Faced with Oracle’s massive AI infrastructure financing needs, Wall Street banks’ balance sheets are nearing their limits. To mitigate risk exposure and free up credit capacity to continue lending, banks are eager to securitize and sell hundreds of billions of dollars in loans related to Oracle data center projects to insurance companies and private credit funds through “securitization” ratings.
Sources familiar with the matter say that at least $56 billion worth of data center construction loans have received investment-grade ratings, supported by future lease income from a $300 billion deal between Oracle and OpenAI. Achieving investment-grade ratings for infrastructure loans still in construction is extremely rare, allowing banks to bring in insurance and private credit funds that typically avoid such risks.
While rating packaging has opened new funding channels, the market has not fully embraced it. As banks rush to sell debt to reduce overexposure to the AI financing boom, related financing costs are rising sharply. Some investors are holding back, expecting higher returns in the future, while borrowing costs for new projects have widened to nearly junk bond levels.
Meanwhile, Oracle is not slowing its aggressive expansion. According to Bloomberg, the company plans to raise $50 billion through bonds and equity refinancing by 2026. This “heavy asset” expansion strategy amid huge debt pressure has not only made banks nervous but also attracted attacks from well-known short seller Michael Burry.
Bank Limits Reached: Borrowing Everywhere Possible
In traditional project financing, banks usually hold infrastructure loans for highways or airports themselves. However, the recent scale of AI data center projects has completely overwhelmed banks’ conventional capacity. Tech giants are urgently seeking new sources of capital, while banks must clean up old loans before continuing to lend.
“We’ve basically knocked on every project finance bank’s door, but the number of banks is limited,” said a banker familiar with Oracle’s financing activities. “If banks want to keep lending, they have to unload these risks.”
This urgency has prompted banks to push rating agencies to rate loans in construction. Insiders say that the $56 billion in loans covers $38 billion worth of data center facilities being built by Oracle in Texas and Wisconsin, as well as a $1.8 billion data center park in New Mexico supported by Blue Owl Capital. Both loans are currently being marketed to investors.
Rating “Reform” and Insurance Capital Entry
Securing investment-grade ratings for projects in construction is seen as a “transformative” move in the industry. It effectively opens a new institutional funding pool—mainly insurance companies and pension funds—that previously avoided non-operational assets due to high risk.
Christine Brozynski, a partner in infrastructure project finance at law firm Norton Rose Fulbright, said that while obtaining credit ratings during construction was once extremely rare in the data center sector, it is now “becoming common.” She noted that almost all large data center deals are now attempting to obtain credit ratings.
In the latest deal structure, over a dozen banks have provided loans secured by Oracle’s long-term lease commitments. STACK Infrastructure is responsible for the New Mexico data center development, and the company confirmed that the deal is currently in syndication and has received an investment-grade credit rating, with progress in line with expectations.
Crowded Market and Rising Financing Costs
Despite passing the rating threshold, concerns about Oracle’s aggressive AI spending commitments and its debt pile are intensifying among investors.
According to a research report from TD Cowen published on January 26, although the current deal is priced at SOFR plus 2.5 percentage points, the borrowing costs for newer Oracle-related data center projects not yet sold to investors have widened to SOFR plus 3 to 4.5 percentage points—approaching the pricing of junk bonds.
Some investors are hesitant about the two Oracle-backed syndicated loans on the market, expecting higher-yield assets to be released in the future. “The elephant in the room is—whether there is enough appetite to invest in these notes, after all, products with higher returns might come out in two weeks?” said a senior project finance banker in the US.
Another investor who purchased bonds from other data center projects noted that banks are nervous about their increasing AI financing exposure and are trying to offload the debts they have committed to. To close deals, banks have had to offer higher-than-expected interest rates to buyers.
