Data Center Acquisitions: What the Money Is Actually Chasing

The deal structure buried in Oi's telecom breakup reveals something most acquisition headlines won't: when a distressed operator sells its remaining stake and the buyer's strategy explicitly bundles submarine cables with data centers, you're not looking at a real estate play. You're looking at a bet on where the internet's physical backbone goes next.

That's the lens worth applying to the broader wave of data center acquisitions reshaping infrastructure investment right now. Not "who bought what" — but *why* these assets are being assembled in specific combinations, and what that reveals about where returns are actually coming from.

What's Actually Being Acquired — and Why It Matters

A data center acquisition sounds straightforward: buy the building, buy the servers, control the revenue. The reality is considerably more complex, and the gap between those two descriptions is where most outside observers miss the point.

What sophisticated acquirers are really buying is interconnection — the ability to move data fast, reliably, and cheaply between facilities, cloud regions, and undersea cable landing stations.

The Oi situation illustrates this precisely. The transaction wasn't structured around a single data center asset. It explicitly combined the operator stake with submarine cable infrastructure. That's not coincidental packaging — it's a deliberate signal about how these assets create value together. A data center sitting on or near a cable landing station commands fundamentally different economics than an identical facility sitting inland. Latency drops. Redundancy improves. And for hyperscale cloud customers, those two factors can justify meaningfully higher contract rates.

This matters for anyone evaluating infrastructure investments right now: the unit of analysis has shifted. It's no longer a single facility. It's the *network* of facilities and how they connect.

The Forces Making These Deals Happen Now

Several structural forces have converged to accelerate deal velocity in this sector, and they're not going away.

Demand for compute and storage capacity has outpaced what organic construction can deliver at the speed buyers need. Building a hyperscale data center from greenfield takes 18 to 36 months in favorable conditions — longer when you factor in power interconnection queues, which have stretched to multi-year timelines in constrained markets like Northern Virginia, Dublin, and Singapore. Acquisition compresses that timeline to zero. You're buying operational capacity today, not a construction promise.

The remote work shift accelerated enterprise cloud adoption by roughly three to five years, depending on which analyst forecast you trust. That acceleration created immediate density requirements that operators couldn't build fast enough to satisfy. The result was a seller's market for operational data center assets — and transaction multiples reflected it, with stabilized facilities trading at EV/EBITDA multiples in the high teens to low twenties in peak years.

There's also a less-discussed driver: the geographic rebalancing of internet traffic. As content consumption grows faster in Southeast Asia, Latin America, and Sub-Saharan Africa than in mature Western markets, data center infrastructure is being acquired and built in locations that weren't on most institutional investors' maps five years ago. Submarine cable routes follow that demand, and data center acquisitions follow the cables.

Submarine Cables: The Infrastructure Nobody Talks About Until It's Critical

Submarine cables carry roughly 95% of international internet traffic. They are, without exaggeration, the physical foundation of the global digital economy. And yet most infrastructure investment coverage treats them as a footnote.

The Oi transaction's explicit bundling of submarine cable assets with the data center stake reflects a more sophisticated understanding of how these two infrastructure categories interact. A cable landing station without nearby data center capacity to terminate traffic is a choke point. A data center without low-latency access to submarine cable infrastructure is, for certain workloads, economically disadvantaged against competitors that have it.

When hyperscalers like Google, Meta, and Microsoft began building their own private submarine cable systems — rather than leasing capacity on consortium cables — they weren't just reducing costs. They were vertically integrating the infrastructure stack, from the ocean floor to the server rack.

Independent data center operators and telecom companies paying attention understood the implication: own the cable adjacency, or risk being commoditized by customers who control their own fiber paths. That logic has quietly driven a meaningful portion of data center acquisition activity in cable-rich markets — coastal facilities, port-adjacent campuses, and assets with existing IRU agreements on key trans-oceanic routes.

For infrastructure investors evaluating data center deals, submarine cable access — owned, leased, or proximate — deserves a line item in due diligence. It's not a bonus feature. In the right markets, it's a core value driver.

Where the Returns Are — and Where They're Not

Data center acquisitions have attracted institutional capital from private equity, sovereign wealth funds, REITs, and pension funds, all of which have different return expectations and hold periods. That diversity creates both opportunity and risk.

Core data center assets — fully leased, long-term contracted, investment-grade tenants — have compressed to cap rates that make traditional real estate investors uncomfortable. Sub-4% cap rates were not uncommon at the peak. At those levels, the return thesis depends almost entirely on rent escalation and occupancy stability, both of which carry real risk in a market where hyperscaler customers negotiate hard and occasionally build competing capacity.

The more interesting return profile is in the value-add and opportunistic buckets: distressed operators like Oi, secondary markets with undersupplied power capacity, and edge data center assets positioned to serve latency-sensitive workloads that centralized hyperscale facilities can't efficiently serve.

Edge infrastructure — smaller facilities distributed closer to end users — is the segment most analysts agree on in theory and most struggle to underwrite in practice. The economics work at scale across a large portfolio. Individual edge nodes often don't pencil as standalone investments, which is why aggregation plays and platform acquisitions have dominated that segment.

ROI analysis on data center acquisitions also needs to account for power costs and infrastructure — increasingly the binding constraint. Facilities with secured, long-term power purchase agreements at competitive rates are worth materially more than physically identical facilities facing power cost uncertainty. In some markets, a 2-megawatt power contract has become the hardest asset to acquire, harder than the land or the building itself.

What Can Go Wrong

Regulatory exposure is real and growing. Cross-border data center acquisitions, particularly those involving telecom operators or submarine cable assets, have drawn national security review in the US, EU, UK, Australia, and an expanding list of other jurisdictions. The CFIUS process in the United States can delay or block transactions involving foreign acquirers, and the definition of "critical infrastructure" has expanded considerably in recent years.

The practical consequence: deals that would have closed in six months now regularly take twelve to eighteen, with conditions attached. Acquirers need to price regulatory risk into deal structures, including potential divestitures, operational restrictions, and ongoing government access requirements.

Market competition has also created execution risk that wasn't present when fewer players were chasing fewer deals. Auction processes for premium assets now routinely attract eight to twelve qualified bidders, which means overpaying is a real and documented outcome — not a hypothetical. The discipline required to walk away from a deal at a given price is genuinely difficult when your investment mandate requires deploying capital into this sector and quality assets are scarce.

There's also integration risk specific to acquisitions involving legacy telecom operators. Oi-type situations — distressed carriers with data center or cable assets mixed into complex corporate structures — often come with deferred maintenance, long-tail liabilities, and workforces operating under collective agreements that complicate operational transformation.

What Comes Next

The acquisition wave isn't finished, but it's maturing. Early-mover advantages in primary markets are largely exhausted. The next phase favors operators who can identify value in secondary geographies, navigate complex regulatory environments, and integrate submarine cable access into coherent infrastructure networks — not just stack individual assets.

Investors sitting on the sidelines waiting for cap rate normalization may be misreading the market. The structural demand drivers — AI compute requirements, continued cloud adoption, and the geographic expansion of digital infrastructure — are compressing the window for opportunistic entry. Assets that look expensive today may look cheap against the demand backdrop of 2027.

The Oi transaction, modest as it may appear in isolation, is a useful reminder: the acquirers who understand that data centers and submarine cables are one integrated infrastructure category, not two separate ones, are the ones writing checks that will look prescient five years from now.

**Explore the InfraSale Marketplace for the latest data center opportunities!**