Meta just did a $14 billion deal to build an AI data center without it showing up on its own balance sheet, and bond investors are starting to demand more for the risk

[Dialer75]

Meta has structured a roughly $14 billion financing deal with BlackRock to fund a nearly one gigawatt AI data center campus in El Paso, Texas, expected to begin operating in 2028. Under the arrangement, a BlackRock owned vehicle will hold 80% of the project while Meta retains a 20% stake, with BlackRock contributing about $4.9 billion in cash, Meta contributing land and building assets worth roughly $2.3 billion in exchange for a $1 billion payout, and the remainder financed through $12.5 billion in debt. The structure specifically keeps this debt off Meta's own balance sheet, a financing approach increasingly common across the AI industry as hyperscalers try to fund enormous infrastructure buildouts without loading all the risk directly onto their own books

That debt is getting noticeably more expensive to raise though. Bond investors are pushing for yields above 7% on this financing, roughly 0.4 percentage points higher than Meta secured on its record breaking $27 billion Hyperion data center bond sale back in October, with bonds from that earlier deal already trading around 96 cents on the dollar. On a raise this size, even a tenth of a percentage point translates into millions of dollars in additional annual interest, and the gap reflects growing caution among lenders after months of heavy borrowing across the entire sector, alongside a broader selloff in AI linked stocks

The financing lands just before Meta's July 29 earnings release, with investors watching closely for updates on capital spending plans, AI revenue and future infrastructure commitments. Meta has said it plans to invest as much as $600 billion in data centers through 2028, even as its own stock sits down nearly 10% for the year. The pattern fits a broader trend across the industry, data center operators and developers are increasingly seeking large institutional investors and off balance sheet structures specifically because the sheer scale of AI infrastructure spending has outpaced what companies want to fund directly themselves, while investors simultaneously grow more selective about the terms they'll accept for taking on that risk

Metauk.finance.yahoo.com