Data Center Joint Venture Structures in 2026: The Capital Architecture Behind the Builds

Hyperscalers, PE sponsors, and developers are no longer operating as landlord and tenant. The capital structure itself has changed.

The Aligned Data Centers acquisition closed the first half of 2026 as the most structurally significant transaction in the data center market in years. The deal was not primarily about the real estate. It was about who owns the infrastructure layer -- and what that signals for every other developer currently building or planning AI data center capacity.

A consortium anchored by BlackRock's Global Infrastructure Partners, MGX, Microsoft, and Nvidia acquired 100% of Aligned at an implied $40 billion enterprise value. Microsoft and Nvidia are not traditional infrastructure investors. Their presence as equity co-investors in an operating platform marks a structural shift: hyperscalers are moving from tenants who write long-term leases to equity participants who share in the development risk and the upside.

That shift is reshaping how institutional data center development gets financed.

Why the Traditional Lease Model Is Insufficient

For the past decade, the standard data center capital structure looked like this: a developer or operator owned the facility, signed a long-term lease with a hyperscale tenant, and used the contracted cash flows to support construction debt and equity returns. The hyperscaler got capacity without owning the asset. The developer got the anchor credit. The structure worked.

It is still being used, and it will continue to be used. But it is no longer the only structure in the market, and increasingly it is not the structure that gets the largest projects done.

Three forces are pushing toward more complex arrangements.

First, the capital requirements have outgrown what single sponsors can deploy. A 300 MW hyperscale campus at $11 to 20 million per megawatt requires $3.3 to 6 billion in construction capital before tenanting. That is a concentration of risk that institutional equity sponsors typically manage through co-investment rather than sole ownership.

Second, hyperscalers are moving faster than development capital can track. Microsoft, Google, Amazon, and Meta collectively committed over $280 billion in AI infrastructure capex through 2027. Conventional developer economics -- acquire land, underwrite on a speculative basis, then sign a lease -- creates a timeline mismatch with hyperscalers who need capacity on a defined schedule for AI product roadmaps that do not move.

Third, the strategic importance of power-certain sites has created a land and infrastructure premium that rewards early mover equity participation over late-stage lease negotiation.

The Joint Venture Structures in Use

Development JV with hyperscaler equity. The Aligned deal is the largest example, but the structure is being used across smaller transactions. A developer or operator brings a site, permits, and development capability. The hyperscale tenant contributes equity capital in exchange for preferential capacity access and a share of the economic upside. The developer retains operational control and development management. This structure aligns incentives on schedule in a way that a conventional lease cannot -- a hyperscaler who is an equity partner is invested in delivery in a different way than a tenant who can invoke force majeure provisions.

Infrastructure fund plus operator. Blue Owl's Abilene project is the clearest current example of this structure: equity and project-level debt invested through a dedicated SPV, backed by a long-term contracted lease with a creditworthy tenant, with the developer managing construction and operations. The infrastructure fund provides patient capital at a lower cost of capital than typical private equity. The developer generates fees and promotes. The hyperscale tenant gets capacity without balance sheet exposure. This is essentially a project finance structure applied to data center development.

Equity recycling through stabilized asset sale. Developers build and lease-up, then sell the stabilized cash-flowing asset to long-duration capital: pension funds, sovereign wealth funds, insurance company real estate platforms. CapitaLand's $982 million acquisition of a 49% stake in an Osaka hyperscale data center in April 2026 is a clean example. The developer (Nippon Telegraph and Telephone, in this case) retains the other half, generates liquidity, and uses the proceeds to fund the next development. JLL refers to this dynamic in the APAC market as "capital recycling" -- the term is accurate.

SPV co-investment structures. For projects where the hyperscale commitment is strong but the development risk is unusual -- an off-grid power plant, a brownfield conversion, a campus that requires utility upgrade coordination over multiple years -- developers are using SPV co-investment structures that allow them to bring in project-specific capital partners without affecting their broader platform economics. This keeps leverage off the developer balance sheet while providing the project-level capital stack depth that unusual site types require.

What PE Sponsors Are Doing

Private equity investment in US data centers reached $45.7 billion in 2025, the highest in at least five years. But the deal count is shifting. In the first half of 2026, $12.3 billion in PE-backed deals closed while deal count fell 17% year-on-year. Median EBITDA multiples compressed to 18.2x from 19.8x in H1 2025.

PE sponsors are concentrating capital on fewer, larger transactions with clearer hyperscale demand signals. KKR acquired STT GDC for an implied enterprise value of $10.8 billion in a consortium with Singapore Telecommunications -- a platform deal rather than a single-asset acquisition. DigitalBridge acquired Yondr Group to build a global hyperscale development platform rather than a portfolio of stabilized assets.

The implication for mid-market developers is that PE as a passive financial sponsor for conventional data center development is becoming more selective. Funds that are deploying at scale want platform exposure, not a single campus in one market. Developers without a multi-market development pipeline or a credible differentiated capability -- site sourcing speed, power certainty, tenant relationships -- are finding the PE bid for their projects thinner than it was two years ago.

The New Due Diligence for JV Partners

The joint venture structure works when partners have aligned incentives and clearly defined responsibilities. It fails when the division of accountability is ambiguous or when one party's risk tolerance is miscalibrated against the project.

For developers evaluating JV arrangements with hyperscale equity partners, three areas require explicit negotiation.

Schedule rights and remedies. A hyperscale equity partner who is also a tenant has interest in accelerating delivery that a purely financial equity partner does not. But that interest can create tension if the development team encounters permitting delays, power delivery timeline shifts, or supply chain disruptions. The JV documents should define what triggers schedule remedies, what the remedy mechanism is, and how disputes are resolved.

Power delivery risk allocation. In markets where utility coordination timelines are running 18 to 36 months, the risk of a power delivery delay materially affecting lease commencement should be explicitly allocated between the parties. This is not theoretical. The Virginia market, the Texas market, and several PJM-adjacent markets have enough queue congestion that a 12-month power delivery slip is a real scenario.

Exit and recapitalization provisions. JV structures for large development projects typically run 10 to 15 years from inception to stabilization and potential monetization. The documents should address what happens if a partner needs liquidity before stabilization, how the asset gets valued for partial buyout, and what the governance structure is for major decisions -- capital calls, refinancing, major capex, lease modifications -- when partners may have evolved priorities.

The Developer's Position

Independent developers without hyperscale relationships or infrastructure fund backing are not shut out of the market, but the capital formation path is more deliberate than it was two years ago.

The developers who are structuring their next projects successfully are approaching capital formation the same way they approach power diligence: early, specific, and with a clear view of which structure serves the project's risk profile.

A 50 MW speculative campus in a secondary market with partial power certainty needs a different capital partner than a 200 MW build-to-suit with a signed pre-lease and utility confirmation in hand. Matching the capital structure to the development risk is the core of the current market.

The Aligned transaction reset expectations for what the largest deals look like. For the rest of the development market, it is also a signal: the projects that will attract capital are the ones where the developer can demonstrate not just that the site works, but that the power certainty, the tenant relationship, and the development capability justify a long-term equity partnership rather than a conventional lease.

That is a higher bar. It is also a more durable business.