The Next Phase of Data Center Infrastructure

For the past several years, building a data center has meant competing for a place in line for the power, cooling, and mechanical infrastructure that help to keep the data center operational. Moody's now projects hyperscaler capital spending will reach $785 billion in 2026, approaching $1 trillion by 2027 (Moody's Ratings, 2026), yet estimates of how much planned 2026 capacity will be delayed or canceled for a variety of issues including a lack of transformers, switchgear, chillers, generators, and other critical infrastructure equipment range from roughly a third to nearly half of new capacity (Sightline Climate, cited in Bloomberg, 2026; ChargedUp!, 2026). For the manufacturers supplying that equipment, across power, cooling, and HVAC alike, the result has been the best pricing environment in a generation, but it will not last.

Scarcity-driven markets like this one always rebalance eventually, and the data center OEM suppliers that treat today's conditions as permanent will be poorly positioned for what comes next. The suppliers that use this window to identify which product lines are headed for commoditization, and which still command differentiated value, will be the ones setting the pace once the market rebalances.

The Scarcity Window Will Close

Power transformers offer the clearest illustration of a pattern playing out across the supporting infrastructure stack. Lead times for high-voltage power transformers, which ran 24 to 30 months before 2020, now stretch to as long as five years (Sightline Climate, cited in Bloomberg, 2026), and demand for generator step-up transformers has grown 274% since 2019 (Wood Mackenzie, cited in Power Magazine and IndustrialSage, 2026). Cooling and HVAC equipment face comparable constraints as data centers move to liquid cooling and higher-density thermal management, and generator and switchgear lead times have stretched in parallel. That imbalance is why buyers currently compete for allocation rather than suppliers competing for customers, and why price has stopped being the deciding factor in most purchase decisions across these categories.

As new manufacturing capacity comes online across these categories, the buyer relationship will change faster than most suppliers expect, because buyer behavior shifts as soon as alternatives emerge, not after the market has fully rebalanced. The question a data center developer asks will move from “can you deliver” to “why should I buy from you specifically,” and total cost of ownership, service quality, reliability, and upgradeability will move from secondary considerations to primary decision criteria. Suppliers who wait for unambiguous signs of oversupply before adjusting their commercial approach will be reacting to a shift their customers made months earlier and will potentially place them at a disadvantage going forward.

From Permanent Assets to Planned Refresh Cycles

The bigger structural change goes beyond buyer behavior, reaching into how the equipment itself gets bought and replaced, as compute density, cooling requirements, and power architectures evolve fast enough that the electrical and mechanical systems supporting data centers, transformers, and switchgear as much as chillers and cooling towers, will likely be upgraded on a much shorter cycle than the buildings that house them. That shorter upgrade cycle reflects a change in procurement philosophy. In conversations with data center operators and developers, BCE has increasingly heard the same expectation: infrastructure will move toward a rip-and-replace cycle of roughly five to seven years, shifting the mix of future work away from greenfield construction and toward planned replacement. Under that model, equipment no longer needs a thirty-year lifespan to be a sound purchase. Customers are looking for equipment that performs with total reliability for the length of that shorter cycle, not equipment engineered to outlast it. They are buying performance through the next planned modernization, at which point replacement happens by design rather than by failure. This shift will lead to a greater reliance on price-competitive products that are more commoditized than the ones that are required today.

Repositioning for a Post-Scarcity Market

Competing in a mature, post-scarcity version of this market will require change across four areas, regardless of product category:

1. Product design: Build for modularity and fast replacement, not maximum lifespan. Equipment engineered to last thirty years untouched, whether a transformer or a chiller, is the wrong asset for a customer planning to reconfigure power and cooling every five to seven years. That same flexibility needs to anticipate where computing architectures are headed, since equipment specified for today's power density and cooling requirements may already be undersized for the next generation of compute.
2. Manufacturing: Prioritize cost position and inventory flexibility now, before price sensitivity returns. Once availability stops being the constraint that wins deals, manufacturers running scarcity-era cost structures will quickly lose share.
3. Commercial strategy: Future competition requires that vendors compete on lifecycle value, service, and speed. Once multiple suppliers can deliver on time, suppliers can no longer differentiate on speed, and customers will adapt their considerations to align with key operational considerations.
4. Service model: As the market shifts from greenfield construction to planned upgrades, the suppliers with a service organization built to sell recurring replacement work will capture that revenue. The suppliers without one will watch it go to a competitor.

The OEM suppliers that get ahead of these four areas will protect margins as scarcity pricing fades. Suppliers that do not will face specific and predictable risks as commoditized product lines will see margin compression first, sales cycles will lengthen as buyers start comparison shopping on total cost of ownership, and revenue that once came from allocation-driven greenfield orders will not be replaced automatically by modernization revenue unless the commercial organization is built to sell it. The combination of margin dilution in the product lines most exposed to commoditization and a sales motion still calibrated for a scarcity market is where failing to prepare for this transition will do the most damage.

Where BCE Can Help

The hardest part of this transition is not recognizing that it is coming. It is knowing, component by component and across power, cooling, and mechanical infrastructure alike, which product lines are closest to commoditization and which still command a defensible position, while also building the commercial and product roadmap around that answer before competitors do. BCE works with equipment manufacturers to map that exposure, quantify the margin risk under a normalized-supply scenario, and design the transition plan needed to compete on lifecycle value rather than availability. The shift from scarcity to scale is coming either way. The only open question is whether your organization meets it with a plan or gets caught unprepared.