Building a Business that Outlasts the Data Center Boom

September 3, 2026

We think the data center buildout has a long runway left, well into the 2030s, by most reasonable accounts. Power constraints, permitting, and grid buildout mean this isn't a market that spikes and disappears; it's more likely to run hot for years and then settle into something closer to a normal industry cadence. The scale involved is real: the four largest hyperscalers, Microsoft, Alphabet, Amazon, and Meta, are guiding to a combined $720 billion to $745 billion in 2026 capital spending, up roughly 77% from about $410 billion in 2025, based on their own first-quarter 2026 earnings reports. Spending at that scale doesn't get built out in a single year; a meaningful share of the capacity planned for 2026 is already sliding into 2028 because of power and grid bottlenecks, which tells you this is a multi-year story, not a next-quarter one. And even after growth eventually settles into a more traditional pace, data centers aren't going anywhere. They'll need replacement parts, upgrades, retrofits, and new capacity in specific markets indefinitely. This isn't a story about an ending. It's a story about what a smart business owner does with the years in between.

Here's the structural point we'd want every owner, and every adviser to an owner, to sit with: the capacity being built right now (the plant, the equipment, the workforce, the supplier relationships) is scaling to a historically unusual level of demand. That's fine while the demand is there. But growth rates by definition come back down to earth eventually, even in a market that never disappears. When that happens, all the manufacturing capacity, all the crews, all the shop-floor hours that got built up to serve peak data center demand need somewhere else to go. Call it backfill. The businesses that already have somewhere else to send that capacity will barely notice the transition. The ones that don't will spend a very uncomfortable year or two figuring it out in real time.

The good news is that backfill isn't something you scramble to build after the fact. It's something you build in parallel, starting now, while the data center business is still strong enough to fund it.

Build capability that flexes, not capability that's locked in

Not every dollar you spend building out for data center demand needs to be single-purpose. If you're investing in new equipment, cross-training your workforce, or standing up new production lines to meet data center backlogs, ask a second question before you commit: could this same line, this same crew, this same certification also serve a different customer? A machining cell tooled to make only one highly specific part, built exactly to one customer's specifications, is a data-center asset. The same cell, tooled with enough flexibility to also produce parts for grid equipment or industrial machinery, is a business asset. The capital cost is often similar. The optionality is not.

Build the pipeline in adjacent markets before you need it

The core capability most data center suppliers have built, power density, redundancy, precision cooling, fast-turn fabrication, and EPC (engineering, procurement, and construction) execution, is exactly what's needed in a handful of other markets investing heavily right now: grid and utility modernization, industrial and manufacturing reshoring, hospital and healthcare facility expansion, and EV and battery/energy-storage buildout. These aren't hypothetical someday markets. They're active today, often with the same specification requirements and many of the same engineering firms involved. The mistake is treating this as a project for whenever data center orders slow down. By then, you're starting a sales cycle from zero, against competitors who started years earlier. The businesses that will handle the transition well are building quoting relationships and getting qualified as a vendor in these adjacent markets right now, even if the volume today is small relative to data center revenue.

Build the annuity on top of everything you've already installed

Here's the part that should make this whole conversation less challenging, not more: every generator, switchgear panel, cooling unit, and mile of structured cabling that goes into a data center today becomes a piece of equipment that needs maintenance, monitoring, spare parts, and eventual replacement for the next 15 to 20 years. New-build volume may normalize, but the installed base only grows. A supplier who builds out service capability (maintenance contracts, retrofit and upgrade programs, monitoring, parts supply) alongside their new-build business is setting up a second revenue stream that doesn't depend on the pace of new construction at all. It depends on the size of the fleet already in the ground, which only goes up. This is, in a sense, the most natural backfill there is: the boom you're living through today is quietly building your replacement-and-upgrade business for the 2030s and beyond.

Diversify inside the data center world too

None of this requires walking away from data centers, and it shouldn't. But even within the sector, there's a difference between a business that depends on three or four hyperscalers' new-build capex and one that also serves colocation operators (companies that rent out shared data center space to multiple tenants), enterprise data centers (facilities a company builds and runs for its own internal use rather than to sell to outside customers), edge deployments (smaller sites placed close to where the data is actually used, to cut latency), and international buyers on different investment cycles. Spreading across customer types and geographies inside the same end-market is a lower-effort version of the same diversification logic, and it's worth doing alongside the moves above.

Why this matters at the closing table, not just on the shop floor

We spend most of our time on the M&A side of businesses like this, and we can tell you plainly how this shows up in a sale process or a recapitalization. A business with 90% of revenue tied to new-build data center capex, no service book, and no presence in adjacent markets gets priced and discussed like a project business: good today, uncertain tomorrow, and buyers will build that uncertainty into the multiple whether you agree with it or not. A business with the same core capability, a growing service and parts business, and real traction in two or three adjacent end-markets gets priced like a platform: durable, diversified, worth a premium. The work of building that second business doesn't just prepare you for a slower-growth decade. Dollar for dollar, it's also some of the highest-return work an owner can do before a sale.

The data center boom is a genuine, multi-year opportunity, and we'd tell any client to lean into it hard. We'd also tell them that the years when this market is strongest are exactly the years to fund the parts of the business that will matter most once it normalizes. That's not hedging against the good times. It's what makes the good times worth something ten years from now.

If you're an owner, or advising an owner, in this position and want to think through what a durable, diversified business actually looks like for their specific company, we are always glad to have that conversation.