Circular Finance May Set Up Systemic Risk in Artificial Intelligence

Legis1

Source published: Site published:

Original article

Why It Matters

A new Congressional Research Service (CRS) report says that financial deals in which artificial intelligence companies invest in businesses that then buy their products or services could put the wider economy at risk of financial trouble.

This practice is called circular financing. It has become more common as technology companies spend billions of dollars on AI infrastructure. According to CRS, experts estimate that investment in AI-related infrastructure in the United States will reach about $500 billion this year and $10.3 trillion from 2025 through 2032.

These deals have led Congress to ask whether the AI industry's fast growth could cause economic or financial problems, and whether circular financing could make those problems worse. The report does not say that an AI investment bubble definitely exists or that circular financing would automatically cause a wider financial crisis.

The Big Picture

Circular financing happens when one company gives money or other financial support to another business, expecting that business to buy its products or services. These deals are becoming more common among AI developers, chipmakers, and cloud computing companies. The Bank for International Settlements found that 28.7% of investments made by AI companies from 2021 to 2025, based on the value of the deals, went to other AI companies. It also found that other AI businesses provided 55.2% of all investment money received by AI companies.

The report describes three kinds of circular financing. In investor-supplier deals, a company helps pay for a customer's purchase of its products. In investor-purchaser deals, a customer invests money in the company that supplies it. In two-way deals, companies make financial promises to each other. One major example involves Nvidia and OpenAI. On Sept. 22, 2025, Nvidia announced that it planned to invest up to $100 billion in OpenAI as the two companies worked to build computing infrastructure using Nvidia technology.

However, Reuters reported in January, citing The Wall Street Journal, that the investment plan had stalled. This raised questions about whether the planned deal would go ahead. Another example involves Amazon and Anthropic. In April, Amazon announced that it would invest another $5 billion in the AI company, with the possibility of adding another $20 billion. Anthropic also promised to spend $100 billion over 10 years on Amazon's chips and cloud computing services.

The report says these deals are a larger version of vendor financing. This is a long-standing practice in which suppliers help customers pay for the products they buy. These deals can give companies access to money they might have trouble getting elsewhere. Suppliers may also understand their customers' financial risks better than traditional lenders do.

This kind of funding is especially important for major AI developers. Many are startups that spend more money than they earn and depend on outside investors to grow their businesses. But these deals can make it harder for investors to tell whether demand for AI products and computing infrastructure comes from real market growth or from spending supported by the companies selling those products.

The report warns that complicated financial deals can hide how much AI companies depend on one another. This could make their exposure to financial problems elsewhere in the industry look smaller than it really is. The report compares this situation with earlier technology investment booms. During the dot-com era, telecommunications equipment suppliers and their customers also used circular financing deals. When investment money dried up, companies that depended on continued funding struggled to stay in business, and some failed.

The Bottom Line

The growing financial connections between AI companies could allow problems at one company to spread to its suppliers, investors, and customers. If major banks or other financial institutions have large investments in these deals, trouble in the AI industry could also affect wider financial markets.

The International Monetary Fund's October Global Financial Stability Report said that a possible drop in the value of AI-related investments was a major concern for financial market participants. However, the level of risk partly depends on how companies pay for their investments. If businesses use money earned from their own operations instead of borrowing from banks or other financial institutions, the immediate risk to the financial system may be smaller.

For Congress, the main unanswered question is whether circular financing mainly helps companies build the infrastructure needed to develop AI or creates financial connections that could make an industry downturn more damaging.