BDC Weekly

BDC Weekly: How Meta’s AI Data Centers Are Pulling Private Credit Into Infrastructure

AI used to reach private credit through software borrowers. Now it arrives as a gigawatt-scale infrastructure project with billions of dollars of debt.

Last updated: July 22, 2026.

AI used to reach private credit through software borrowers.

Now it arrives as a gigawatt-scale infrastructure project.

BlackRock is leading a debt sale targeting at least $12 billion for a Meta-backed data-center campus in El Paso, Texas. BlackRock-related investors reportedly own 80% of the project. Meta retains 20% and secures the computing capacity. The complex is expected to reach roughly 1 gigawatt.

That structure matters because it turns future demand for computing power into a financeable infrastructure asset. A hyperscaler supplies the demand. Outside investors provide most of the capital. Banks arrange the debt. Long-duration lenders underwrite construction, power, contracts, collateral, and refinancing risk.

The AI race has reached the liability side of the balance sheet.

For BDC investors, the opportunity is real—but so is the need for precision. BlackRock is not a BDC, and a manager-level AI infrastructure deal does not automatically belong to a public BDC. The same distinction applies to Blue Owl, Blackstone, Ares, KKR, Sixth Street, Oaktree, Barings, New Mountain, and Carlyle. The public vehicle owns only what its filings say it owns.

What does the BlackRock-Meta financing mean for BDC investors?

The El Paso transaction shows that AI infrastructure is becoming a major private-market capital-formation business. The reported structure combines an 80/20 ownership split, a project-level entity, hyperscaler demand, and at least $12 billion of debt.

Meta used a similar structure for its Hyperion campus in Louisiana. Blue Owl-led investors reportedly own 80%, Meta owns 20%, and a special-purpose vehicle raised more than $27 billion of debt and about $2.5 billion of equity. That is roughly $11 of debt for every $1 of equity in the reported package.

The Texas terms are not fully public, so investors should not assume the Louisiana leverage ratio applies to El Paso. The important point is the pattern: AI projects are now large enough to require dedicated entities, outside equity partners, bank distribution, private credit, infrastructure capital, and long-duration debt investors.

That can enlarge the opportunity set for alternative-asset managers. It does not make every affiliated BDC a direct AI data-center investment.

The financing model behind Meta’s AI data centers

The two reported structures are easier to understand when ownership and financing are separated.

Project	Ownership structure	Capital structure	Why it matters
El Paso AI campus, Texas	BlackRock-related investors reportedly own about 80%; Meta retains about 20%.	Debt sale targeting at least $12 billion.	Outside investors can provide most of the capital while Meta secures long-term computing capacity without funding the full project itself.
Hyperion campus, Louisiana	Blue Owl-led investors reportedly own about 80%; Meta retains about 20%.	More than $27 billion of project debt plus roughly $2.5 billion of equity through an SPV.	Most of the financing can sit at the project level, making the asset look more like infrastructure finance than ordinary corporate borrowing.

One project can be an exception. Two projects with the same 80/20 ownership split begin to look like a repeatable financing template.

The model lets Meta preserve corporate flexibility while controlling the strategic outcome: access to enormous computing capacity. The outside capital partner receives a majority interest in a long-lived infrastructure asset connected to a hyperscaler. Banks distribute the debt. Institutional lenders underwrite the contracts, construction plan, power supply, collateral, asset value, and refinancing path.

The model is digital. The financing is concrete.

How AI infrastructure moves risk through private markets

Outside capital does not make project risk disappear. It divides the risk among different balance sheets.

Meta creates the demand. Its need for computing capacity helps make the project financeable, but lenders still need durable contracts and enforceable cash flows.

The equity partner absorbs first-loss project risk. Infrastructure and private-market investors commit capital, oversee development, and earn returns if the asset performs. Their equity sits behind the debt when losses arrive.

The project entity borrows. A special-purpose vehicle can isolate the assets, contracts, debt, and cash flows. That protects the sponsor’s broader balance sheet, but it also forces lenders to underwrite the project on its own merits.

Banks distribute the financing need. Large arrangers can connect a single multibillion-dollar project with bond buyers, insurers, private-credit funds, infrastructure-debt investors, and other institutions.

