The BlackRock Tell: Why the $12 Billion El Paso Deal Marks the Moment AI Shadow Debt Went Mainstream

John Grey — Private credit and macro desk

BlackRock, JPMorgan and Morgan Stanley began pitching bond investors on Monday for a debt package of more than $12 billion to fund a Meta-anchored hyperscale data center in El Paso, Texas, with pricing expected early next week. The structure is now the familiar one. A joint venture holds the physical asset. BlackRock's Global Infrastructure Partners and HPS Investment Partners funds sit at 80 percent equity. Meta sits at 20 percent, retains operational control, and signs a long-term operating lease that backstops the SPV debt with a residual value guarantee. The debt does not appear on Meta's consolidated balance sheet. It sits in the SPV, and the SPV is now being funded in the public bond market at scale. This is the second $10-billion-plus off-balance-sheet AI infrastructure vehicle to price inside 30 days, following the $27 billion Blue Owl-Meta Hyperion joint venture that syndicated in June and July. Consensus is reading the deals as validation of the private-credit AI thesis. The correct read is the opposite. When the largest asset manager on Earth is willing to arrange the mainstream distribution of what the Bank for International Settlements has explicitly labeled "shadow borrowing," the structure has stopped being novel and started being systemic. That is the tell.

The scale deserves a slower reading. On Tuesday, Nikkei published a study putting the combined off-balance-sheet debt of Alphabet, Microsoft, Amazon, Meta and Oracle at roughly $1.65 trillion, up eightfold since 2022 and now larger than the $1.35 trillion those companies report as on-balance-sheet debt. Morgan Stanley's estimate puts the total off-balance-sheet AI exposure across the industry at $1.8 trillion, decomposed into nearly $1 trillion in purchase commitments, more than $800 billion in leases that have not yet begun, and roughly $110 billion in accounts-payable financing. Aggregate hyperscaler leverage ratios have doubled, from 0.9x to 1.8x, and annual AI-related debt issuance is now on pace to exceed $570 billion a year. Those are the numbers of a structural rewiring of investment-grade credit around a single thematic bet, executed largely through structures that do not appear on the balance sheets of the entities most exposed to the bet.

What the BIS has said, and how it has been read

The most important primary-source reading on this comes from the BIS Annual Economic Report 2026, released on June 28, which placed AI investment sustainability on its short list of interlocking pressures to global financial stability, alongside persistent inflation, sovereign debt fragility and financial system vulnerability. The report's language on the financing structures is unusually direct for a central bank of central banks.

"Disappointments in returns could trigger a sudden pullback in financing and turn the capex boom into a protracted investment bust, with potential knock-on effects on financial conditions," the BIS wrote. On the specific structures, it added: "The terms of such deals are typically poorly disclosed, with risks of the same asset being pledged multiple times."

The report also made the transmission-mechanism argument explicit through Zhang Tao, the BIS chief representative for Asia and the Pacific, who told the South China Morning Post that a correction could unwind "much faster than previous banking crisis episodes" precisely because so much of the financing now runs through hedge funds, private credit vehicles and non-bank intermediaries carrying less regulatory oversight than traditional banks. The BIS Bulletin No. 128, released July 14 and authored by Fernando Avalos, Giulio Cornelli and Egemen Eren, added the specific data-point that business development companies alone had lent approximately $115 billion to software firms, about one fifth of all BDC lending and more than 80 percent of their fast-growing technology portfolios.

Read as a coordinated intervention, the BIS is telling the market three things. Leverage in the AI supply chain is systematically understated because of off-balance-sheet structures. The transmission mechanism from AI capex disappointment to broader stress runs through non-bank channels that regulators cannot easily see or slow. And the correct policy response, disclosure, is currently absent. None of those three has been priced into either equity valuations or IG credit spreads.

The sell-side hedge that clarifies the risk

The most useful sell-side reading of the situation is Morgan Stanley Investment Management's July 20 note "AI Dispersion in Credit," authored by its Broad Markets Fixed Income Team. The framing is instructive because it works to reconcile the constructive fundamental view with the structural technical concern, and in doing so it lays out the specific inflection points that the market should be watching.

"In that sense, AI is as much a capital allocation story as it is a technology story, and for investment grade (IG) credit, the funding dynamics are likely to matter sooner than the margin benefits," Morgan Stanley wrote. On the near-term impact: "Despite the constructive long-term backdrop, the AI investment cycle is likely to remain modestly credit-negative in the near term until investors see clearer evidence of returns and stabilization in credit metrics." And on the technical setup: "Persistent supply may contribute to sector-level spread pressure and greater sensitivity to capex guidance, particularly if issuance outpaces incremental demand."

