Data centers craving food and service

Perspectives from BofA Global Research’s Leading Analysts

 

July 20, 2026

Head shot of Curtis Nagle

Curtis Nagle, Senior Research Analyst, Business & Information Services

It takes a (small) city to feed the hyperscaler buildout

We see the global food and facilities service providers as an under-the-radar winner from the rapid expansion of data centers driven by unrelenting demand for AI compute and workloads.

By 2030, data-center capacity is expected to double to 200GW, supported by an estimated $7tn in capital investments, potentially bringing thousands of new data centers online. Much of this demand is being fueled by the hyperscalers, who are expected to increase investments by ~70% y/y in 2026, highlighting the need to bring capacity online with high immediacy. Compared to build-outs over the past several years, many hyperscaler data centers are significantly larger in capacity (with some planned campuses in the several-gigawatt range), require immense construction workforces, are located in rural locations that are difficult to serve and can span years to complete. This presents opportunities for the global service providers who have the scale and expertise to serve large, operational complex and remote worksites.

A recipe for a nearly $100bn market in the making

Data-center construction is highly labor-intensive, with large-scale projects requiring up to 10,000 workers. While infrastructure constraints ultimately dictate the pace of deployment, workforce dynamics are becoming an increasingly important layer of execution as projects scale in size, duration and geographic dispersion. Large campuses require thousands of workers across multiple shifts, often in remote or industrial locations with limited access to food, housing and basic amenities, which can create additional bottlenecks if not proactively managed. As hyperscalers and contractors look to maintain productivity across multi‑year build cycles, we see a growing recognition that the workforce experience, which includes food services, accommodation, transportation and on‑site amenities, can support more stable and efficient site execution over time and, potentially, at a lower cost. This is where integrated workforce and facilities service solutions come into play.

Based on projected data-center capacity needs, we believe this could represent a new market opportunity of nearly $70bn in the U.S.

In the post build-out, operational phase that requires significantly smaller workforces, we still see an estimated $20bn market and upside to both estimates from opportunities outside of the U.S.  

Double-digit earnings upside opportunity for scaled providers

The data-center services ecosystem today remains highly fragmented, with most operators relying on a mix of specialized providers across food, housing, transportation and facilities management. This fragmented delivery model becomes increasingly complex as data-center construction evolves into a multi‑year, multi‑phase process. Coordinating separate providers for dining, housing, logistics and amenities adds operational burden at a time when contractors are focused on managing timelines, safety and technical execution. We see a growing need for simplified, integrated service models that can reduce coordination complexity and provide more consistent workforce support across the build cycle. We see the global food and facilities service providers as uniquely well positioned to capture this growth opportunity given scale and decades of experience in operating facilities services in remote and logistically challenging sites such as mining camps or offshore drilling rigs. While it’s still very early days, the first (of what we expect to be many) integrated services contract for a hyperscaler data-center build-out is in the low hundreds-of-millions-of-dollars range. On relatively modest market share assumptions, we believe data centers could represent an opportunity for a double-digit earnings upside for the global food and facilities service providers.

Labor constraints, supply chain & NIMBYism are risks

We see risks as largely based on supply and logistical constraints. As an example, permitting timelines (typically 2–3 years, but up to 7–10 years with bottlenecks), particularly for grid interconnection and transmission infrastructure, can extend project start dates. Labor dynamics are also a material risk given that data centers are increasing demand for specialized technicians and electricians, in the context of already high structurally driven shortages of industrial workers. In parallel, supply-chain bottlenecks, including long lead times for critical equipment such as transformers, may further delay construction schedules. Finally, the broader data center build-out faces increasing scrutiny around resource intensity, particularly related to power and water usage, which has led to rising political and community opposition in certain regions. Measures such as moratoriums, including the first one signed in New York state a year ago, could slow the pace of development in key markets and are likely to be a key swing issue for many states going into midterm elections.

bofa global research

High stock in style for off-price retail

In light of strong comp momentum and gains in apparel retail stocks, we see limited risk to building inventory for off-price retailers.