Long-duration investors take the final underwriting test. They must decide whether tenant strength, contract duration, power availability, construction risk, collateral, maturity, and refinancing assumptions justify the return.

This is the hidden private-credit mechanism underneath the AI boom: risk is sliced, assigned, priced, and moved to another balance sheet.

For the full capital stack, read How AI Infrastructure Gets Financed. For the lender map, read Who Finances AI Data Centers?. For the collateral layer, read Asset-Backed Finance And AI Infrastructure.

Why BDC investors should care without direct project exposure

Most public BDCs are not designed to hold a $12 billion data-center financing. Their portfolios are generally diversified across middle-market corporate loans, sponsor-backed companies, venture borrowers, lower-middle-market businesses, structured investments, and occasional equity positions.

The AI infrastructure boom still matters in four ways.

First, it creates a new capital-demand arena for the asset managers around BDCs. Large platforms can route opportunities across infrastructure funds, private-credit funds, insurance accounts, asset-backed strategies, real-estate vehicles, separately managed accounts, and public BDCs.

Second, it increases the value of origination scale. A platform that can underwrite power, construction, leases, equipment, software, services, and corporate borrowers can follow AI spending across the stack rather than depend on one loan product.

Third, it changes competition for institutional capital. A pension fund or insurer allocating billions to contracted data-center debt may demand different spreads from ordinary middle-market loans.

Fourth, it can tempt lenders to stretch. When every borrower claims to be part of the AI buildout, underwriting can drift from cash flow toward narrative. The familiar questions still matter: Who pays? What is the collateral? Is power secured? Can the asset be reused? Who takes the first loss? What happens at refinancing?

A manager can win an extraordinary AI mandate while its public BDC still faces ordinary pressures: lower base rates, spread compression, PIK income, non-accruals, NAV marks, funding costs, and dividend coverage.

BDC earnings preview: ARCC, OBDC, BXSL, MAIN and 10 more companies

Large-platform BDCs: scale helps, but portfolio evidence still matters

Ares Capital (ARCC) reports second-quarter results on July 29. ARCC remains the clearest public benchmark for large-scale middle-market direct lending. Its recent financing actions—including a $1 billion commercial-paper program and an $800 million 5.55% note issuance due 2030—show that capital access is part of its competitive advantage. The next test is whether scale produces attractive originations while lower base rates pressure asset yields. The current verified source set does not establish direct ARCC exposure to the Meta projects.

Blue Owl Capital Corporation (OBDC) has the closest thematic connection because Blue Owl is the majority equity partner in Meta’s Louisiana project. That is a platform-level fact, not proof that OBDC owns the project debt or equity. OBDC’s first-quarter adjusted net investment income of $0.31 per share matched its reset base dividend of $0.31. Investors should watch cash coverage, portfolio marks, repurchases, non-accruals, and whether new assets offset lower base rates and tighter spreads when OBDC reports on August 5.

Blackstone Secured Lending (BXSL) reports on August 6. Blackstone’s broader platform participates across infrastructure, real estate, insurance, and credit, but BXSL remains a senior-secured BDC. The useful metrics are first-lien performance, NAV stability, dividend coverage, and new-lending economics. No direct ownership of the Meta project financings has been established.

FS KKR Capital (FSK) enters earnings season with less room for thematic distraction. First-quarter NAV fell to $18.83 per share from $20.89, net debt-to-equity rose to 131% from 122%, and the second-quarter distribution was set at $0.42 per share. Its August 6 report should be judged on NAV, non-accruals, realized losses, leverage, portfolio exits, and the durability of the reset payout.

Carlyle Secured Lending (CGBD) reports on August 6. Carlyle has broad global-credit and infrastructure capabilities, but CGBD remains a senior-secured middle-market lender. Investors should focus on origination spreads, regular-dividend coverage, and whether credit quality supports NAV.

Underwriting comparators: where AI may appear through borrowers

Hercules Capital (HTGC) reports on July 30. HTGC is the covered BDC most naturally connected to the innovation economy through venture-growth lending. That is different from financing a hyperscale data center. Watch commitments, fundings, early repayments, realized gains, non-accruals, and whether venture demand improves without weaker underwriting.