Aggressive Expansion: $50 Billion New Financing Plan
While banks are eager to find “new owners,” Oracle has not slowed its capital expenditure. Bloomberg reported that in a statement on February 1, Oracle announced plans to raise up to $50 billion through bonds and equity financing by 2026 to meet cloud infrastructure demands from major clients like AMD, Meta, Nvidia, OpenAI, TikTok, and xAI.
According to the plan, Oracle will raise about half of the funds through issuing mandatory convertible preferred securities and a $20 billion market equity plan, with the rest to be raised via bonds in early 2026. This will further increase the company’s debt burden. Bloomberg data shows Oracle currently carries about $95 billion in debt, making it one of the largest corporate bond issuers outside the financial sector.
Oracle stated that this move is to “build additional capacity” and pledged to maintain its investment-grade rating by keeping its debt load within manageable limits.
Short Seller Attacks: Fragile “AI Bubble Carrier”
Oracle’s aggressive shift from a “light asset” software company to a “heavy asset” cloud infrastructure provider, along with the resulting deterioration of its balance sheet, has raised market concerns.
According to Wall Street Insights, Michael Burry, the real-life inspiration for the movie “The Big Short,” recently disclosed that he has shorted Oracle. He criticized the company for engaging in “unnecessary heavy asset expansion,” attempting to compete with cloud giants through expensive data center construction, and called it a “pure AI bubble carrier.”
Burry pointed out that unlike tech giants like Microsoft, Alphabet, and Meta, which have strong core businesses, Oracle lacks sufficient safety margins. Its high-risk transformation under heavy debt makes its financial structure particularly fragile. If AI demand falls short of expectations, Oracle’s very low tolerance for errors could pose a significant survival risk.
Risk Warning and Disclaimer
Market risks exist; investments should be cautious. This article does not constitute personal investment advice and does not consider individual users’ specific investment goals, financial situations, or needs. Users should consider whether any opinions, views, or conclusions in this article are suitable for their particular circumstances. Invest accordingly at your own risk.
View Original
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
Afraid of borrowing! Banks are rushing to find a buyer for Oracle's $56 billion debt claim, with insurance funds and private equity stepping in.
Faced with Oracle’s massive AI infrastructure financing needs, Wall Street banks’ balance sheets are nearing their limits. To mitigate risk exposure and free up credit capacity to continue lending, banks are eager to securitize and sell hundreds of billions of dollars in loans related to Oracle data center projects to insurance companies and private credit funds through “securitization” ratings.
Sources familiar with the matter say that at least $56 billion worth of data center construction loans have received investment-grade ratings, supported by future lease income from a $300 billion deal between Oracle and OpenAI. Achieving investment-grade ratings for infrastructure loans still in construction is extremely rare, allowing banks to bring in insurance and private credit funds that typically avoid such risks.
While rating packaging has opened new funding channels, the market has not fully embraced it. As banks rush to sell debt to reduce overexposure to the AI financing boom, related financing costs are rising sharply. Some investors are holding back, expecting higher returns in the future, while borrowing costs for new projects have widened to nearly junk bond levels.
Meanwhile, Oracle is not slowing its aggressive expansion. According to Bloomberg, the company plans to raise $50 billion through bonds and equity refinancing by 2026. This “heavy asset” expansion strategy amid huge debt pressure has not only made banks nervous but also attracted attacks from well-known short seller Michael Burry.
Bank Limits Reached: Borrowing Everywhere Possible
In traditional project financing, banks usually hold infrastructure loans for highways or airports themselves. However, the recent scale of AI data center projects has completely overwhelmed banks’ conventional capacity. Tech giants are urgently seeking new sources of capital, while banks must clean up old loans before continuing to lend.
“We’ve basically knocked on every project finance bank’s door, but the number of banks is limited,” said a banker familiar with Oracle’s financing activities. “If banks want to keep lending, they have to unload these risks.”
