How Data Center Acquisitions Change EBITDA Margins
Data center acquisitions are reshaping EBITDA margins—discover how this affects the industry landscape!
Acquiring a data center seems straightforward: buy capacity, gain revenue, and grow your balance sheet. However, what actually happens to margins is considerably more complicated — and the gap between expectation and reality is where deals are won or lost.
A recent Fitch analysis flags this dynamic. The rating agency's model explicitly accounts for a projected mid-2026 acquisition as a variable in its EBITDA margin forecast, alongside contributions from data center computing and defense/aerospace revenue. That's not a throwaway assumption. When a major credit rating agency builds an acquisition into its base case and treats it as a margin event — not just a revenue event — it signals something worth paying close attention to.
What Acquisitions Actually Do to EBITDA
The common assumption is that buying a revenue-generating data center is accretive from day one. Sometimes it is. More often, the acquired asset brings a cost structure that doesn't match the acquirer's, a customer base with different contract terms, and a capital expenditure backlog that the seller was deferring.
EBITDA margin isn't a static number you inherit — it's a ratio you have to rebuild.
Consider the mechanics. When a company acquires a data center mid-year, that facility's revenue gets consolidated into the parent's financials from the acquisition date forward. But the associated integration costs — new management overhead, systems migration, potential redundancies — hit immediately. Depreciation schedules reset. Power purchase agreements may need renegotiation. The result is a post-close margin that almost always looks worse than either entity's standalone performance before the deal closed.
Fitch's approach of modeling a "Fitch-adjusted EBITDA margin" is precisely because reported EBITDA post-acquisition can be misleading. Adjustments strip out one-time transaction costs, normalize the timing of acquired revenues, and account for synergies that haven't materialized yet. For infrastructure investors evaluating a target's credit profile, these adjustments aren't accounting gymnastics — they're the difference between understanding what you actually bought and believing a pro forma.
The Revenue Timing Problem Nobody Talks About
One of the least-discussed complications in data center M&A is revenue recognition timing. Hyperscale tenants — the Microsofts, Amazons, and Googles of the world — sign long-term leases that often include rent abatement periods, stepped rents, and tenant improvement allowances. An acquired facility might be 90% leased on paper while generating 60% of stabilized revenue because half the leases are still in their abatement window.
This creates a scenario where the acquirer consolidates the asset at below-run-rate revenue while carrying full operating costs. Margins compress. Analysts downgrade the combined entity's near-term outlook. The deal looks worse than it is — temporarily.
The investors who profit from data center acquisitions are almost always the ones with the patience and liquidity to weather 12 to 24 months of margin drag before stabilized economics appear.
That timeline matters enormously for infrastructure investment underwriting. A deal modeled on trailing EBITDA multiples without adjusting for abatement burn-off will systematically overprice stabilized assets and underprice value-add ones. It's one of the more persistent mispricings in the sector.
What's Driving the Acquisition Wave
Demand for data center capacity is structural, not cyclical. AI inference workloads require fundamentally different compute density than traditional enterprise IT — more power per rack, more cooling per square foot, more redundancy. Building greenfield facilities that meet these specs takes two to four years in most major markets, assuming you can secure the power interconnect, which increasingly you cannot.
Acquisition is faster. Buying an existing facility with permitted power capacity, established fiber routes, and an existing customer base compresses that timeline dramatically. That's why cap rates on stabilized, well-located data centers have compressed to levels that make traditional real estate investors wince — sub-5% in primary markets — while transaction volume keeps climbing.
The competitive dynamics amplify this. Hyperscalers are building their own capacity at scale, which means third-party colocation and wholesale operators face a shrinking addressable market in core enterprise IT. The response has been consolidation: acquire the platforms, the customer relationships, and the land positions before competitors do.
Defense and aerospace — specifically the A&D revenues referenced in the Fitch analysis — add another dimension. Government and defense-adjacent workloads increasingly require physically separate, security-cleared facilities. Operators with cleared facilities command premium pricing and face virtually no hyperscaler competition. An acquisition that brings DoD-adjacent revenue streams into a portfolio meaningfully changes the risk profile of that EBITDA, not just its magnitude.
The Margin Math in Practice
Walk through a simplified version of what Fitch is likely modeling. Assume an operator generates $500M in revenue at a 30% EBITDA margin — $150M EBITDA — before a mid-2026 acquisition. The acquired facility adds $80M in annualized revenue but is only generating $60M in the stub period post-close. Integration costs run $12M in year one. The acquired asset carries a 22% standalone EBITDA margin (it's a value-add play, not stabilized).
Combined revenue in the acquisition year: roughly $540M. Combined reported EBITDA: somewhere around $155M before integration costs, $143M after. Reported margin: ~26.5%, down from 30%. On paper, the deal destroyed margin. Fitch's adjusted view would normalize the revenue to a $580M run-rate, strip integration costs as one-time, and project stabilized EBITDA at $168M — a 29% margin trending back toward the acquirer's baseline as synergies land.
That spread between reported and adjusted is where analysts earn their pay. It's also where opportunistic buyers find their edge — acquiring assets that look margin-dilutive at close but normalize favorably 18 months out.
What the Next Few Years Look Like
The acquisition pipeline in data centers isn't slowing. Power constraints in Northern Virginia, Silicon Valley, and the Chicago corridor are pushing development to secondary markets — Phoenix, San Antonio, Columbus — where land is available but operational expertise is scarce. That creates a natural acquisition opportunity for scaled operators who can buy smaller regional platforms, bring operating leverage, and improve margins through procurement scale and management efficiency.
The risk is overpaying. Cap rate compression has been so severe that even modest interest rate sensitivity creates negative leverage in acquisition models. Operators who assumed a financing cost of 5% and are now looking at 6.5%+ on acquisition debt face a structurally different return profile than their original underwriting suggested.
The Fitch framing is instructive here: treating an acquisition as a margin event, not just a revenue event, is the right analytical lens. Infrastructure investment in this sector rewards those who understand that buying capacity is easy — integrating it, stabilizing it, and growing EBITDA margins through the noise of post-close disruption is the actual job.
The deals that look transformative in press releases often look painful in the first two quarterly earnings calls. The ones that look like disciplined, boring infrastructure plays tend to be the ones still generating returns five years later.