AI Funding Rush Nears $1 Trillion as Borrowing Swells and Circular Finance Spreads

Companies Turn to Bonds, IPOs and Share Sales to Raise Cash

About $3 Trillion of AI Investment Needed by 2028

Roughly $1 Trillion Must Be Raised Each Year

As the race to build artificial-intelligence infrastructure intensifies, Wall Street is mobilizing every major funding channel available — including corporate bonds, initial public offerings, share sales and private credit — to pour massive sums into technology companies.

Some on Wall Street say additional bond issuance and other fundraising could push the total to as much as $1 trillion this year. Morgan Stanley projects about $3 trillion of AI infrastructure capital spending by 2028, implying roughly $1 trillion of financing a year will be needed to support that investment.

Still, concerns about overheating are growing. Some companies, including Oracle, have sold debt so aggressively that their bonds are trading at levels typically associated with junk-rated issuers.

About $3 Trillion Needed for AI Infrastructure

In a recent report, Morgan Stanley estimated that global data-center construction costs will reach about $3 trillion from 2025 through 2028. More than 80% of that spending has yet to be deployed. On a simple annualized basis, that points to investment of more than $800 billion a year starting this year.

The bank also said intensifying competition in AI is fueling the mergers-and-acquisitions market, as companies pursue deals to secure AI technology, talent and customer bases. Once M&A-related financing is included, Wall Street expects companies to need about $1 trillion a year for AI-related investment.

Even so, fundraising that can be anticipated this year amounts to only about $500 billion. Major technology companies including Alphabet, Amazon, Meta and Oracle are set to issue a combined $159 billion of debt this year to secure large AI data centers and GPU infrastructure. OpenAI, Anthropic and SpaceX are projected to raise a combined $200 billion through IPOs and investment. Add private credit and bond issuance by other AI companies, and the total could reach $500 billion.

Funding Methods Are Becoming More Diverse

To meet that surge in demand, Wall Street and technology companies are deploying a full range of financial-engineering tools. Alphabet recently launched an unusually large $85 billion equity-issuance program to help fund AI infrastructure.

This year, Alphabet has sold bonds not only in US dollars but also in Canadian dollars, Japanese yen, euros, Swiss francs and British pounds. It also drew market attention with a 100-year sterling bond sale. In California's municipal-bond market, the company is also raising $1 billion for energy projects. Amazon has likewise tapped the Canadian-dollar bond market this year after issuing debt in the US, Europe and Switzerland.

Oracle, which raised $5 billion, and AI cloud company CoreWeave, which raised $4 billion, are also making active use of the convertible-bond market, where debt carries the right to convert into equity.

'Oracle Faces Late-Mover Risk'

Among them, Oracle is the company drawing the closest scrutiny on Wall Street. Since September last year, Oracle has issued $43 billion of corporate bonds. The company plans to spend tens of billions of dollars over the next several years as it shifts from a software company to an AI cloud-infrastructure provider. The investment is aimed at expanding its business of leasing large GPU clusters to customers including OpenAI.

That transition, however, is expected to drive a sharp increase in cash burn. Despite retaining an investment-grade credit rating, Oracle's bonds are effectively trading at speculative-grade levels in the bond market.

There is also growing concern over heavy capital spending by Big Tech. Bank of America estimates capital expenditures by hyperscale cloud companies including Microsoft, Amazon, Google and Meta will reach about 100% of operating cash flow by the end of this year, up from roughly 40% in 2023. BofA also said investment-grade bond issuance and equity supply have increased, while share buybacks have slowed and cash-generation capacity has stagnated.

Mutual Financial Dependence Is Rising

Interdependence across the AI industry is deepening. Google signed a long-term TPU supply agreement with Anthropic last year and made an additional $10 billion investment two months ago. More recently, it signed a large contract with SpaceX. SpaceX said Google agreed to pay $920 million a month from October 2026 through June 2029 to secure computing capacity.

Wall Street views that structure as a form of circular finance within the AI industry. Companies are repeatedly investing in one another, buying one another's products and supporting one another's fundraising, rapidly deepening mutual dependence across the sector.

Some worry that the arrangement could amplify systemic risk. If one company's investment plans falter or its funding runs into trouble, related companies could be hit in a chain reaction. The Federal Trade Commission launched an investigation last year into investment and partnership structures between Big Tech and AI companies. The agency is examining whether some deals create arrangements in which companies effectively prop one another up.

Park Shin-young, New York correspondent, Hankyung.com, nyusos@hankyung.com