This urgency has prompted banks to push rating agencies to rate loans in construction. Insiders say that the $56 billion in loans covers $38 billion worth of data center facilities being built by Oracle in Texas and Wisconsin, as well as a $1.8 billion data center park in New Mexico supported by Blue Owl Capital. Both loans are currently being marketed to investors.
Rating “Reform” and Insurance Capital Entry
Securing investment-grade ratings for projects in construction is seen as a “transformative” move in the industry. It effectively opens a new institutional funding pool—mainly insurance companies and pension funds—that previously avoided non-operational assets due to high risk.
Christine Brozynski, a partner in infrastructure project finance at law firm Norton Rose Fulbright, said that while obtaining credit ratings during construction was once extremely rare in the data center sector, it is now “becoming common.” She noted that almost all large data center deals are now attempting to obtain credit ratings.
In the latest deal structure, over a dozen banks have provided loans secured by Oracle’s long-term lease commitments. STACK Infrastructure is responsible for the New Mexico data center development, and the company confirmed that the deal is currently in syndication and has received an investment-grade credit rating, with progress in line with expectations.
Crowded Market and Rising Financing Costs
Despite passing the rating threshold, concerns about Oracle’s aggressive AI spending commitments and its debt pile are intensifying among investors.
According to a research report from TD Cowen published on January 26, although the current deal is priced at SOFR plus 2.5 percentage points, the borrowing costs for newer Oracle-related data center projects not yet sold to investors have widened to SOFR plus 3 to 4.5 percentage points—approaching the pricing of junk bonds.
Some investors are hesitant about the two Oracle-backed syndicated loans on the market, expecting higher-yield assets to be released in the future. “The elephant in the room is—whether there is enough appetite to invest in these notes, after all, products with higher returns might come out in two weeks?” said a senior project finance banker in the US.
Another investor who purchased bonds from other data center projects noted that banks are nervous about their increasing AI financing exposure and are trying to offload the debts they have committed to. To close deals, banks have had to offer higher-than-expected interest rates to buyers.
Aggressive Expansion: $50 Billion New Financing Plan
While banks are eager to find “new owners,” Oracle has not slowed its capital expenditure. Bloomberg reported that in a statement on February 1, Oracle announced plans to raise up to $50 billion through bonds and equity financing by 2026 to meet cloud infrastructure demands from major clients like AMD, Meta, Nvidia, OpenAI, TikTok, and xAI.
According to the plan, Oracle will raise about half of the funds through issuing mandatory convertible preferred securities and a $20 billion market equity plan, with the rest to be raised via bonds in early 2026. This will further increase the company’s debt burden. Bloomberg data shows Oracle currently carries about $95 billion in debt, making it one of the largest corporate bond issuers outside the financial sector.
Oracle stated that this move is to “build additional capacity” and pledged to maintain its investment-grade rating by keeping its debt load within manageable limits.
Short Seller Attacks: Fragile “AI Bubble Carrier”
Oracle’s aggressive shift from a “light asset” software company to a “heavy asset” cloud infrastructure provider, along with the resulting deterioration of its balance sheet, has raised market concerns.
According to Wall Street Insights, Michael Burry, the real-life inspiration for the movie “The Big Short,” recently disclosed that he has shorted Oracle. He criticized the company for engaging in “unnecessary heavy asset expansion,” attempting to compete with cloud giants through expensive data center construction, and called it a “pure AI bubble carrier.”
Burry pointed out that unlike tech giants like Microsoft, Alphabet, and Meta, which have strong core businesses, Oracle lacks sufficient safety margins. Its high-risk transformation under heavy debt makes its financial structure particularly fragile. If AI demand falls short of expectations, Oracle’s very low tolerance for errors could pose a significant survival risk.
Risk Warning and Disclaimer
Market risks exist; investments should be cautious. This article does not constitute personal investment advice and does not consider individual users’ specific investment goals, financial situations, or needs. Users should consider whether any opinions, views, or conclusions in this article are suitable for their particular circumstances. Invest accordingly at your own risk.