A new customer has signed on to lease up to 3.8 megawatts of capacity at the Hudson Interxchange data center at 60 Hudson Street in New York City, according to Data Center Dynamics. The customer is described as an AI company. It’s a modest deal by industry standards, but it’s a useful window into how AI demand is reshaping even the oldest corners of America’s digital infrastructure.
Why an old telecom building matters
60 Hudson is one of the most storied buildings in American telecommunications — a former telegraph and phone company hub that became one of the internet’s key connection points decades ago. Buildings like this matter because they sit at the crossroads of fiber networks, making them attractive to companies that need fast, reliable connections. When an AI company leases space there, it’s often chasing that same connectivity, not just raw power.
What 3.8 megawatts actually means
To put that in perspective, a few megawatts can power thousands of homes. AI workloads, especially the computing behind large language models and image generators, run on servers that draw far more electricity per rack than traditional office computing or basic web hosting. That’s why even small-sounding leases like this one add up when you consider how many are being signed across the country right now, often invisibly to the public until utility bills or power negotiations show it in local records.
The bigger accountability question
Deals like this one rarely get public debate the way a new suburban data-center campus might, because they happen inside existing buildings in dense cities rather than on greenfield land. But the power draw is real either way. Every megawatt leased to an AI tenant is a megawatt of demand utilities and grid operators have to plan for — and ultimately, ratepayers may help finance that capacity through infrastructure upgrades, whether or not they ever use an AI product themselves.
What you can do
Most people won’t have a say in a single lease inside a private building. But the broader buildout is shaped by decisions people can weigh in on: utility rate cases, local zoning hearings, and state energy planning processes all shape who pays for the grid upgrades AI demand requires. Check GridWatch’s interactive map to see what’s happening near you, browse our explainers on how data centers connect to your power bill, and visit Take Action for ways to engage with the public processes that actually do have room for comment.
Source: Data Center Dynamics