Ivo Ivanov is the CEO of Overflow.
At first glance, New York has everything. Largely considered to be the “capital of the company”, the city has now overshadowed London and Singapore to become the most sought -after financial center in the world. It has a booming talent pool, an exceptionally strong technological sector and a pro-business mentality which attracts not only large companies but also countless small businesses and startups. In 2024, the New York City Economic Development Corporation (NYCEDC) has announced a record number of small businesses, with 1 in 3 of 183,000 small businesses that opened in the past two years. Make no mistake, New York is the very definition of a business center.
However, despite its great success as Center for Business, there is a very important problem threatening its momentum – while companies have found physical or digital houses in New York, many now find it difficult to find a house for their data.
While we enter in 2025, there is a separate lack of data center capacity in New York. According to the latest figures available, the vacation rates of the data center in the three states area have reached A hollow of all time At only 6.5%, creating a headache for growing businesses. The obvious solution is to build more data centers, but New York does not have physical space to make it a reality. Consequently, existing buildings are reused to house data centers, but the energy capacity remains rare, leaving New York to train for other major American cities in terms of deployment of the data center. While the main business centers like Dallas have experienced a 150% increase In the inventory of the data center in the past decade, New York has increased only 4%. Many multi-local data center installations in New York and elsewhere also become unique installations, and many others now exclusively house companies and large-scale hyperscalers. While companies turn their attention to AI and heavy data workloads, this lack of provision of data centers will cause serious problems. So what’s the answer?
One is the deployment of hybrid and multicloud strategies to “distribute the risk” and reduce their dependence on individual physical data centers. This allows companies to distribute workloads between on -site infrastructure, roommate and cloud suppliers like AWS, Google Cloud and Microsoft Azure. EDGE computer nodes are another option, using smaller micro data centers for specific AI workloads. Another option is that companies associate with a roommate installation for a personalized construction, which allows them to deploy their own infrastructure; However, this can be expensive. For many companies – in particular startups or those with limited resources – another path is to look for options for alternative data centers outside the immediate New York region, perhaps long Island or Brooklyn, but any relocation naturally creates concerns about latency and ensuring high performance access to their “domestic networks”, New York partners and customers.
Why geography is important
The relocation of the workloads of the data center can – and often – resulting in significant time that few companies can afford. Although we accustom a digital world, physical distance also contributes to the degradation of the service, which is an Athema to companies that require data processing in real time and low latency connectivity. When companies can use cross -checking and peer networks in their city to eliminate all performance problems, their options when switching to a peripheral data center tend to be more limited. This means that instead of looking directly, they may have to count on a good old -fashioned IP transit (public internet) to transport their data. Needless to say, this creates security problems and leaves businesses without control over their connectivity – Data traffic routes, network hops, performance and latency are simply left at random.
What options are open to businesses?
First, companies can take measures to optimize their existing infrastructure. Approaches such as virtualization and containerization can help make workloads thinner and reduce the imprint of the physical server. A multicloud strategy and the adoption of a “cloud bursting” approach – where private clouds can use public cloud resources when the demand for calculation is high – can also help to mitigate the overcrowding of the data center. However, although some of these options can delete the pressure of the workloads of the individual data center, they do not take up the basic challenge – companies must either invest massively in the rationalization of their operations to reduce their footprint or rely on public connectivity channels for high calculation charges. And with regard to IA inference, coherent and high performance connectivity is essential.
Earlier, I mentioned that the move of workloads in data centers outside the New York region is a viable strategy. However, instead of only counting on traditional IP transit, interconnection in the form of data centers and neutral internet exchanges (IX) are the preferred choice, in particular when companies are increasingly going towards workloads that use latency that uses generative AI.
IXS are data centers hubs where several networks can connect and exchange traffic directly to improve speed, reduce latency and optimize data traffic flow. These direct peering capacities, especially when distributed in the installations of different operators of data centers, can allow cloud connectivity from a variety of data centers, no matter where they are. For example, a company could settle in New Jersey or Long Island and – if it joins an IX compatible data center – it always has quick access to data centers in the heart of New York. Since the connections can be maintained both in the old and new data centers using the same IX, there are few or no interruptions during the move because traffic can move to the new location once operational.
What is the next step for the Big Apple?
As the data center tightens tighter in New York, companies meet at a crossroads: moving, optimizing or risking being exceeded by competitors. For companies with resources to do so, refine their workloads and look at a multicloud solution may be sufficient. For others, the distributed NEUTRE IX model makes relocating a much more accessible option, which allows them to strengthen their connectivity with a larger choice of providers, improved scalability and less risky migration process.
Forbes Technology Council is a community only at the invitation for CIOs, CTOs and world -class technology executives. Am I qualified?