Artificial intelligence is the most powerful disruptor to hit corporate real estate in decades. But here’s the twist: it’s driving demand in some corners of the market even as it guts it in others.
If you’ve been following the headlines, you’ve seen AI companies doubling their Manhattan footprints, snapping up 30-, 60-, even 90-thousand-square-foot leases in trophy buildings. But scratch beneath the surface and you’ll find another story: AI is collapsing the very headcount models that fueled office demand for half a century.
So which is it? A leasing boom or a demand collapse? The truth is: both.
It’s the ultimate paradox: AI is simultaneously shrinking traditional headcount-driven office demand while fueling a surge of new, capital-driven demand in emerging hubs.
For sophisticated tenants, the implications are profound. Let’s unpack both sides of the story, and then connect the dots on what it means for occupiers who need to plan not just for 2025, but for the AI-driven decade ahead.
Where AI Creates Demand: The Rise of the Applied AI Tenant
AI-native companies are scaling at breakneck speed. Tempus AI, Harvey AI, Synthesia, and Sigma aren’t downsizing… they’re expanding, and fast.
- Tempus AI: 39,600 SF at 11 Madison Ave., doubling size.
- Harvey AI: 34,000+ SF at Park Ave. South after two expansions in a year.
- Sigma: Quadrupling its NYC footprint with 64,000 SF at One Madison Ave.
- Synthesia: More than doubling space in Flatiron, anchoring the ad/media corridor.
These deals are not outliers. They’re part of a 21% surge in national tech leasing this year. For landlords, AI tenants are a godsend: well-capitalized, fast-growing, and image-boosting.
The driver here isn’t headcount in the old sense; it’s funding velocity and client proximity.
AI startups flush with VC cash need to establish credibility with enterprise customers. And they need to do it in prime locations with amenity-rich space that attracts scarce talent. So let’s talk about where this is playing out.
San Francisco: The Research Lab of AI
- 5M+ SF already leased, with projections up to 16–21M SF by 2030 (CBRE).
- Core tenants: foundational model developers — OpenAI, Anthropic, Scale AI.
- Leasing is heavily concentrated in SoMa and Mission Bay trophy towers.
- The paradox: while AI firms are growing, other tech headcount is shrinking. Legacy Class B buildings in older stock remain vulnerable, with vacancy stubbornly above 30%.
New York: The Applied AI Capital
- The deals are piling up: Tempus AI (11 Madison Ave.), Sigma (One Madison), Harvey AI (Park Ave. South), Synthesia (Flatiron).
- Focus is applied AI — law, healthcare, finance, media — sectors native to Manhattan.
- Trophy towers and amenity-rich spaces are thriving, even as suburban back-office footprints collapse.
- Active AI demand: 1.7M SF currently being hunted, not far off SF’s 2.5M.
Boston: The AI-Biotech Hybrid Zone
- Kendall Square is the epicenter. Demand is for lab-office hybrids where AI meets biotech and pharmaceuticals.
- Venture-backed biotech-AI crossovers (drug discovery, genomics) are fueling preleases in lab-enabled buildings.
- Space competition here is less about desks and more about wet labs + compute integration.
Austin & Denver: Engineering-Heavy Growth Markets
- Cost advantages + lifestyle draw engineers away from the coasts.
- These markets are absorbing engineering-heavy AI shops, especially those less reliant on proximity to enterprise customers.
- Expect rapid lease-up of modern mid-rise towers, while older suburban parks still struggle.
AI is following the same playbook, clustering where customers and capital are thickest. San Francisco remains the lab, New York the applied hub, Boston the biotech haven, and other cities are burgeoning tech regions with tons of skilled workers and far lower taxes.
AI companies are striking while the balance is still in the tenant’s favor, in the wake of persistent pandemic-era vacancies.
And as other industries struggle and pullback in headcounts, they are able to step into the newest towers, the most flexible fit-outs, and the brand-defining addresses that tell the market they’ve arrived.
