AI startups in New York are signing leases for offices four times bigger than they need, and Futurism AI has the receipts. The pattern points to something deeper than vanity square footage. It’s a tell about where venture money is going and what founders think they need to look like to survive the next funding round.
The numbers Futurism AI pulls from Wall Street Journal reporting are striking. Adonis, an AI health startup, took a 25,000 square-foot office at 3 World Trade Center. At a generous 250 square feet per worker, that space fits 100 people. Adonis had 25 employees when it signed. Today it’s at 50 to 60. Co-founder and CEO Akash Magoon framed it this way: “We were basically giving the company a size-12 shoe, and we were size-4 at the time. We figured it would be motivating.”
Fazeshift, another AI startup with about a dozen people, just opened its second Manhattan office. CEO Caitilin Leksana told the WSJ that customers were asking about in-person presence during diligence calls.
What’s actually happening
Software companies don’t need square footage. The work runs on laptops. Most engineers work from home anyway. So why the real estate spree?
Two pressures are colliding:
- Enterprise buyers want signs of permanence. A real office address shows up as a checkbox in vendor risk reviews. “Are you in-person?” used to be a weird question. Now it’s part of diligence.
- VC capital is still flowing, and founders feel pressure to spend it on something visible. A skyscraper lease photographs better than a higher cloud bill.
Benjamin Bass, vice chairman of JLL’s New York brokerage, told the WSJ that AI tenants routinely lease offices “vastly larger than their staff numbers actually require.” That’s the quiet quote in the story. Brokers see it across deals. It’s not one weird startup. It’s a pattern.
Why this matters now
The AI hype cycle has moved into the phase where everyone is asking the awkward question: when do these companies actually make money? Most don’t. Venture capital still subsidizes the majority of AI companies, as Futurism AI notes. Trophy space buys time, signals seriousness to enterprise buyers, and burns cash that has to be replaced.
There’s a precedent here. WeWork, Theranos, and a long list of dot-com casualties all leaned on the office-as-credibility trick. It works until the money slows. Then those long leases turn into anchors.
Practical takeaways
If you’re a founder:
- Don’t confuse looking real with being real. A 25,000 square-foot lease doesn’t ship product.
- If enterprise diligence is asking about your office, that’s a procurement artifact, not a real signal of company health. A virtual office address plus a co-working node in your customer’s city solves the same checkbox at a fraction of the cost.
- Cash runway matters more than headquarters square footage. Long leases shorten runway in ways founders consistently underestimate.
If you’re an enterprise buyer:
- Stop treating office presence as a proxy for stability. SOC 2 reports, named references, and revenue history tell you more than a Class A address.
If you’re a VC:
- Watch the burn line. Real estate decisions inside portfolio companies are now a useful tell. A skyscraper lease in year two is a bet that the next round closes on time.
The correction, if it comes, will look familiar. Empty Manhattan floors. Broken leases. Founders explaining why a size-12 shoe was the right call. For now the music is still playing, and the offices are still being signed. Full reporting available at Futurism AI.