Control of compute is becoming a capital-allocation problem. Meta Platforms’ proposed El Paso transaction shows the emerging solution: institutions own and finance the physical infrastructure while Meta retains operational access through a long-term lease.
Under the reported structure, BlackRock-managed funds would own 80% of the project, Meta would retain 20%, and Meta would lease the entire facility 1,4,17,21,25. The proposed campus is described as a roughly $14 billion, 1-gigawatt AI data center 11,16. Its importance extends beyond one site. It offers a template for financing hyperscale AI capacity without requiring the technology company to fund every asset directly from its own balance sheet.
The core ownership structure is relatively well supported. Five sources identify the 80% BlackRock-managed-fund and 20% Meta split 1,25. Three support the existence of the Meta–BlackRock joint venture 4,17,21, and two identify BlackRock as Meta’s strategic financial partner 20. The broader economics remain unverified. The transaction’s debt terms, lease duration, accounting treatment, recourse, and return thresholds are not established by the available claims.
How the structure works
The proposed arrangement is a financing structure, not a transfer of operating control to BlackRock. BlackRock-managed funds would provide the majority of the capital and own most of the physical infrastructure 4,11,22. Meta would retain a minority stake and secure dedicated capacity through a long-term lease 4,11. One source describes BlackRock as a financial sponsor rather than the data-center operator 5. Another cautions that majority ownership by investment funds does not necessarily confer direct operational control over Meta’s facilities 16.
That distinction matters. BlackRock would own the asset. Meta would use and operate the capacity. Control is divided between asset ownership and infrastructure utilization. Meta gains access to strategic compute without making BlackRock an operating competitor or assuming sole responsibility for the campus.
The immediate benefit to Meta is capital sharing. The joint venture could accelerate data-center capacity while distributing the capital requirement with BlackRock 4. It is also described as mitigating near-term funding pressure from Meta’s accelerated AI investment 4. In practical terms, the model lets Meta secure AI capacity without relying exclusively on its own balance sheet 16.
This is the larger signal. Data centers may be moving from assets funded and held exclusively by hyperscalers toward institutionally owned infrastructure 11,16. The old model concentrated ownership, financing, and operational risk inside the technology company. The new model separates those functions and brings long-duration institutional capital into the infrastructure layer.
Debt is doing the heavy lifting
The proposed $14 billion venture appears heavily dependent on leverage. Claims describe it as primarily debt-financed 22 and characterize the infrastructure joint venture as using heavy debt financing 22. BlackRock reportedly sought billions of dollars of project debt 6. Separately, BlackRock, JPMorgan Chase, and Morgan Stanley were said to have marketed approximately $12.3 billion of high-grade corporate bonds through specialized holding companies 19. These financing claims are not independently corroborated within the cluster and should be treated as indicative, not definitive.
The math is simple. Debt can reduce Meta’s upfront capital burden, but it does not remove the cost of infrastructure. It converts part of that cost into interest expense, refinancing exposure, and lease obligations. The resulting economics depend on interest rates, construction execution, power availability, facility utilization, and the ability of the campus to remain productive as AI hardware changes.
Risk therefore migrates rather than disappears. It moves from Meta’s corporate balance sheet into project vehicles, institutional portfolios, private-credit markets, and long-term contracts with Meta as lessee. That shift raises governance and fiduciary questions involving risk disclosure, valuation, leverage, conflicts of interest, asset allocation, and the suitability of these investments for pension and insurance capital 22. It also creates broader systemic and regulatory concerns if technology companies increasingly depend on institutional financing for essential compute infrastructure 22.
BlackRock’s role as institutional capital provider
BlackRock’s broader platform makes its participation plausible. The firm manages institutional and retail assets 12, has linked its initiatives to long-term institutional capital 10, and expanded its private-asset capabilities in infrastructure and credit through acquisitions completed through 2024 3. It has also prioritized long-term, stable capital in arranging the Meta financing rather than investors seeking quick profits 6.
Those facts establish capacity, not favorable economics. They do not show that Meta receives cheap financing, that BlackRock’s funds achieve superior risk-adjusted returns, or that the project improves Meta’s return on invested capital. Sentiment is noise. The relevant question is the after-cost return on the capacity.
The transaction also fits a wider capital-markets trend. BlackRock reportedly raised capital for AI infrastructure and plans to raise more to support U.S. leadership in AI 24,27. Its participation in Nvidia’s financing effort places it among several major financial institutions backing the buildout of AI-related infrastructure 26,27. This could deepen the funding pool available to Meta. It could also intensify competition for institutional capital, power, data-center sites, and financing capacity across Microsoft, Amazon, Google, Nvidia-backed projects, and other AI developers.
