Blackstone’s Jon Gray Urges More Equity Financing as AI Infrastructure Debt Risks Grow

Table of Contents

• You might want to know


• Main Topic


• Key Insights Table


• Afterwards...

You might want to know

• Why is Blackstone’s president calling for more equity to finance artificial intelligence infrastructure?


• Can financial markets absorb the enormous investment expected for AI and data centers?

Main Topic

The rapid expansion of artificial intelligence is driving a major buildout of computing infrastructure, including data centers, chips, and the power and connectivity needed to operate them. That expansion requires substantial capital. At the Global Financial Leaders’ Investment Summit hosted by the Hong Kong Monetary Authority, Blackstone President Jon Gray argued that a greater share of the financing should come from equity as lenders take on more exposure to the companies and projects behind the buildout.

Gray’s concern is not simply that AI-related investment is large. It is also that borrowing may concentrate risk among debt investors. Exposure can build across the AI sector as a whole and within individual companies, particularly when many projects depend on continued growth in demand and successful access to essential infrastructure. If financing relies too heavily on debt, lenders could face substantial exposure without receiving a corresponding share of the potential gains.

The difference between debt and equity helps explain Gray’s argument. Lenders generally receive interest and repayment of principal, so their returns are limited even if a borrower becomes much more valuable. Shareholders, by contrast, may benefit when a company grows, although they also bear the risk of losses if its performance disappoints. Equity can therefore share more of the upside and downside of an uncertain, capital-intensive expansion, while reducing the pressure to fund every project with debt.

Gray did not suggest that debt should disappear from AI financing. Instead, he described a range of funding sources that could contribute to the buildout: public markets, banks, and private capital. Companies are also increasingly using structured transactions to limit the amount of debt they take on. In addition, cash generated by hyperscalers—large technology companies operating extensive computing platforms—could pay for some of their infrastructure spending.

Gray said that the scale of investment makes creativity in financing necessary. A combination of funding channels may help distribute exposure and match the financial structure of a project to its risks. For example, equity can provide a buffer against uncertainty, while debt may still support projects with more predictable cash flows. The mix can vary by company, asset, and stage of development.

At the same summit panel, Pimco CEO Emmanuel Roman warned that the total amount of financing required could be too large for markets to absorb comfortably. Roman cited General Atlantic CEO Bill Ford’s estimate that AI-related capital spending could reach around $10 trillion over the next five years. He also said about $550 billion in AI debt had been issued in various forms so far this year.

Roman emphasized that financing the anticipated spending would remain a major challenge even if businesses used both debt and equity. He described the required amount as “gigantic” when compared with a global corporate bond market of roughly $20 trillion. His warning was about the scale of the demand for capital relative to the capacity of financial markets, rather than a claim that any single financing route would necessarily fail.

Separately, Goldman Sachs wrote in a report Monday that meeting global data center demand would require about $8.5 trillion in capital spending over the next five years. These estimates point to the breadth of the infrastructure challenge. AI investment is not limited to developing models or buying computing chips; it also involves securing sites, constructing facilities, obtaining permits, and arranging reliable power and financing.

Roman’s warning and Gray’s comments highlight related but distinct questions. Roman focused on whether markets can supply the total amount of capital required. Gray focused on the composition of that capital and the risks created when lenders become highly exposed. Together, their remarks suggest that the funding challenge is both quantitative and structural: the industry needs vast sums, and those sums must be allocated through arrangements that remain manageable for companies and investors.

Despite these concerns, Gray remained positive about the outlook for AI infrastructure. He said demand for computing power was “running well ahead of supply.” That imbalance helps explain why companies are continuing to invest, even as the cost and complexity of development increase. Strong demand, however, does not eliminate the need to assess whether individual projects can be completed and financed on sound terms.

Evidence of business interest can also be seen in spending on AI models among Blackstone’s portfolio companies. Gray said that this spending had risen 20-fold over the past year. He connected the increase to financial benefits businesses were seeing from applications such as coding, customer engagement, and content creation. These uses can support investment by giving companies practical reasons to adopt AI tools, rather than relying only on expectations about future demand.

Yet converting interest into operational capacity remains difficult. Gray identified sites, permits, chips, power, and capital as increasingly challenging requirements. Each can constrain development: a project may have funding but lack a suitable location, or a site and equipment but insufficient power. Delays or shortages in one area can affect schedules and costs across the entire project.

The practical implication is that AI infrastructure finance cannot be considered separately from the physical and regulatory conditions that enable construction. Investors must evaluate not only the projected demand for computing but also the feasibility of obtaining the resources needed to meet it. For companies, careful sequencing and diversified financing may help reduce the chance that a bottleneck or a change in market conditions disrupts a large investment plan.

These comments do not establish that the AI investment boom will end, nor do they show that debt financing is inherently unsuitable. Rather, they identify risks that become more significant as capital requirements grow. Equity, debt, private investment, public markets, structured transactions, and internally generated cash can each contribute, but the appropriate balance depends on the financial resilience of the borrower and the characteristics of the project.

Key Insights Table

Afterwards...

The financing debate points to several areas that businesses, policymakers, and investors will need to examine as AI infrastructure expands. One is how to assess the financial resilience of projects whose costs arrive well before their expected returns. Better disclosure of project funding, debt exposure, and reliance on future demand could help market participants understand where risks are concentrated.

Another area is the connection between computing infrastructure and energy systems. Data centers require dependable power, and the availability and timing of that power can shape where projects are built and how quickly they become operational. Further work on grid capacity, energy efficiency, storage, and lower-impact cooling could help infrastructure providers meet computing needs while managing pressure on energy resources.

Research into more efficient computing may also influence how much capital the industry ultimately needs. Improvements in chips, software, model design, and data center operations could allow more computing work to be completed with fewer resources. Efficiency gains would not necessarily remove the need for new infrastructure, but they could affect the balance between demand, operating costs, and construction requirements.

Finally, financial innovation deserves careful evaluation. Structured transactions and a broader mix of funding can help distribute risk, but complexity should not obscure who bears that risk or how obligations may change under different conditions. The most durable path forward will depend on pairing technological progress with transparent financing, realistic project planning, and infrastructure that can be built and powered reliably.