- Equity has to carry more of the load. Blackstone President Jon Gray said Tuesday that financing the artificial intelligence buildout will require more equity because lenders are becoming increasingly exposed to the companies driving it. Speaking at the Hong Kong Monetary Authority’s Global Financial Leaders’ Investment Summit, he said the scale of capital required is creating concentration risk for debt investors across the sector and at individual companies.
- The numbers are enormous. Pimco CEO Emmanuel Roman, speaking on the same panel, cited General Atlantic CEO Bill Ford’s estimate that AI-related capital spending could reach around $10 trillion over the next five years. Roman said about $550 billion of AI debt has been issued in various forms so far this year, against a global corporate bond market of roughly $20 trillion.
- Lenders have limited upside. Gray noted that, unlike shareholders, debt investors earn only interest and principal, so they do not share in the growth that justifies the risk. Goldman Sachs estimated in a report Monday that global data center demand will require about $8.5 trillion in capital spending over five years.
- Demand is not the problem. Gray said demand for computing power is “running well ahead of supply” and that AI model spending across Blackstone’s portfolio companies has risen 20-fold over the past year. The bottlenecks are sites, permits, chips, power and capital.
- Wall Street’s bank chiefs say the money is there. At the same summit, Morgan Stanley CEO Ted Pick estimated the largest hyperscalers could deploy $800 billion to $1 trillion a year over a seven- to 12-year horizon, with cumulative AI spending potentially reaching $15 trillion over the next decade, according to Bloomberg. Goldman Sachs CEO David Solomon said the main constraint on deal size is the cost of capital rather than its availability: “There will be robust financing to meet the demand.”
What Happened?
Two of the world’s biggest investors used a Hong Kong summit on Tuesday to warn that the debt-heavy model for funding AI infrastructure is reaching its limits. Gray said financing will have to come from a combination of public markets, banks and private capital, with companies turning to structured transactions to limit the debt they take on, and with hyperscaler cash flows funding part of the bill. “The creativity is needed because the scale of capital is enormous,” he said. Roman was blunter: “At some point in time, it’s going to be too much.”
Why It Matters?
The comments land as credit markets push back on AI borrowers. Oracle filed Monday to sell $10 billion of stock alongside $10 billion of bonds, the kind of equity-heavy financing Gray describes, according to CNBC. When the largest private credit and bond managers say concentration risk is building, that signals tighter terms and higher costs for data center developers and neoclouds that depend on borrowed money. A shift toward equity also means more dilution for shareholders of companies building AI capacity.
What’s Next?
Watch for more AI companies following Oracle by pairing debt with stock sales, and for growth in structured deals such as joint ventures and off-balance-sheet vehicles. Pricing on new AI-linked bonds and private loans will show whether lenders are demanding more compensation for concentration risk. If capex estimates keep rising toward the $8.5 trillion to $10 trillion range, the gap between what companies want to spend and what markets will fund could become the main constraint on the AI buildout.
