Can the AI boom save NY’s ailing commercial properties?

In December, President-elect Donald Trump and SoftBank CEO Masayoshi Son announced $100 billion in US investments to create more than 100,000 new jobs in artificial intelligence and related infrastructure.

Then, last week, Hussain Sajwani, the chairman of Dubai-based luxury property developer Damac, stood by Trump at Mar-a-Lago to announce a $20 billion investment in the Midwest and Sunbelt for AI data centers and chip making.

Chip makers are also investing $112 billion around New York State with the Albany NanoTech Complex becoming the home of CHIPS for America EUV Accelerator, one of three National Semiconductor Technology Centers. It will unlock $825 million in federal funds for research and development.

For brokers and building owners yet to recover from the pandemic, artificial intelligence investment is a godsend.

“New York has the second-largest number of AI companies after California,” said Rahul Mewawalla of Mawson Infrastructure Group, which builds and maintains the digital infrastructure for high-energy data centers that support AI, High Performance Computing (HPC), crypto mining and blockchain growth.

Over the last decade, AI firms have expanded exponentially in New York from less than 450,000 square feet to over 4.8 million square feet and counting. And although many AI firms are headquartered in Silicon Valley, they are finding the talent they want to lure in the Big Apple.

“New York attracts both graduates and the employee base,” added Jamie Katcher of JLL. “Almost 40% of relocators to New York are coming from the West Coast and the venture money pouring into New York is driving part of this surge.”

CBRE reports there are already nearly 190,000 high-tech jobs in Manhattan, with “more help” wanted ads posted each day. 

“AI needs access to a highly skilled workforce with a distinct set of skills apart from the broader tech industry,” explained Sacha Zarba of CBRE. “The fact that employees want to come to the office to work in New York is a huge driver for overall demand.”

Sam Altman’s OpenAI just leased 90,000 square feet at Kushner Companies’ Puck Building at 295 Lafayette St. That deal was repped by CBRE with asking rents ranging from $125 to $145 per foot. “OpenAI not only capitalized on the unique, 30,000-square-foot floorplate, which is rare in Soho, but on a building with characteristics that a young technology company craves,” Zarba said of space which had been occupied by NYU. 

San Francisco-based Harvey recently leased 17,000 square feet at 315 Park Ave. South as it broadens its focus from AI for legal services to all professional services companies. Meanwhile, the generative video and editing AI firm Captions leased 15,000 square feet at 71 Fifth Ave. in the Flatiron area near Union Square.

But Zarba believes the ability for these companies to scale up and grow in the same building is going to be key. That’s why AI startups are moving from submarkets with smaller floor plates like Soho, Union Square and Park Ave. South, to neighborhoods like Chelsea and Hudson Square where they will have room to grow.

London-based Genius Sports, for instance, is moving its local headquarters from the glass and steel institutional-like 825 Third Ave. to Chelsea and doubling in size to 22,454 square feet at 512 W. 22nd St. next to the High Line and near the sports fanatics at Chelsea Piers. The Vornado and Albanese-owned property’s last remaining space had an asking rent of $107 per foot.  

In Hudson Square, with help from the state’s Excelsior Jobs Program tax credits, ecommerce company Rokt agreed to expand by 34,000 square feet to 100,000 square feet at 175 Varick St. and will add an R&D hub to help other companies and workers.

Canadian AI company Cohere.com — with offices in Toronto, Palo Alto and London — opened a city space at 1 Little W. 12th St. in the Meatpacking District. That’s not to be confused with Cohere.io, with offices at 285 Fifth Ave., which was just acquired by Ramp at 28 W. 23rd St. 

With Amazon investing $8 billion into Anthropic, that San Francisco-based firm is expected to enlarge its Gotham presence out of a coworking site, while seeking dozens more “human engines” here. Similarly, Grammarly, a helpful writing tool with hubs in San Francisco, Kyiv, Seattle, Toronto and in NoMad at 29 W. 30th St., is on the prowl for more NYC elbow room, sources say.

“It’s no surprise these companies are sitting in the core of Midtown South,” said Katcher. “They want to have flexibility to expand and contract depending on where they are in the growth cycle. They are cautious, growth-wise, and signing short-term leases and not oversubscribed in their initial location.” 

In the meantime, building out energy resources to support AI is imperative, AI tenants often need 10 times more electrical capacity than traditional tenants because of their heavier computational draw — and they need it 24/7.

“You will need double the electricity you have now,” said Trump. “We will build better ones.” He is also telling companies: “Build your electric facility alongside your plant and people are loving that idea.”  

And right on schedule, DataBank will complete a 45 megawatt center in upstate Orangeburg that will link to its Manhattan locations at 60 Hudson and 111 Eighth — along with 165 Halsey St. in Newark.  

Mewawalla’s Mawson Infrastructure Group is also actively investing in and looking for former industrial sites, including steel plants and paper mills, because they used to consume a lot of energy and much of that power equipment and infrastructure remains.

“We need more support for data and dark fiber,” he said. “Energy will be critical moving forward. Robotics and all data users fundamentally will require more powerful sources of energy.”

Moving all that data also needs its own underground highways and companies “are buying fiber optic infrastructure at a terrific pace,” said CBRE broker John Needham, who specializes in the sale and leasing of rights of way, dark fiber and ducts across the US and Mexico.

The need for processing power and speed is also creating all kinds of construction jobs. One data line now being offered by Needham, for instance, has  “termination optionality at meet-me-rooms … or at mutually agreed-upon manholes in Jersey City and Lower Manhattan.”

Of course, behind the scenes, everyone has their fingers crossed that the growth of these companies won’t be a repeat of debacles like the dotcom bubble — when companies sold goods for more than they cost, took too much space and quickly burned through their venture money.

“We are very conscience of the upfront cost and term elements and all those variables so we don’t get into the dotcom situation where they were long on space,” said Katcher.