If you use generative AI, you know it can seem like magic. Chatbots and multimedia models can easily conjure up poems or high-definition videos with the click of a finger.
But the fast results and sleek interfaces of AI models hide the massive amount of physical infrastructure behind them, and as AI continues to grow, the data centers and power plants on which AI is built are beginning to attract widespread attention outside the industry.
Earlier this week I took the train to Orangeburg, New York, a quiet suburb of the lower Hudson Valley just 25 miles from the city. LuckNewsroom in midtown Manhattan. I was there to visit one of the waves of artificial intelligence infrastructure projects popping up across the country: Orangeburg is the future home of DataBank’s newest data center facility, called LGA3.
DataBank already operates two data centers in the New York metro area: one in Newark, New Jersey, and one in Chelsea, Manhattan. But LGA3 will by far be its largest facility, a $250 million, 200,000-square-foot facility that will consume up to 45 megawatts of energy to power five massive data centers crammed with computer chips.
The facility won’t open until next year, but tenants are already booked, most notably New Jersey-based artificial intelligence startup CoreWeave, which recently received a staggering $19 billion valuation and has already secured nearly half of LGA3’s capacity.
“The explosive growth of artificial intelligence has required a complete re-evaluation of traditional data centers to meet the demand for next-generation computing requirements, and this new data center campus provides some of the most advanced new technologies that will allow us to deliver to our customers,” he wrote in a note to me. CoreWeave founder Ben Venturo.
My taxi from the station made no fewer than four wrong turns as we drove past farmhouses and office parks toward the LGA3 construction site, wedged between an electrical substation and the New Jersey state line. Stepping out of the car, my first impression was the sheer size of the building. LGA3 looked the size of a New York City block: a massive single-story hall with high ceilings—I wouldn’t be surprised if it could fit a commercial jet inside.
“Addicted to Technology”
I was greeted at the entrance by DataBank CEO Raoul Martynek, wearing clear-framed glasses and a lavender button-down shirt. Martinek has been working in Internet infrastructure for decades, almost since the advent of the commercial Internet in the 1990s. He has been with Databank since 2017, overseeing the company’s 69 data centers in the US and UK. Martynek told me that he hasn’t seen an explosion in demand for digital infrastructure like that created by artificial intelligence since the dot-com bubble. 90s.
“People are addicted to technology, period. And ultimately, for the data center sector, we’re enabling people to use more technology,” Martynek told me. “If you deploy more technology, you need more fiber, more cell towers, more data centers. And at this particular stage that we are at with AI, the bottleneck is the data centers.”
Nathan Howard/Bloomberg
The huge surge in demand for AI has seen data centers making headlines. Martynek explained to me that most of the things we do online now—from accessing images on our phones to scrolling social media to running ChatGPT—involve physical hardware to a greater extent than we think. Wi-Fi routers and cell towers constantly send signals through underground fiber optic cables to data centers and remote servers, accessing stored information and keeping the Internet running smoothly.
“The Internet is a network, right? Information is transmitted through fiber optic cables in the form of photons. And they travel around the world at close to the speed of light,” Martynek said. “I was hanging out with a guy online last night and I was like, ‘What do you do?’ He said, ‘We’re plumbers, right?'”
And being a plumber is a good business these days. The exponential growth in the amount of data generated by and through the Internet over the past 20 years, and the expectation that AI will only speed things up even further, means that storage space for all this information is in high demand.
“This device didn’t exist until 2007,” Martynek told me, pointing to his iPhone. “So think about how much content and how many applications have been created [by it.] It all ends up in the data center… It’s a physical ecosystem.”
Courtesy of SourceCode Communications
AI increases the need for space for construction
Data centers may not be the most glamorous projects, but a surge in demand from artificial intelligence companies is bringing big money, widespread press coverage and the emergence of some of the biggest names in construction. DataBank alone has spent about $4.5 billion on data center projects since 2016. Tishman Speyer, the real estate company building LGA3, is one of the biggest names in the business: it worked on the World Trade Center and the John Hancock Center in Chicago, and it also owns Rockefeller Center. A low-rise data bank next door to a suburban Little League baseball complex may seem like an odd addition to its portfolio, but the company is betting that data centers will prove as important as skyscrapers.
When I first received an invitation to visit this place, I was surprised. NY? Does it have some of the highest real estate and energy prices in the country? Wouldn’t it be cheaper to build this? in the center a desert where land is cheaper and there is access to renewable energy at minimal cost?
But Martynek explained that for many clients, it’s simply not practical to be thousands of miles away from one of the most important parts of your business. New York City is one of the largest data center markets in the country, with about 800 megawatts of capacity currently online, much of it serving financial and technology companies that depend on nearby computing power to create and trade around the clock.
“It’s not practical for a data center to be in the middle of nowhere—there’s too much latency,” Martynek said, referring to delays in response times between computers and remote data centers. “Too many things can happen along the way.”
“Data centers tend to cluster around metropolitan areas,” he continued. “New York has always been a pretty big data center market. It really depends on the population and the business: If you’re JPMorgan, you don’t need a data center in Omaha.”
Government policy is also an important factor. New York Governor Kathy Hochul earlier this year introduced the government initiative Empire AIwhich has committed more than $400 million to finance, among other things, infrastructure such as data center projects.
Donning a hard hat and reflective vest, I walked around the unfinished structure with two construction managers. They pointed to an area the size of an airplane hangar where pods of computer chips will eventually be installed, along with ventilation and water cooling to prevent them from overheating.
Once construction is complete, other CoreWeave and DataBank customers will begin installing their chips, and DataBank expects the facility to be fully operational by early next year. Once CoreWeave is online, it will begin leasing its computing power to tech startups and other artificial intelligence companies. Martynek told me that DataBank had no problems finding clients.
“We signed a contract with CoreWeave last year. The building didn’t even exist back then—it was just land. That’s how in demand this product is,” Martynek said. “It’s crazy.”
As they say, strike while the iron is hot: DataBank is already drawing up plans for another facility next door, LGA4. Next time you’re driving around town, take a look at the tatty buildings in your neighborhood: the AI data center boom may be closer than you think.