Where to play in the increasingly diverse data centre economy

In the recent past, the types of companies involved in data centres looked familiar: technology vendors, operators, and specialty construction. But that’s no longer the full story. Previously unpublished data from PwC’s AI performance study reveals that more than one-third of companies across 25 sectors already participate in the supply side of the data centre economy. A further 11% expect to soon. Of these participants, three-quarters come from sectors outside of technology. 

The pull is obvious: PwC expects some US$5.1 trillion in capital expenditures on data centres globally by the end of 2030. However, broad participation in pursuing a share of this massive market shouldn’t be confused with durable advantage. The buildout will be shaped less by who wants to participate and more by who can solve the constraints that determine what actually gets built, financed, and used. 

Those constraints cluster in a few familiar places. One is electricity: deliverable power, grid interconnection, backup generation, and increasingly scarce equipment such as transformers and turbines. U.S. data centres already consumed 176 TWh in 2023, and global data centre electricity use could more than double by 2030. A second is the license to build at all: permitting, water, emissions, zoning, and community acceptance. Yet another is what customers and regulators now require once a facility is running: demonstrable resilience, security, auditability, and, increasingly, a credible data sovereignty posture. In PwC's study, sovereignty and residency rank as a top buyer priority for 44% of end users—nearly tied with total cost of compute (46%).

The upshot: a company can have exposure to the buildout without a clear right to win. It may own land but lack power. It may provide capital but inherit permitting and interconnection risk. It may sell a product into the value chain without gaining pricing power or repeatable demand.

Looking at the market through the lens of constraints also clarifies why the most obvious entry point may not be the safest. The largest share of prospective participants in our study expect to enter through financial investment or ownership. The appeal is understandable. Data centres can look like infrastructure assets with long-duration demand and stable cash flows. But financing may be the hardest place to play passively. If power slips, permits are contested, equipment is scarce, or sovereignty requirements tighten, the underwriting case changes—financing decisions must track these issues.

For leadership teams, the challenge today is to determine whether data centres can be a credible extension of the business you already have—or a new business worth investing in—and what kind of moves may be justified. Here are three ways to consider the issue:

Identify your best opening

Start with the source of advantage. Power access, grid relationships, land, fibre, cooling technology, engineering depth, long-duration capital, local operating credibility—the strongest positions come from solving a bottleneck others can’t easily work around.  

Decide what to own and what to partner for

The right shape of a move depends on where you start. A real estate player will usually need utility, cloud, or operating partners. An industrial manufacturer is often better positioned as a critical supplier than a platform owner. A financial investor may need a deeper power-market diligence than a traditional infrastructure deal requires.

Build for repeatability and speed to market

One site, one customer, or one attractive deal may generate returns. A stronger position is one that can be repeated across a portfolio, a customer base, or a capability set. That means being able to quickly keep meeting the electricity, permitting, and assurance requirements as they evolve.