By Ray Sidler, CEO & Co-Founder, DataVerge
Picture a trading desk in Midtown. A radiologist at NYU Langone is pulling up a scan. A broadcaster in Midtown is cutting a live feed. Each of them, right now, is depending on an AI model to make a decision in real time.
And somewhere in Brooklyn, the infrastructure making that possible is humming along in a building most New Yorkers have never heard of.
New York’s AI Moment
I’ve been running a data center in Brooklyn for over two decades. In that time, I’ve watched New York slowly, then suddenly, become the most important city in the world for applied AI. Not for building AI. Silicon Valley still owns that. But for putting AI to work in ways that generate measurable business value, there is no city on earth better positioned than New York.
The reason is simple. New York’s greatest industries, such as finance, healthcare, and media, run on high-stakes, real-time decisions. They’re also concentrated within a few square miles of each other. That makes New York one of the most target-rich environments on earth for AI with real business impact. The AI companies gaining traction here are running inference for Wall Street, hospital networks, and media workflows.
In fact, the city is already home to more than 2,000 AI startups and 40,000 AI workers, the largest concentration in the US outside of the Bay Area. In June 2025, IBM opened its WatsonX AI Labs in Manhattan. And McKinsey projects that AI adoption could add 200,000 net jobs to the New York region by 2030.
The deployments are already underway. Banks and financial services providers are using AI to accelerate client onboarding, streamline lending workflows, and automate regulatory reporting. Many of New York’s major medical institutions, including NYU Langone, NewYork-Presbyterian, and Mount Sinai, are deploying AI for everything from early cancer detection to catching conditions in CT scans that would otherwise go undetected. And broadcasters are running AI inference against live video in real time, extracting context, talent, and audience signals to power targeting and personalization at scale.
As Cecilia Kushner, Chief Strategy Officer at the New York City Economic Development Corporation (NYCEDC) put it:
“California and Silicon Valley will always be the place where innovation in the technology itself is coming, but we think New York is where this innovation will be applied to real-world problems and real business opportunities.”
The Infrastructure Problem Nobody Is Talking About
From where I sit, New York’s AI ambitions have a hardware problem.
AI in production is a fundamentally different infrastructure challenge than AI in development. You can train an AI model anywhere with enough power and GPU clusters. But deploying one and running it in real time against live users and business systems is another matter entirely.
The numbers tell the story. Racks that once drew 30 to 40 kilowatts are now measured in hundreds. Cooling systems designed for enterprise IT are being asked to manage heat loads they were never engineered for. And that’s before you even get to the network. AI inference has to reach clouds, carriers, and data sources fast enough to be useful. In New York, where a delayed transaction, a missed diagnosis, or a dropped media stream carries real consequences, every layer of the stack has to perform.
Most existing facilities in the city were never designed with that in mind. And building enormous new campuses from scratch isn’t feasible in New York. The grid is too constrained, space is too limited, and the regulatory environment is too complex. What the city needs instead is smaller, denser, distributed edge infrastructure built close to the users, networks, and business systems that inference serves.
That’s the problem DataVerge was built to solve.
Brooklyn as the Edge
Manhattan has long served as New York’s primary connectivity hub, but its legacy carrier hotels were designed for a different era, one of enterprise IT, not sustained high-density AI inference.
Brooklyn offers what Manhattan increasingly cannot: physical flexibility, power capacity, and room to build the kind of dense, purpose-built compute environments that modern AI workloads demand. Combined with millisecond-level proximity to financial institutions, media companies, healthcare networks, and enterprise headquarters, the outer boroughs have become a serious part of how New York solves its AI infrastructure problem.
Walk into the DataVerge facility at Industry City in Sunset Park and you’re standing inside one of the most connected buildings in the northeastern United States. The campus itself is a 16-building complex housing more than 550 businesses, and it has undergone a $450 million redevelopment. Hundreds of businesses across technology, media, and creative industries operate there, connected by a dark fiber network with access to carriers including Verizon, AT&T, Lumen, Cogent, and Zayo, as well as direct connections to internet exchanges like DE-CIX and cloud on-ramps through Megaport.
We built DataVerge here because this is where the infrastructure math works. It’s minutes from Midtown by subway, with more power and physical flexibility than Manhattan’s legacy carrier hotels. The facility was purpose-built for the density and interconnection that AI inference demands, not retrofitted to accommodate it after the fact.
The workloads coming through our doors tell the story of where New York’s AI economy is heading: applications running continuously against live data, where the margin for delay is measured in milliseconds and the consequences of getting it wrong are immediate and real.
New York has the talent, the capital, and the use cases. If you’re building AI in this city, your infrastructure should be here. The infrastructure that makes it possible is already in place, and it’s closer to Midtown than most people realize.
