Everything About AI
NextFin News - Blue Owl Capital is working to create a publicly traded real estate investment trust focused on data centers, seeding it with roughly $6.5 billion of existing digital-infrastructure assets and taking the vehicle public through an initial offering, according to people familiar with the matter. The plan, which remains under discussion and could still change, would make Blue Owl the latest private-capital giant to convert AI-era data-center holdings into a listed company — and it lands at a moment when the public market is sending mixed signals about how much it will pay for that exposure.
The Deal: A $6.5 Billion Seed, an IPO, and a Follow-On Engine
The proposed REIT would start with approximately $6.5 billion of data-center assets already held by Blue Owl, rather than launching as a pure blind pool. The company intends to raise fresh capital through an IPO and subsequent share sales, using the proceeds to acquire additional facilities and expand the portfolio, the people said. A Blue Owl representative declined to comment.
The structure matters. Several data-center vehicles backed by large investment firms have entered the public market this year as artificial-intelligence infrastructure spending accelerates. Blackstone's Digital Infrastructure Trust raised $2 billion in its May debut as a blind pool, letting investors buy shares before knowing which assets the REIT would eventually acquire. Csquare, backed by Brookfield, priced its IPO at $21 a share and raised $1.05 billion in July; its shares fell about 3% on their New York Stock Exchange debut, leaving the company valued at roughly $3.24 billion. Csquare sold 50 million primary shares, valuing its equity at about $3.25 billion.
Blue Owl's planned vehicle would operate along similar lines to Blackstone's blind-pool REIT but would begin with initial assets already in place as a foundation — a hybrid that gives investors both a visible starting portfolio and a committed acquisition pipeline. The design is a direct response to the central problem of the AI data-center boom: private capital built the stack, but private funds have finite lives, and their limited partners are increasingly asking for liquidity.
Why Blue Owl, and Why Now
Blue Owl is not a newcomer to digital infrastructure. In May it closed its third digital-infrastructure fund at $7 billion, with plans to develop, acquire, and own data centers to meet AI-driven power demand. The previous September, the Qatar Investment Authority partnered with Blue Owl to launch a digital-infrastructure platform carrying more than $3 billion of data-center assets; the sovereign fund's chief executive said in May that the fund would commit an additional $500 billion to U.S. investments over the next decade, particularly in AI and data centers.
The firm's largest single exposure is the Hyperion campus in Richland Parish, Louisiana. In October 2025, Meta struck a $27 billion financing deal with Blue Owl — Meta's largest-ever private-capital agreement — under which funds managed by Blue Owl took an 80% stake while Meta retained about 20% and leases back the facilities. By July 2026, Meta said the Hyperion supercluster would be a 5-gigawatt site costing more than $50 billion, up sharply from the original $27 billion figure. That structure is the archetype of the current boom: a hyperscaler keeps the project off its balance sheet, a private asset manager owns the real asset, and the tenant signs a long-dated lease.
But the architecture has a timing problem, and it is not unique to Blue Owl. In February 2026, the firm sold roughly $1.4 billion of direct-lending investments across three funds and said it would return 30% of investors' capital at book value within 45 days — a move that rattled private-credit peers and pushed Blue Owl's shares down more than 5% in a single session. In December, Oracle said Blue Owl was not contributing equity to a data-center financing in Michigan, a split that rattled markets banking on the AI boom. Private funds need to return capital. Public markets offer permanent capital. The REIT is the bridge between the two — and, for Blue Owl, a way to turn a series of finite-life fund positions into a perpetual fee-generating asset base.
The Public Market Is Not Yet Convinced
Here is the tension Blue Owl must resolve. The AI infrastructure trade has been one of the most crowded in markets, yet the companies that actually own the buildings have not captured the gains. As of early 2026, Equinix carried a market capitalization of about $78 billion, Digital Realty about $55 billion, and Iron Mountain about $27 billion — and analysts noted that the market believed the economic profits of AI would accrue to chip designers and hyperscalers, not to what one observer called "mercenary" data-center landlords.
Public REIT shareholders are also more risk-averse than private-infrastructure investors. They do not want these companies to lever up into deals with a non-negligible default risk — which, in today's data-center market, is often exactly what building a facility requires. REITs must distribute 90% of taxable income, and they face quarterly scrutiny that private funds do not. Brookfield-backed Csquare underscored the point: the company paid its owner a $785 million distribution last year, more than four times its $172 million of operating cash flow in 2025, before coming to a public market that priced the offering below its marketed range.
Even so, the operators insist demand is overwhelming supply.
"Essentially, we're in a place where demand is well outpacing supply, so the speculative data center builds, you can't build it fast enough for the customers," said Andy Power, chief executive of Digital Realty, adding that vacancies at the company are the tightest they have ever been.
The Second-Order Trade: Who Really Owns the AI Stack?
The first-order read of this news is simple: Blue Owl wants a cheaper, permanent source of capital. The second-order implication is more consequential. A successful Blue Owl REIT would begin shifting the ownership of AI data centers from private credit and off-balance-sheet joint ventures toward public equity — and with it, the transparency that comes with quarterly reporting, mark-to-market pricing, and minority shareholders.
Consider the Meta-Blue Owl Hyperion structure. Meta keeps the debt off its balance sheet and books a right-of-use asset rather than the full facility; Blue Owl's funds own 80%. That arrangement works smoothly while capital is private and patient. Once the asset sits inside a public REIT, the lease terms, the credit of the tenant, and the valuation of the campus become visible to every analyst with a spreadsheet. The accounting convenience that made the private structure attractive — the ability to look asset-light while still controlling the infrastructure — comes under pressure.
