What Makes a Data Center Developer Competitive in 2025?

Being good at everything has all but topped the ‘pipe dream’ list since time immemorial.

Yet, as demand skyrockets, requirements become more sophisticated and resource competition grows fiercer, developers must continue pursuing all-around optimization. Our designs must be hyper-efficient, our supply chains airtight, capacity pipelines always stocked, design and construction practices near faultless — and that’s just the beginning.

Does a developer really need to be a Renaissance man to succeed? In some ways, yes. We must deliver the best solutions and strategies for every goal and have rock solid contingencies when something doesn’t go to plan. But ideal data center deployments aren’t born perfect — they’re cultivated by years of experience and honed by smart people that know how to pair that experience with technical innovation (and a lot of trust and collaboration).

So, as data center demands evolve in 2025, here’s how we’re pursuing internal innovation and ecosystem enrichment to deliver great data center solutions and experiences.

Keeping Your Cool (Efficiently) Amid an AI Explosion

AI is hot — both literally and figuratively. As one expert states in a recent Fast Mode article, “The latest AI server racks cram the heat of 16 gas barbecue grills into the space of a phone booth.”

While the advanced, specialized hardware used for today’s intelligent workloads are increasing heat loads in data centers, direct liquid cooling (DLC) options are growing in relevance. From a sustainability perspective, power-intensive AI workloads are creating efficiency gains thanks to the liquid cooling improvements they’re driving.

It’s well understood that liquid far outperforms air in its ability to transport heat, but with the right IT configuration, cooling fluid temperatures can also be raised. This affords operators the ability to use economization for more hours a year and can help reduce the energy needed for the mechanical part of a data center’s operations.

Plus, as servers are redesigned for liquid cooling and onboard server fans get removed or reduced in quantity, more of the critical power delivered to the server is being used for compute. This means that liquid cooling is also driving an improvement in overall compute productivity even if it’s not noted in facility PUE metrics.

Counter to air cooling, liquid cooling certainly has some added management challenges related to fluid cleanliness, concurrent maintainability and resiliency/redundancy, but once those are accounted for, the clusters become stable, efficient and more sustainable with improved overall productivity.

Still, supporting AI and moving toward liquid cooling can create other kinds of efficiency gaps. For many, evolving to suit new AI applications could mean stranding existing air cooling resources in the jump to liquid to the rack options. For others, traditional workloads aren’t going anywhere, which means they will need both cooling mediums and will want to support the flexibility to deploy either type as the IT stack destined for that space demands.

To smooth the transition between mediums, reduce resource stranding for greater efficiency and offer flexibility for customers that need both air and liquid cooling (or want to adjust their methods in phases over a longer term), the answer is modularity.

For the past two years, in response to demands from some of the most sophisticated data center users, Stream has been developing a configurable air and liquid cooling system that can handle the highest densities in both mediums. It’s a scalable, mix-n-match front-end thermal solution that gives us the ability to late bind the equipment we need to meet our customers’ changing cooling needs.

To learn more about how Stream’s data center operations experts have crafted a solution capable of meeting evolving requirements, check out the Stream Thermal Unit (STU).

Necessity is the Mother of Invention Exploration

You can’t have a data center if you don’t have grounds for it (in the truest sense of the word). As more real estate is absorbed rapidly, site selection processes must become more thorough as data center developers search for available capacity and other must-haves.

It can be difficult to know where new load can be served, but developers are flocking to areas where power is still available. For example, we’re now seeing hyperscalers expanding beyond Northern Virgina into central/rural Virginia, and some hyperscalers have entered non-traditional data center markets in Mississippi or Indiana, among other rural areas in the central U.S. As technology journalists from Data Center Frontier report, both AWS and Meta have planned major investments in those markets, respectively.

To make the best of a crowded market, you need to be more than smart and process-driven in your decision making — you need to be creative and exploratory. Having a great location strategy team and a proprietary GIS tool makes the haystack easier to sift through and the needle easier to spot, so we’ve spent a lot of time finding the best method that is now supporting us (and our customers) well in our development efforts.

Going forward, everyone – customers, developers, investors – should expect timelines to be a bit more drawn out as space becomes more difficult to find. But with the right methods, due diligence processes and in-market relationships, we can be more discerning while shortening our time to the right site.

As our site selection and strategy location experts say, success depends on our ability to reduce the cone of uncertainty.

Partnerships: Playing (and Winning) The Long Game

In complex data center developments, the ‘it takes a village’ idea rings truer than ever. Construction costs are rising, supply chains are still challenged, demands are shifting and growing — and data centers take years to construct, leaving lots of time for the rules to change mid-game. Business models, designs and procurement processes need to provide structure without compromising flexibility. To bend without breaking, we need trust and collaboration on all sides.

Having long-term, deep partnerships with industry-leading vendors, focusing on owner-furnished, contractor-installed (OFCI) processes and having standardized-but-configurable designs adds both flexibility and replicability to data center development. With these processes and partnerships in place, long-lead equipment timelines can be insulated against change while giving customers added time to decide what they really need (and still offering precise, optimized results).

Still, while data center developers and customers are deciding what’s real for their needs, utilities are doing the same. This is why partnerships with utilities and an awareness of how power demands are changing the landscape of development are equally as important for avoiding challenges and delays.

Recently, Scientific American published findings from the International Energy Agency, noting, “Countries are building power plants and upgrading electricity grids to meet the forecasted energy demand for data centres. But the IEA estimates that 20% of planned centres could face delays being connected to the grid.”

To help utilities identify real demand and navigate crowded queues, developers and customers must often cover the upfront costs of power interconnection while also committing to minimum power usage fees for 10+ years. This will require significant capital and long-term commitment, and it’s going to become harder to buy down lead time and stay in the interconnection queue while working with customers to land a lease and get their commitment to the power ramp.

Establishing energy security long-term now means learning how we can be a good partner to utilities while creating clarity around when there’s real demand. What’s more, it’s also becoming clear that deep pockets, strong financial partners and firms that are able and willing to take some financial risks to secure power will win out.

But more importantly, these shifting market requirements and heightened levels of competition underline the importance of trust, transparency and a strong track record of development. These assets will be crucial for data center providers that want to keep (and deepen) their relationships with investors, customers and utilities.

Ultimately, no one person can be good at everything — but an organization full of smart, passionate people building strong processes while strengthening their trust-based partner ecosystem? Well, that’s where our team of data center pros proves the truth is in the track record.

Insights shared by Michael Lahoud, Stream’s Co-Managing Partner, as part of Data Center Frontier’s Q1 2025 Executive Roundtable.