Note the specific claim in the third quote. Sector-level spread pressure and greater sensitivity to capex guidance are the two symptoms that would signal the market beginning to reprice AI-linked IG credit for the actual leverage, not the reported leverage. Morgan Stanley's own colleagues on the sell-side desk, in a note picked up by Forbes on July 17, made the point more directly. Investment-grade bond order books for hyperscaler paper have compressed from nearly five times cover in February to less than double by July. Demand is fatiguing. The BlackRock El Paso deal will be a real-time test of where the marginal buyer of AI-linked SPV paper sits, and at what spread. If the deal prices wide of guidance, or breaks in the aftermarket, the market has begun the repricing. If it prices tight and trades well, the private-credit and IG desks will structure ten more of these deals through year-end, and the systemic exposure will keep growing.

The PIMCO frame, and why it matters more than it reads

The other primary-source reading worth weight is PIMCO's May 2026 credit analysis report, which anchors the current bull case for the structure among the largest institutional buyers of the paper. PIMCO's framing is careful and worth quoting in full.

"AI is in the midst of a capex boom with genuine risks: uncertain monetization, potential overbuild, shortening asset lives, and growing reliance on debt. But for now, it is a more disciplined and far more financeable cycle than the late-1990s telecom boom," PIMCO wrote.

Two features of that quote matter. First, PIMCO acknowledges every substantive risk the BIS has flagged: uncertain monetization, potential overbuild, shortening asset lives, growing debt reliance. It is not defending the fundamentals of the AI capex bet. Second, the comparison to the telecom boom is the correct historical parallel, and it is not benign. The 2000-2002 telecom bust was a private-credit and IG-credit event as much as an equity event, destroying roughly $2 trillion in market capitalization and producing a recession that took the Fed to zero. PIMCO's argument that this cycle has better collateral discipline is the same argument every large institutional buyer of subprime mortgage paper made in 2006 about the improvements over the 1994 mortgage cycle.

The El Paso deal in three data points

The BlackRock deal isolates the market signal in three data points. First, the deal is anchored by BlackRock's own Global Infrastructure Partners and HPS Investment Partners funds, both acquired within the last two years for a combined $24.5 billion of purchase price. BlackRock is now using those platforms to originate what the market will treat as vehicle-agnostic AI infrastructure paper. Second, the arrangement replicates the Blue Owl-Meta Hyperion structure almost line-for-line, which suggests the structure is now standardized and can be issued in size at ten or more sites over the next 24 months. Third, JPMorgan strategists estimated in June that hyperscalers will spend roughly $5.5 trillion on AI infrastructure through 2030, much of it borrowed. Even under conservative funding-mix assumptions, that pipeline implies $2 trillion or more of new off-balance-sheet issuance through the end of the decade. The BlackRock deal is not a one-off. It is the second scaled instance of what will become the dominant funding structure for the AI infrastructure buildout.

Our view

The correct positioning for this regime is not short AI equity. AI equity has fundamentals, cash flows and monopoly dynamics that reward optionality. The correct positioning is defensive on IG credit, and specifically on AI-linked hyperscaler paper and the utility-issuance complex that has doubled in tandem. Morgan Stanley's projection of $130 to $150 billion in incremental hyperscaler IG issuance in 2026, up 30 to 50 percent from 2025's $100 billion print, will meet a rates market that is not currently pricing spread widening. As demand for the paper compresses from five-times to less-than-two-times cover, and as new issuance continues to hit at scale, the marginal price becomes wider spreads. Long-duration hyperscaler IG is the specific expression. Utility IG is the second-order expression, running at the record $145 billion 2026 issuance pace Morgan Stanley has quantified. Both should widen relative to broader IG. Both are underweights.

The larger structural point is that the AI capex cycle has now completed the transition every prior structural credit cycle has completed at exactly this stage. It began as an equity story. It moved to a corporate bond story. It has now moved to a private-credit and off-balance-sheet story that is being distributed to the mainstream IG market by the largest asset manager on Earth. Every prior cycle that reached this stage produced an eventual repricing, and the repricing was always concentrated in the credit structures rather than in the underlying equity thesis. This one will follow the same pattern. The BlackRock El Paso deal is the tell. The trade is to be positioned for the credit repricing, not the equity one.

This note reflects the views of Solomon Grey Capital's Private credit and macro desk as of the date of publication and is provided for informational purposes only. It does not constitute investment advice, an offer, or a solicitation to buy or sell any security. Past performance is not indicative of future results.