AI Now Wants Half of America's Tech Office Space — And It's All Landing on a Few Streets

If you want to see where the AI boom is physically going, don't look at funding announcements. Look at building tours. A new report from commercial real estate platform VTS offers one of the clearest signals yet of where AI companies are actually planting their flags, and the numbers are striking: AI firms now account for nearly half of all the office space America's tech industry is chasing.

The Numbers

VTS tracks live, forward-looking leasing pipelines rather than deals already signed, which makes its data an early warning system rather than a rearview mirror. Across the 17 US markets it follows, AI tenants make up 34% of active tech leasing requirements — but a disproportionate 46% of the actual square footage being sought. National AI office demand is up 85% year over year, and a startling 179% in the biggest AI hubs.

One caveat: VTS sells software to landlords, so this isn't a neutral academic study. But the platform handles more than 60% of Class A office space in the US, so the pipeline it's watching is real.

The capital fueling this is enormous. VTS points to the first quarter of 2026 as the largest AI funding quarter in history, with $226 billion in global private investment. OpenAI alone raised $122 billion at a roughly $850 billion valuation. Anthropic and xAI added tens of billions more. That money is now converting into desks — AI tenants are signing for bigger spaces too, averaging about 37,000 square feet per deal, some 37% larger than a typical tech lease. Leases above 50,000 square feet, once a shrinking category, are back.

Three Cities, One Street

Despite talk of AI demand "spreading," the reality is concentration, not dispersion. San Francisco, Silicon Valley, and New York together hold 63% of all active AI square footage nationwide. San Francisco alone accounts for 5 million square feet — nearly a third of the national total.

Zoom in further and it gets almost absurd: a single San Francisco submarket, the SoMa and Mission Bay corridor, holds a quarter of all active AI office demand in the entire country. This is where the foundation-model labs cluster. The report notes that Sierra, the agentic-AI startup co-founded by OpenAI chairman Bret Taylor, is set to take about 300,000 square feet in one building there, while OpenAI and Anthropic keep expanding nearby.

New York's AI tenants look different — leaning toward application-layer and infrastructure firms serving banks and law firms. The standout New York deal came from Nscale, an AI cloud company, which signed space at One Vanderbilt for $320 per square foot — the highest office rent Manhattan has ever recorded, according to VTS. Silicon Valley, meanwhile, keeps pulling in chip and hardware firms that want proximity to the semiconductor supply chain.

The New Map Forming Around the Edges

Beyond the big three hubs, a secondary geography is emerging — driven by talent scarcity, San Francisco's punishing real estate costs, and the sheer crowding of that one SoMa corridor. Seattle is the fastest-growing market VTS tracks, with AI office demand up 390% year over year, pulled by the region's deep engineering talent pool (a trend visible in Anthropic's own recent 113,000-square-foot Seattle lease). Northern Virginia is up 169%, riding defense and government AI work. Austin, Chicago, and Atlanta are all climbing too, though driven more by enterprise AI adoption than AI-native labs setting up shop.

Washington, DC is its own category entirely. In several of its submarkets, AI now accounts for more than 80% of all active tech office demand — not chatbot startups, but defense and intelligence contractors like Anduril, Shield AI, and Palantir, whose growth tracks the Pentagon budget rather than venture capital. Notably, two major tech markets sit out the boom altogether: Los Angeles and Boston, where AI office demand is actually falling.

Is This 2022 Again?

The obvious worry is a repeat of the last cycle's bust, when companies over-hired and over-leased ahead of revenue, then dumped staff and space when funding dried up. VTS argues this time is structurally different — today's leading AI tenants are scaling revenue faster and from a higher base, and many have already committed to build-outs years from completion.

That argument is reasonable, but it invites its own scrutiny. Demand this concentrated is also this exposed: if the capital propping up a handful of labs slows down, the same few streets carrying a quarter of national AI demand would feel it first. The broader office market recovery remains patchy nationally, and this wave leans heavily on a small number of companies raising enormous sums with comparatively little profit underneath them.

Why It Matters

The real value of the VTS report is what it reveals about hidden signals that citywide vacancy figures miss. Overall vacancy numbers still look soft in most markets — Seattle's, for instance, sits around 28% — which tempts landlords into pricing space at the citywide average. But that average is misleading. In a handful of corridors, concentrated AI demand is quietly tightening supply while the rest of the market notices nothing.

There's a forward-looking signal buried in the data too. In past tech cycles, law and consulting firms have historically followed the tech industry into new office space with roughly a year's lag. San Francisco's professional-services leasing demand is already up 33%, an early hint that spillover effects are starting. Legal-sector demand, by contrast, is down sharply — possibly a sign that AI tools are already eating into that industry's own workload.

Where AI companies choose to put their desks, in short, is a map of where they think the future goes — and that map tends to draw itself about a year before the rest of the economy catches up. Right now, it's pointing at a remarkably small number of streets.

Based on reporting by Cristian Dina, The Next Web, July 9, 2026, and Bloomberg.