Data centre PERE deals are equity investments where the value sits in real estate and contracted cash flows from data-centre property, even when an operator helps run the site. Enterprise value (EV) is the full value of a business – equity plus net debt – so it lets you compare deals that use different capital stacks. A “platform” is a set of assets, people, contracts, and development rights that can compound by adding capacity, not merely by collecting rent.
When people ask for the top data centre PERE deals in 2025, they usually mean “biggest.” I read that as EV or an implied platform valuation where it’s disclosed. If the market only quotes an equity check and stays quiet on debt, you’re not looking at a comparable number. In 2025, many announcements were selective, so the useful work is understanding structure: who took what risk, when cash goes in, and what has to go right to earn the return.
Scope: what counted as data center PERE in 2025
In 2025, the investable perimeter included stabilized colocation (retail and wholesale), powered shells with pre-leases, and land-plus-power positions that were de-risked enough to finance. Real estate still anchored the thesis because buyers could underwrite title, lease-like income, and security packages that lenders can enforce.
It did not include venture-stage compute businesses, chip or equipment makers, or most cloud service providers. Those businesses sit nearby in the value chain, but unless the transaction sat on property interests and long-lease income, it was corporate growth capital, not PERE.
One more point mattered in committee: “biggest” league tables did not tell you whether a deal was smart. In 2025, the hard part was not finding appetite. The hard part was making sure the purchase price was not quietly paying for dreams about power, permits, and delivery schedules.
2025 market context: why deal structure beat deal volume
In 2025, three conditions dominated underwriting and documentation. Together, they made “deal shape” a key driver of outcomes, even when the headline valuation looked unchanged.
Power and grid access became the gating item
Power and grid access became the gating item more often than land, capex, or customer demand. In plain terms, buyers were often paying for a queue position and contracted connection capacity. As a result, purchase agreements leaned harder on step-in rights, assignment mechanics, and tight conditions tied to energization dates. Timing risk moved from “minor delay” to “whole thesis.”
Data centers underwrote like capital programs, not just rent rolls
Underwriting shifted from “data center as a rent roll” to “data center as a capital program with execution covenants.” Even stabilized portfolios carried forward capex because customers asked for more power in the same footprint. Consequently, sponsors pushed governance that looked closer to infrastructure than core real estate. Returns depended on delivery discipline, not only occupancy.
Conservative construction leverage pushed more risk into equity
The debt stack stayed conservative on speculative development. The consequence was that equity absorbed more execution risk, and equity structures did the heavy lifting: preferred equity, delayed-draw equity, JVs with staged capital, and downside triggers in governance documents. Headline valuations held up, but the waterfall often changed who got paid first.
Against that backdrop, the largest disclosed transactions leaned toward platform control, development pipelines, and power-secured clusters. Pure “core” cash-flow trades still happened, but fewer were big enough to dominate the year’s conversation.
Top disclosed and reference data center PERE deals shaping 2025
These transactions were not a complete league table. Instead, they are the disclosed “markers” that shaped how 2025 deals were priced, documented, and explained to investment committees.
Brookfield Infrastructure: Colonial Enterprises’ data centre business (Australia)
Brookfield Infrastructure announced an agreement to acquire Colonial Enterprises’ data centre business for AUD 9.1 billion (May 2025; Brookfield press release). This was a clearly disclosed control acquisition and a pricing marker in a market where power and land scarcity were constraints you could measure.
This was not a simple cap-rate trade. A buyer paid for in-place contracts and for embedded expansion options. The key committee question stayed simple: how much of the price was supported by contracted cash flows today, and how much depended on pipeline optionality that required permits, grid upgrades, and customer pre-commitments. Optionality can pay well, but it rarely pays on schedule.
On structure, expect a “whole business” perimeter with carve-outs for non-core services, a transition services agreement (TSA) for shared functions, and real focus on customer assignment consents. In Australia, FIRB considerations can shape timeline and conditionality, even when the assets are not framed as critical infrastructure. Regulatory timing can become the closing critical path.
Blackstone: AirTrunk as a pricing anchor for hyperscale platforms
Blackstone agreed to acquire AirTrunk in a deal widely reported around AUD 24 billion (September 2024; Reuters). It was not a 2025 signing, but it cast a long shadow over 2025 auctions and refinancings. Sellers loved to reference “AirTrunk-like” growth, while buyers had to separate marketing from math.
AirTrunk mattered for 2025 underwriting because it showed that the market would capitalize long-dated hyperscale cash flows at infrastructure-like multiples when a platform could keep winning power and pre-leases. At the same time, it highlighted concentration and renewal risk. A long contract can behave like a shorter one if the landlord must fund density upgrades to keep the tenant. Renewal economics often moved downside cases more than vacancy assumptions.
It also reminded everyone that “PERE” labels can mislead. Real estate, power contracting, design-and-delivery capability, and customer relationships all contribute to value, and they do not carry the same warranties, financing terms, or exits. Committees should force that separation because remedies depend on which component breaks.
