan for repositioning, conversion, or disciplined exit rather than hoping for a demand wave that will not arrive. Investors who want to assess where a specific market or asset sits on this spectrum, and how to underwrite AI tenant demand realistically, can reach out to Avi Hacker, J.D. at The AI Consulting Network. Research from JLL and Colliers is useful for tracking where AI leasing is actually landing.
The Caveats: Concentration and Durability
The biggest caveats are concentration and durability, and investors should weigh both before extrapolating the headlines. AI office demand is heavily concentrated in a few markets and a narrow band of high quality buildings, so national averages understate the strength in hubs and overstate it everywhere else. A portfolio outside those hubs sees little of this demand regardless of how strong the aggregate numbers look.
Durability is the second question. Many AI tenants are young, fast growing, and dependent on continued funding, which makes their lease credit less seasoned than a blue chip occupier's. A pullback in AI funding or a consolidation among labs could slow leasing or leave space back on the market. The demand is genuine and large, but it is best underwritten as a powerful, concentrated, and still maturing force rather than a permanent fixture, and that distinction is what separates careful investors from those chasing a headline.
Frequently Asked Questions
Q: How much office space are AI tenants actually leasing?
A: According to Colliers, AI firms accounted for more than a third of the technology sector's office demand in the first quarter of 2026, leasing roughly 670,000 square feet, up from a 12% share in 2025. JLL Research estimates the Bay Area already holds about 17 million square feet of AI-occupied space and could add 20 to 35 million more square feet by 2030.
Q: Does AI increase or decrease office demand?
A: Both, depending on the building and market. AI is driving genuine new leasing in high quality buildings in innovation hubs, while accelerating obsolescence in weaker, commodity assets as other tenants consolidate. It is best understood as a sorting force that widens the gap between premier and secondary office space rather than a single directional trend.
Q: Why do AI tenants lease space so quickly?
A: Well funded AI companies prioritize speed and certainty over price because they are racing to scale teams. Industry brokers report deals moving from first tour to signed lease in weeks rather than months. These tenants typically want move in ready, high quality space with the power and infrastructure to support dense technical teams.
Q: Should investors underwrite AI office demand into their models?
A: Cautiously and specifically. AI demand is a targeted tailwind concentrated by geography and building quality, so it should only support occupancy and rent assumptions for assets genuinely in its path. Investors should also weigh the durability of young, capital dependent AI tenants rather than treating their demand as a permanent fixture.