Where AI Destroys Demand: The Collapse of the Apprentice Economy
But turn to traditional white-collar occupiers and the picture flips.
AI is hollowing out the roles that once justified vast office floors:
- Law firms: AI tools like Harvey cut legal research time 80%, reducing junior associate demand.
- Software shops: Copilot reduces project teams from 50 to 10 developers.
- Finance & accounting: McKinsey estimates nearly 40% of accounting tasks are automatable.
- Call centers & admin hubs: AI chatbots and claims processors are already eliminating thousands of seats.
For decades, corporate real estate was tied to the apprenticeship model with fresh graduates filling floors of desks, learning by repetition.
That system is collapsing. AI has vaporized the need for large entry-level cohorts, and with them the need for massive cubicle farms in suburban and Class B buildings.
Fewer employees = fewer desks = smaller leases.
Reconciling the AI Leasing Paradox
So how do we reconcile this contradiction?
- AI-Native Firms (Expansion): Creating new demand in premium buildings. They’re young, aggressive, cash-rich, and strategically positioning in global cities.
- AI-Impacted Firms (Contraction): Slashing legacy demand in routine back-office functions. They’re consolidating, renegotiating, or exiting space.
The paradox resolves when you realize AI is reallocating demand rather than simply destroying or creating it. The winners are high-quality, innovation-ready buildings in global hubs. The losers are Class B/C suburban cubes and legacy training floors.
What An AI Leasing Wave Means For Sophisticated Tenants
For corporate occupiers, the message is clear:
- Decouple Space from Headcount:
Headcount forecasts are obsolete. Model real estate needs against automation curves and functional demands, not seat counts. - Act on B/C Asset Exposure:
Vacancy in Class B assets is already >25% in markets like Chicago and San Francisco. These assets are in structural decline. Exit while you can. - Exploit Landlord Competition:
Landlords are desperate for AI tenants but still under stress elsewhere. Smart occupiers are locking in termination rights, TI allowances, and rent concessions. - Embed Flexibility in Portfolios:
Replace 10–15 year commitments with 3–5 year leases, auto-renewals, flex integrations, and swing space. Agility is the new currency. - Follow AI Clusters Strategically:
For law, finance, healthcare, and media tenants, proximity to AI firms could mean ecosystem advantages. For others, secondary growth markets (Austin, Denver) may offer value without the AI inflation premium
For 50 years, real estate planners forecasted needs by multiplying employees × square feet per desk. That formula is dead. AI has rewritten the math:
- AI-Native Growth: Funding + client access = demand for prime, flexible, expandable space.
- AI-Impacted Shrinkage: Automation + efficiency = demand collapse in back-office hubs.
The new model? Space demand = business model × AI adoption curve.
The Bottom Line
The two narratives aren’t contradictions. They’re two sides of the same coin:
- AI giveth: fueling massive new leases in Manhattan, Boston, and San Francisco’s top towers.
- AI taketh away: collapsing headcount in legal, finance, software, and customer support, vaporizing space in Class B suburban stock.
For tenants, the mandate is clear: stop treating real estate as a static headcount equation. Start treating it as a dynamic AI-adjusted strategy.
That’s how you avoid being stuck with bloated, obsolete space while your competitors position themselves in the hubs of tomorrow. And powerful, AI-empowered software can help you achieve this. Tenants need real data, real modeling, and real market intelligence to future-proof your portfolio.
That’s why REoptimizer® technology exists: to give tenants the intelligence edge in an AI-driven market. Our platform helps you identify underutilized space, run “what if” scenarios that account for automation’s impact on headcount, and model future needs with precision. It arms you with the leverage to renegotiate bloated leases, exit obsolete buildings, and embed flexibility into every deal.
With REoptimizer®, you don’t just react to AI’s disruption, you use it to re-architect your portfolio for speed, savings, and competitive advantage.
Learn more today.