Sixth Street Specialty Lending (TSLX) reports on August 4. TSLX reduced its base dividend to $0.42 per share after first-quarter net investment income also came in at $0.42. The next quarter is a direct test of recurring cash coverage. Watch NAV, non-accruals, fee income, repayments, and whether new spreads rebuild the dividend cushion.

Golub Capital BDC (GBDC) reports on August 3. Golub’s portfolio remains heavily tilted toward first-lien, sponsor-backed lending. The next report should show whether middle-market activity translated into stronger originations and fee income or merely more competition. Software exposure deserves special attention because AI can strengthen some borrowers while disrupting others.

Main Street Capital (MAIN) provided the strongest early operating read. Preliminary second-quarter NII was $0.95 to $0.99 per share, distributable NII was $1.02 to $1.06, and estimated NAV was $33.88 to $33.96, up 1.2% to 1.5% after a $0.30 supplemental dividend. Non-accruals were estimated at 1.1% of fair value. MAIN reports final results on August 6. Its broader lesson is that control investments and equity gains can reinforce income when direct-lending spreads become less generous.

Capital Southwest (CSWC) estimated fiscal first-quarter 2027 pretax NII of $0.57 to $0.58 per share, NII of $0.58 to $0.59, and NAV of $16.55 to $16.65. Final results arrive August 3. Its internally managed model puts capital allocation at the center: can CSWC protect the base dividend, preserve NAV, and keep supplemental payouts tied to earned income?

Funding and repair watchlist: the ordinary credit cycle still matters

Prospect Capital (PSEC) continued using its InterNotes program, offering fixed-rate maturities with coupons from 6.00% to 6.50%. Retail debt access provides flexibility, but it is not cheap capital. Investors should compare liability costs with asset yields, credit losses, NAV movement, and the cash quality of dividend coverage.

Oaktree Specialty Lending (OCSL) reports on August 5. OCSL’s public test remains NAV trust and portfolio repair. Watch non-accruals, amendments, realized losses, PIK income, and whether Oaktree’s credit discipline is producing better marks and stronger recurring earnings.

New Mountain Finance (NMFC) reports on August 3. Its defensive-industry positioning should help if growth weakens, but labels are not substitutes for borrower performance. Watch cash coverage, PIK, non-accruals, portfolio concentration, and repayments.

Barings BDC (BBDC) reports on August 5. BBDC must show progress through NAV, non-accruals, portfolio exits, new originations, and dividend coverage. The current verified source set does not establish direct participation in the Meta financings.

The common thread across all 14 names is not direct AI exposure. It is balance-sheet discipline while the managers around them pursue a much larger opportunity set.

BDC earnings calendar: what investors should watch

The Federal Reserve meets on July 28-29. That matters because BDC assets are predominantly floating-rate while much of their funding is fixed-rate or slower to reset. Lower short-term rates can reduce portfolio income before all funding costs adjust.

Date	Companies or event	Key metrics to watch	Why investors should care
July 28-29	Federal Reserve meeting	Policy rate and forward guidance	The rate path determines how quickly floating-rate asset income may compress.
July 29	ARCC	NII, portfolio yield, funding costs, originations	ARCC will show whether scale and funding flexibility can offset lower base rates and tighter spreads.
July 30	HTGC	Commitments, fundings, non-accruals, prepayments	The report will test whether venture-credit demand is improving without a parallel rise in credit stress.
August 3	CSWC, GBDC, NMFC	Dividend coverage, NAV, portfolio yields, leverage	These reports provide an early read on lower-middle-market and sponsor-backed lending conditions.
August 4	TSLX	NII, fee income, repayments, non-accruals	Investors will see whether underwriting discipline is rebuilding the dividend cushion.
August 5	OBDC, BBDC, OCSL	Cash coverage, NAV, PIK, realized losses	Large platforms and repair cases will show whether recurring income and portfolio marks are stabilizing.
August 6	BXSL, FSK, CGBD, MAIN	Credit quality, leverage, originations, dividend sustainability	This is the clearest cross-sector comparison of scale, balance-sheet pressure, and underwriting performance.
August 7	MAIN and CGBD calls	Management commentary and portfolio detail	The calls will test whether preliminary strength and stronger origination terms hold up under scrutiny.