BlackRock’s cryptocurrency ETFs and tokenized-fund initiatives provide additional context, but they are peripheral to the El Paso project. BUIDL tokenizes Treasuries 2,8. BlackRock has launched tokenized funds on Ethereum and Solana 9 and supports broader real-world-asset tokenization 23. These initiatives show an asset manager building financial infrastructure around digital assets, alternative assets, and long-duration physical infrastructure. The cluster provides no evidence that blockchain financing is being used for El Paso. Any relevance to Meta is indirect.
The economics and risks for Meta
For Meta, the proposed transaction could improve strategic flexibility while AI infrastructure becomes a central constraint on product development and competitive position. Access to a dedicated 1-gigawatt facility could support training and inference capacity, help Meta compete for scarce compute, and preserve capital for model development, applications, and other initiatives. Sharing the funding requirement with institutional owners reduces the need for Meta to finance every dollar of infrastructure directly 4.
The trade-off is a shift from balance-sheet ownership risk to contractual and operating leverage. Meta would lease the entire campus while holding only 20% ownership 14,25. Long-term lease payments could become a material fixed obligation regardless of AI monetization, utilization, power costs, or the pace of hardware obsolescence. The arrangement may improve near-term liquidity and make reported capital intensity appear more attractive. It also increases the importance of future free cash flow, lease-adjusted leverage, and the facility’s economic useful life.
Strategic control requires equal scrutiny. Meta may control use of the facility through its lease and operating role, but the majority owner would be BlackRock-managed funds rather than Meta itself 5,16. Investors need clarity on control over expansion, hardware refreshes, power procurement, maintenance standards, refinancing, and disposition. They also need to know whether the lease is cancellable, whether Meta bears completion and cost-overrun risk, and whether the project is consolidated or treated as an operating lease, finance lease, joint venture, or another structure.
The campus creates concentration risk as well. A 1-gigawatt AI build concentrated at one site creates dependence on local power, grid access, construction execution, cooling systems, and operating performance 16. Delays, grid constraints, technology obsolescence, cooling failures, or underutilization could affect Meta’s AI deployment schedule and the returns earned by institutional owners. The available claims do not establish construction status, lease economics, financing recourse, or the allocation of cost overruns. Those are not footnotes. They are the deal.
Institutional participation is not proof of value
BlackRock’s market signals are strong. Institutional ownership is reported at 83.8% across 3,025 institutions, with net buying detected 12,13,15. BlackRock also captured a dominant share of recent cryptocurrency ETF inflows 7,18. But institutional accumulation does not establish undervaluation or durable fundamentals. One claim says the bullish case rests on accumulation and market conditions rather than demonstrated undervaluation 12. Another warns that the current score may reflect short-term flows rather than durable fundamentals 12.
The same discipline applies to Meta’s infrastructure deal. Institutional participation validates the availability of capital. It does not validate the project’s returns. The return calculation must include rent, interest, depreciation, maintenance, power, refinancing, and contractual commitments. The asset creates value only if Meta converts the added compute into higher engagement, advertising efficiency, new revenue streams, and defensible AI products.
Strategic implication
The El Paso venture is a case study in the financialization of compute. Hyperscalers retain operational access while institutions own, leverage, and monetize the underlying physical assets. This can accelerate capacity buildout and expand Meta’s financing options. It also pushes more risk into lease contracts, project-finance vehicles, credit markets, and institutional portfolios.
The model could give Meta a financing advantage. It does not automatically give Meta a cost advantage or a durable moat. Its strategic value depends on the terms of control, the cost of capital, the reliability of the facility, and Meta’s ability to monetize the resulting capacity faster than its fixed obligations accumulate.
Bottom line
The reported 80% BlackRock-managed-fund and 20% Meta ownership structure marks a meaningful shift toward institutionally owned AI infrastructure, with Meta retaining access through a lease 1,4,17,21,25. The model may ease near-term funding pressure 4, but debt and long-term rental commitments can replace upfront capital intensity with fixed contractual and refinancing risk 4,22.
The 1-gigawatt single-campus concentration adds execution, power, utilization, and site-dependence risk 16. The cluster does not establish the lease economics, recourse, accounting treatment, or allocation of cost overruns. Those gaps determine whether the structure is a capital-allocation victory or merely a postponement of risk.
Meta should treat the transaction as strategically supportive but financially conditional. The company must disclose the debt terms, lease obligations, ownership rights, construction milestones, and return thresholds before investors can judge the economics. Institutional capital is available. Control is the prize. The best hedge is ownership—but only when the asset earns more than the financing costs imposed to secure it 12.