The transmission channel runs through the cost of capital. If public REIT investors accept data-center assets at cap rates below what private funds demand, hyperscalers gain a cheaper financing option and the REIT sponsors gain fee-generating assets under management. If public investors demand a risk premium — as the Csquare debut suggested they might — the private structure retains its advantage, and the REIT becomes a vehicle for exiting mature, contracted assets rather than funding the next wave of speculative builds.
There is a third-order effect worth naming. A wave of data-center REIT supply changes who sets the price of AI infrastructure. Today, valuations are set in private transactions between a sponsor and a hyperscaler, with little daylight. Tomorrow, a publicly traded multiple on funds from operations becomes the reference point for every new deal. That is good for market discipline and bad for the opacity that made the boom's financing so easy to arrange.
The Counter-Thesis: The REIT May Be the Wrong Vehicle
The strongest case against Blue Owl's plan is that the REIT format is structurally mismatched to the asset class it is trying to monetize. Data-center development in the AI era is a high-leverage, long-gestation, power-constrained business. REITs are engineered for stable, contracted, income-producing property. They must pay out 90% of taxable income, they cannot easily retain earnings to fund speculative development, and they are punished by rising discount rates — the very environment that has shadowed the sector since 2022.
Private-infrastructure funds, by contrast, can sit on development risk, recycle capital quietly, and underwrite to a hyperscaler's credit rather than a public multiple. The Csquare IPO priced below range and opened lower for a reason: public investors are asking whether these assets are scarce infrastructure or leveraged bets on a handful of technology tenants whose capex plans can change with the next earnings call.
There is also a supply question. If Blackstone, Brookfield, and Blue Owl all bring data-center vehicles to market in the same window, public investors get to choose — and choice compresses pricing. The blind-pool model, which dominated 2026's debut class, transfers acquisition risk to shareholders who cannot see the assets until after they buy. Blue Owl's decision to seed its REIT with existing holdings is a concession to that skepticism; it says, in effect, that a blind pool alone will not clear the market.
The falsifying signal is specific: if Blackstone Digital Infrastructure Trust trades sustainably above net asset value and Csquare recovers above its $21 IPO price within six months of listing, the "public market will not pay for data centers" thesis is wrong — and Blue Owl's timing would look prescient rather than risky.
Cyclical Exit, Structural Demand
This is the call the piece rests on: the demand for AI data-center capacity is structural, but the financing wave that built it is cyclical — and Blue Owl's REIT is a cyclical exit mechanism dressed in structural clothing.
The structural leg is real. Power consumption is rising on a trajectory that does not revert. The U.S. Energy Information Administration has projected electricity use to reach record highs — roughly 4,199 billion kilowatt-hours in 2025 and 4,267 billion in 2026, up from 4,110 billion in 2024 — with data centers, including AI and crypto, cited as a primary driver. Hyperscalers have committed to tens of gigawatts of capacity this decade. Leases are long, tenants are investment-grade, and the physical constraints of power and land are genuine moats.
The cyclical leg is equally real. The capital stack that financed the boom — private credit, sale-leasebacks, off-balance-sheet joint ventures — was a product of a specific moment: abundant private liquidity, hyperscalers desperate for speed, and accounting rules that rewarded opacity. That moment is passing. Fund vintages are maturing, limited partners want liquidity, and the public window, however narrow, is open. A cyclical wave of exits does not invalidate the structural demand; it simply means the owners are changing, and the price of capital is being rediscovered in a more transparent venue.
Getting this distinction wrong flips the conclusion. If an investor reads the REIT as proof that the AI infrastructure boom is peaking, that is a category error: the boom's demand side is intact, but its financing side is rotating from private to public. The rotation itself is the story.
What to Watch
Short term (sentiment and pricing): The IPO terms. A seed portfolio valued at $6.5 billion with a modest premium to net asset value would signal that Blue Owl is prioritizing execution over extraction. A richly priced deal in a weak tape would risk a Csquare-style reception.
Medium term (fundamentals): Lease-up and power delivery. The key metrics are contracted occupancy, weighted-average lease term, and — critically — whether facilities can actually be energized on schedule. Power constraints, not demand, are the binding constraint on this trade.
Long term (structural): The cost-of-capital gap between public REITs and private funds. If public markets consistently price data centers below private implied values, the REIT model wins and more sponsors follow. If the gap persists, the assets stay private and the REIT becomes a niche exit rather than an industry reset.
Base case: Blue Owl prices a modestly sized IPO, the stock trades flat to slightly positive, and the vehicle becomes a steady acquirer of contracted assets. Upside case: the hybrid seeded-plus-blind-pool model is embraced, Blackstone's and Csquare's vehicles rerate higher, and the REIT becomes the template for the next wave of private-infrastructure exits. Downside case: public investors balk at the pricing or at hyperscaler concentration, the IPO is delayed or downsized, and the assets remain in private funds longer than planned.
The signal that would prove the bullish read wrong is not a bad quarterly print — it is a sustained discount of public data-center REITs to private transaction values. As long as that discount persists, the smart money stays private.
Blue Owl's REIT is not a bet that the AI boom is ending. It is a bet that the people who built it are ready to sell their seats — and that the public market is finally willing to buy them.
Explore more exclusive insights at nextfin.ai.