The Fed sets the price of short-term money. The BDCs show how that price moves through loan yields, funding costs, NAV, credit quality, and dividends.

Five questions that will define BDC earnings season

The first is whether wider origination spreads can offset lower base rates. A lower all-in yield can still be attractive if the spread, covenants, collateral, and documentation improve.

The second is whether funding access is a competitive weapon or an expensive necessity. ARCC’s commercial paper, FSK’s leverage, and PSEC’s retail notes all sit on the same spectrum. Access matters. Cost matters more.

The third is whether base dividends are covered by recurring cash income. Reported NII should be read alongside PIK income, fee income, realized gains, and the distinction between regular and supplemental dividends.

The fourth is whether NAV and non-accruals confirm the earnings story. Stable NII can coexist with weakening marks, rising amendments, or more PIK. The strongest reports will show cash coverage, contained non-accruals, credible marks, and enough liquidity to remain patient.

The fifth is whether any BDC discloses real AI-infrastructure exposure. Investors should look for named borrowers, industries, collateral, leases, data-center services, power, cooling, equipment, or software exposures. Manager-level headlines are not portfolio evidence.

For the core BDC dashboard, read BDCs: The Public Door Into Private Credit. For the broader system, read Private Credit. For the next stress test, read The Private Credit Refinancing Wall and Private Credit’s Discipline Cycle Has Started.

Investor quick answers

Does OBDC own Meta’s Louisiana data-center financing?

The current verified source set does not establish that OBDC owns the Louisiana project debt or equity. The reported transaction involved a Blue Owl-led project structure. Blue Owl’s manager-level activity and OBDC’s public portfolio are different things unless OBDC filings connect them.

Does BXSL own BlackRock’s Meta data-center financing?

No connection has been established. BXSL is managed within the Blackstone platform, while the El Paso transaction is reported as a BlackRock-led financing. BlackRock and Blackstone are separate firms.

Are BDCs direct investments in AI data centers?

Usually not. BDCs may lend to software companies, equipment providers, infrastructure services, power-adjacent businesses, data-center suppliers, and other private borrowers touched by AI spending. Their value is as a credit window, not as pure data-center ownership.

Why is Meta using outside capital for data centers?

Outside capital lets Meta secure computing capacity while sharing ownership, construction exposure, and financing needs with infrastructure investors. A project entity can connect debt to specific assets and cash flows rather than place all borrowing directly on Meta’s corporate balance sheet.

What is the biggest risk in AI infrastructure private credit?

The biggest risk is financing the AI narrative faster than contracts, power, construction, collateral, and cash flows can support it. If assumptions miss, pressure can appear through cost overruns, delayed completion, weaker asset values, refinancing risk, portfolio marks, or credit losses.

What comes next

This earnings season is about more than quarterly numbers. It will show whether BDCs can protect income as interest rates normalize while their parent platforms pursue a new generation of infrastructure-scale credit opportunities.

BlackRock’s Meta financing is important because it makes the capital shift visible. AI infrastructure is moving beyond ordinary corporate capex and into project companies, outside equity partnerships, bank-led debt sales, private-credit allocations, insurance portfolios, and long-dated assets tied to compute demand.

That creates a larger opportunity for private markets—and a new way for public investors to get confused.

The manager may touch the project. The public BDC may not.

The winning platforms will route each risk to the right balance sheet. The winning BDCs will refuse to let an extraordinary theme weaken ordinary credit discipline.

AI is pulling private credit toward infrastructure-scale ambition. The next earnings wave will show which public lenders can keep their underwriting feet on the ground.

Source notes

This analysis draws on The Drift’s BDC Weekly research packet for the seven days ended July 20, 2026; reporting from The Wall Street Journal and Reuters on Meta’s El Paso and Louisiana financing structures; Federal Reserve calendars; issuer press releases and SEC filings; and The Drift’s AI infrastructure, private-credit, and BDC coverage.

Primary sources and current references:

Disclosure

The Drift is published by Drift Research LLC for informational and educational purposes only. Nothing published by The Drift constitutes personalized investment advice, financial advice, tax advice, legal advice, or a recommendation to buy, sell, or hold any security. The Drift is not a registered investment adviser, broker-dealer, or financial planner. All investments involve risk, including possible loss of principal.