The artificial intelligence industry is rapidly amassing capital using an unexpected but highly effective tool: corporate bonds. Based on my analysis of market flows, this sector is turning into a giant debt bubble, which, however, is currently working in favor of the largest players. By 2026, the volume of "AI debt" placements could reach $570 billion, with Morgan Stanley becoming the main beneficiary of this process.
At the center of attention are Nvidia and the new Chinese neural network Kimi K3 from Moonshot AI. The financing of the technological breakthrough is being provided by pension funds and insurance organizations, operating without much fanfare. They are eagerly buying up debt securities structured around future contracts for computing power.
How Morgan Stanley Outperforms Competitors
The volume of placed "AI debt" by the end of May reached $236 billion — four times more than in the same period last year. In late 2025 alone, Morgan Stanley executed bond issuance deals for AI companies worth $65 billion. The result: $2.3 billion in fees over six months, compared to no more than $1.4 billion a year earlier. Thanks to this surge, the bank surpassed Goldman Sachs and now trails only JPMorgan Chase.
Financiers are structuring credit obligations of IT giants and contracts for computing power. The ready bonds are eagerly bought by conservative investors who previously avoided high-risk technology projects.
Google's Support and Meta's "Hidden" Debts
Former bitcoin miner TeraWulf clearly demonstrates how this scheme works. The company repurposed its capacities for neural network needs. A $3.2 billion debt securities placement attracted $10 billion in bids, with a yield of 7.75% per annum. The high demand is explained by Google's involvement: the tech giant guaranteed $3.2 billion in lease payments, receiving in return the right to purchase 14% of TeraWulf's shares.
A similar deal was struck by Cipher Mining. Shares of mining companies are growing faster than cryptocurrencies. Meta chose large-scale projects: with Morgan Stanley's assistance, the company raised $27 billion for the Hyperion complex in Louisiana. The deal set a record for the private credit market. Partner Blue Owl's share is 80%, so the liabilities do not appear on Meta's balance sheet.
Investors Begin Demanding Payment for Patience
Buyers of debt securities are showing restraint. In February, refinancing by IT giants exceeded supply by five times. By July, the ratio had fallen below the two-fold level. In late 2025, the cost of insurance against Oracle's default rose to its highest since 2009. Market participants' concerns are growing alongside talk of the industry overheating.
Infrastructure costs continue to rise. Experts predict new large-scale injections. Required investments in data centers until 2028: $2.9 trillion. IT companies' own funds cover only 50% of needs. The remaining portion will have to be raised through debt capital.
The development of railways and telecommunications in past eras relied on bonds. The debt market is financing the advancement of neural networks. Every microchip operates on credit funds. The cost of borrowing will determine the ultimate speed of technology adoption.
Expert commentary from Cryptalist: We are witnessing a classic "tech bubble" cycle financed by debt. As long as rates are low and demand for AI infrastructure is high, the scheme works flawlessly. But if bond yields continue to rise and investors start demanding a risk premium, the market could face a serious correction. In the crypto industry, we've already been through this — miners overloaded with debt are the first to be hit when market conditions shift.