KKR and GIP: CyrusOne as a governance and credit template
KKR and Global Infrastructure Partners led the USD 15.0 billion take-private of CyrusOne (August 2022; company announcements). This was not 2025 activity, but it stayed relevant in 2025 as a template for operating covenants, reporting, and capex governance for mixed colocation and hyperscale portfolios.
Sponsors and lenders leaned on the “CyrusOne model” when negotiating customer concentration reporting, capex approvals, and incentive design for management teams that operate inside real-estate security packages. Good templates shortened execution time and reduced avoidable disputes.
Legacy mega-deals were pricing anchors, not 2025 volume. In many markets, 2025 large-cap trading ran into bid-ask spreads. Deals that did clear often did so by changing risk allocation through structure, not by resetting headline EV.
Stonepeak and EQT: Radius Global Infrastructure (adjacent digital real estate)
Stonepeak and EQT agreed to acquire Radius Global Infrastructure for about USD 3.3 billion (May 2024; company press release). Radius is adjacent rather than pure-play data centers, but it competed for the same capital and often sat in the same “digital infrastructure real estate” sleeve.
The 2025 lesson was how investors priced contracted cash flows when operational complexity and counterparty dependency were higher than traditional property. For data centers, that translated into tighter scrutiny of tenant credit, step-in rights, and termination mechanics, and less reliance on simple market-rent narratives. Legal terms started to drive value more than “rent growth.”
DigitalBridge-linked consolidation: recapitalizations and asset recycling
DigitalBridge’s reported USD 1.0 billion acquisition of Yondr Group (October 2024; Reuters) influenced 2025 structures because it underscored an execution-heavy truth: value creation depends on delivering power-secured developments to hyperscalers on time.
In 2025, the “top deal” dynamic often showed up as repeatable recapitalizations rather than a single blockbuster. Minority equity sales, project-level JVs, and preferred equity funded pipelines while avoiding full exits at valuations sellers viewed as cyclical lows. These structures can be investable when control deals slow, but the waterfall can punish delays.
What changed in 2025 underwriting and deal documents
In 2025, the strongest investment memos matched the legal documents. That alignment mattered because “development risk” was often hidden inside assets marketed as stabilized.
Power became the first diligence workstream
A credible memo started with power. Investors treated grid connection letters, queue position, curtailment terms, upgrade responsibilities, and expected energization dates like title diligence. Where utility frameworks were opaque, buyers used conditions tied to energization milestones and required independent engineer sign-offs. You cannot lease megawatts you cannot deliver.
This moved purchase agreements closer to development agreements. Sellers resisted open-ended milestone conditions, while buyers wanted walk-away rights if energization slipped beyond defined tolerances. Many deals responded with phased closings or drop-down structures, which improved close certainty but increased complexity.
“Land plus power” traded as a de-risked development right
Land pricing followed power rights and permitting, not land comps. Buyers negotiated covenants to preserve queue position, restrict seller actions between signing and closing, and indemnify interconnection misstatements. Where the power position could not be cleanly transferred, buyers demanded step-in rights and post-close cooperation covenants. One drafting error could strand the most valuable asset: the power.
Hyperscale contracts pulled technical obligations into “core” underwriting
Committees still like to think a lease is a lease. Hyperscale agreements often include bespoke service levels, expansion rights, landlord-funded improvements, and termination regimes linked to delivery milestones. In 2025, buyers pushed to diligence side letters, change orders, and customer-specific technical obligations, not just headline rent and term. Hidden obligations became unbudgeted capex.
The practical modeling shift was to run cash flows net of “keeping the tenant” capex. A long lease can be economically shorter if retention requires repeated upgrades that rents do not fully pay for. In that case, IRR can vanish without a vacancy event.
Structured equity replaced price cuts
When sellers held to 2024-style valuations, deals still cleared by shifting risk through structure: preferred equity with current-pay and PIK toggles, staged equity tied to permits, energization, and pre-leasing, and minority sales with downside-trigger governance that shifts control. Downside distribution changed even when the press release did not.
These tools were not harmless. Preferred accruals could create refinancing cliffs if stabilization lagged. Staged equity could create funding gaps if milestones slipped. Minority deals could turn into control fights under stress. A “cheap” entry price could become expensive if time worked against you.
Credit and security packages tightened
Even with stable headline leverage, lenders demanded tighter cash control and stronger remedies: cash sweeps, debt service reserve accounts (DSRA), distribution lock-ups tied to construction milestones, and detailed reporting covenants. Equity should treat this as equity risk because restrictive debt can force equity cures, capex deferrals, or covenant-driven sales at bad moments.
Energy reporting became a deliverable, not a marketing slide
Tenants, lenders, and regulators pushed energy efficiency and emissions reporting into covenants and information undertakings. The diligence question became practical: can the platform produce audited, asset-level energy and emissions data, and does failure create tenant default risk or financing consequences?
