Off-balance-sheet financing of $120 billion! Tech giants team up with Wall Street to navigate AI infrastructure, with risks shifting towards private credit

In order to maintain a lead in this expensive artificial intelligence (AI) arms race while keeping perfect financial statements, Silicon Valley tech giants are using complex financial instruments to shift massive infrastructure spending off their balance sheets.

According to an analysis by the Financial Times on the 24th, tech companies including Meta, xAI, Oracle, and data center operator CoreWeave have utilized special purpose vehicles (SPVs) to transfer over $120 billion in data center financing debt to Wall Street investors. This strategy not only protects the credit ratings of tech giants but also raises concerns in the market about risk opacity and potential financial contagion effects.

In this financing feast, financial institutions such as Pimco, BlackRock, Apollo, Blue Owl Capital, and banks like JPMorgan have injected at least $120 billion into computing infrastructure projects operating through the SPV structure in the form of debt and equity. This innovative financing structure allows companies like Meta and Oracle to secure the massive funds needed to build AI data centers without significantly increasing their own on-balance-sheet debt. This practice was nearly unimaginable 18 years ago but has now become the industry norm.

However, this financing frenzy may obscure the actual risks borne by tech groups. Although the SPV structure nominally isolates the debt, it remains uncertain who will ultimately bear the cost if AI demand falls short of expectations. Market participants are concerned that if AI operators face financial pressure in the future, this pressure could spread unpredictably through the SPV structure to Wall Street and even the private credit market. Morgan Stanley previously estimated that tech companies' AI plans would require about $1.5 trillion in external financing, suggesting that the scale of such financing models may continue to expand.

The "Perfect" Balance Sheet Game

Silicon Valley giants have traditionally been known for their ample cash flow and extremely low debt, which gives them excellent credit ratings and investor trust. However, the competition for advanced AI computing power has forced these tech groups to bear unprecedented borrowing pressure. To avoid showing excessive leverage on their balance sheets, thereby protecting credit ratings and beautifying financial metrics, introducing private capital through off-balance-sheet structures has become the preferred choice.

According to reports, a senior executive at a large financing institution stated that due to their strong credit status, the tech industry can obtain more capital than any other sector. The basic logic of this structure is: tech companies do not borrow directly but raise funds through SPVs to build data centers, and then sign lease agreements with them. In the event of default, lenders' recourse rights are typically limited to the assets under the SPV—data centers, land, and chips—rather than the parent tech company managing these sites.

Specific Maneuvering Paths of the Giants

Meta completed a representative transaction last October. By partnering with New York financing company Blue Owl Capital to establish an SPV named "Beignet Investor," Meta raised $30 billion for its planned Hyperion facility in Louisiana, including about $27 billion in loans from institutions such as Pimco, BlackRock, and Apollo This transaction effectively allowed Meta to borrow $30 billion without showing any debt on its balance sheet, paving the way for its refinancing of $30 billion in the corporate bond market a few weeks later.

Oracle is also actively leveraging third parties to construct large debt transactions. The tech group led by Larry Ellison has reached multiple agreements with partners such as Crusoe and Blue Owl Capital. Among them, Blue Owl and JP Morgan invested approximately $13 billion in the SPV holding its Texas data center, which includes $10 billion in debt financing. Additionally, the company arranged two bundled financings for several data center projects located in Texas, Wisconsin, and New Mexico, with scales of approximately $38 billion and $18 billion, respectively. In these cases, Oracle agreed to lease facilities from the SPV.

Moreover, Musk's startup xAI has adopted a similar structure as part of its $20 billion financing, which includes up to $12.5 billion in debt. The SPV will use these funds to purchase Nvidia graphics processing units (GPUs) and lease them to xAI.

Concerns in the Private Credit Market

As private capital investors are eager to participate in the AI boom, there has been a surge in "project financing" deals focused on long-term infrastructure financing. According to UBS data, by early 2025, tech companies had borrowed approximately $450 billion from private equity funds, an increase of $100 billion year-on-year. So far this year, about $125 billion has flowed into project financing deals like those of Meta and Blue Owl.

This trend has intensified concerns in the private credit industry, which has rapidly ballooned to $1.7 trillion, primarily focused on soaring asset valuations, lack of liquidity, and borrower concentration risk. A banker close to data center financing transactions pointed out that risky loans and potential credit risks have already accumulated in private credit, as the AI data center boom largely relies on a few clients like OpenAI (which has made over $1.4 trillion in long-term computing commitments in the industry). If major tenants encounter issues, lenders across multiple data centers may face shared risk exposure.

Additionally, these investors face uncertainties in power acquisition, risks from changes in AI regulation, or hardware obsolescence due to technological iterations. Wall Street has even begun attempting to securitize AI debt, packaging loans for sale to a broader range of investors (such as asset management companies and pension funds), with the current scale of such transactions amounting to several billion dollars.

Risk Exposure and Differentiated Strategies

Although the financing structure is designed to isolate risks, in many cases, if demand for AI services declines and facility values are impaired, the ultimate financial risks often still fall on the tech companies leasing the sites. For example, in the "Beignet Investor" case, Meta owns 20% of the SPV and provided other investors with a "residual value guarantee." This means that if the data center's value falls below a certain level and Meta decides not to renew the lease, it must repay the SPV investors' funds Matthew Mish, head of public and private credit strategies at UBS, pointed out that while most investors believe that ultimately owning "hyperscalers" is a good thing, SPV financing actually increases the outstanding liabilities of tech companies, meaning that the overall credit quality of these enterprises may be worse than current models suggest.

Not all giants have joined the trend of off-balance-sheet financing. Companies like Google, Microsoft, and Amazon, which had established data center businesses before the AI boom, are still primarily financing their construction through cash or direct bond issuance and have not disclosed any significant SPV financing plans. The divergence in market strategies reflects different considerations for risk control among players when facing the expensive AI gamble