Fresh angle for 2025: teams increasingly treated “data exhaust” as diligence. Metering coverage, data retention, and auditability were reviewed like financial controls because they affected billing accuracy, SLA compliance, and even the ability to refinance with tighter lender reporting. In short, operational data quality became a finance issue.
Deal mechanics practitioners actually negotiated
Most large deals used holding-company acquisitions with property-owning subsidiaries and intermediate holding structures to ring-fence liabilities. OpCo-PropCo splits remained common: employees and service delivery in OpCo, real estate in PropCo. Lenders pushed bankruptcy remoteness through separateness covenants, limits on additional debt, and independent director requirements.
Funds flow often ran in two steps. First, purchase price went into escrow or closing, and change-of-control debt was repaid. Then, post-close capitalization funded capex reserves and working capital with lender-controlled accounts. Under-construction assets often came through phased closings: stabilized first, developments later after permits, energization, or pre-leases.
Security packages focused on real estate, shares in SPVs, key contracts (customer, utility, construction), and bank accounts under strict cash dominion. Step-in rights were heavily negotiated because enforceability determined recovery. A step-in right that requires a counterparty’s consent at the moment of distress is not a right; it is a request.
Consents and transfer restrictions in customer contracts often set the timetable. Where consents could not be obtained pre-close, parties used interim structures, “silent consent” mechanics where feasible, and drop-dead dates. Committees should identify contracts where consent is required and delay-prone, and then stress-test denial because indemnities rarely replace a lost anchor contract.
Economics: where returns were won or lost
Returns were frequently won in the “unsexy” parts of the model. Transaction costs rose because diligence widened: power engineers, grid consultants, cyber and physical security assessments, and heavier legal drafting for milestones and technical schedules. Financing fees remained meaningful, and ticking fees on acquisition debt added carry. Delays directly taxed equity IRR, so sponsors who managed timeline risk often outperformed sponsors who only negotiated price.
Ongoing fees mattered, especially in platform structures. Asset management fees, development management fees tied to capex, and property or technical operations fees sometimes flowed to affiliates. Minority investors pushed for fee caps, related-party approval, and audit rights because leakage and conflicts showed up over time, not at signing.
Tax leakage remained deal-specific. Cross-border platforms relied on intermediate holdings and shareholder loans, but withholding tax, hybrid mismatch rules, and interest limitation regimes required early structuring. European deals demanded attention to transfer taxes and beneficial ownership substance. The wrong structure turned “good yield” into “good gross yield.”
Governance and downside control: the 2025 playbook
Governance provisions expanded in JVs and minority deals because operating decisions drove outcomes. The provisions that protected equity were practical, not philosophical, and they showed up as reserved matters and reporting rights.
- Capex control: Require budget approvals, change-order thresholds, and objective triggers for “must-spend” upgrades tied to customer retention.
- Customer concentration: Set limits and enhanced reporting for top tenants, including renewal economics and upgrade obligations.
- Merchant power limits: Restrict exposure to uncontracted power costs and define who bears curtailment risk.
- Milestone reporting: Track permits, interconnection steps, equipment lead times, and energization dates with escalation paths.
Deadlock clauses looked good on paper, but valuation disputes and tight financing often made put-calls hard to use. More workable tools were escalation ladders, independent expert determinations for technical disputes, and step-in rights tied to objective milestone failures. Workable remedies beat elegant ones.
Edge cases stayed real and specific. Antitrust clean teams showed up when competitors bid in consortia, a theme that also overlaps with cross-border M&A execution. Export controls and CFIUS-style reviews hit certain counterparties and equipment flows. PII and HR files required cross-border notifications and strict access controls. Small workstreams could still hold up a closing.
What “top deals” signaled for 2026 pipelines
The largest disclosed transactions reinforced a simple point: platform scarcity and power access still earned premium pricing when the assets were executable. They also reinforced bifurcation. Stabilized, well-located cash flows stayed liquid, while speculative pipelines demanded structure, governance, and technical diligence that resembled infrastructure underwriting.
If there is one durable lesson from 2025, it is that “data centre PERE” is a spectrum. Some deals look like contracted real estate, while others look like power projects with a building attached. The teams that did well priced that spectrum explicitly in covenants, documents, and staged capital, and they refused to pay tomorrow’s price for next year’s permission.
Before you archive the deal, close it like you mean it: index the final set (versions, Q&A, user lists, and full audit logs), hash the archive, apply a written retention schedule, obtain vendor deletion plus a destruction certificate, and remember that legal holds override deletion.
Conclusion
In 2025, the “top” data center PERE deals were less about headline size and more about how risk was allocated across power, delivery, contracts, and the capital structure. If you can compare EV consistently, diligence power like title, and negotiate governance that matches the execution plan, you are far more likely to turn a fashionable platform into a durable return.