Saturday, November 21, 2026

55% of AI Investment Funds Circulate Within the Ecosystem... Warning Signs of an AI 'Debt-Fueled Spending Spree'

Input
2026-11-21 06:00:00
Updated
2026-11-21 06:00:16
AI-generated image

[Financial News] Could the astronomical revenues of artificial intelligence (AI) companies really be generated solely by pure market demand? Things look different if a customer buys a company’s products with money that the company itself invested.
According to financial investment industry sources on the 21st, warnings are emerging that behind the explosive growth of the global AI industry lies a so-called “circular finance” structure in which companies prop up one another’s investments and revenues.
The concern is that investment flows back into product purchases, boosting reported revenues, but could pose financial risks across the industry if not backed by genuine end-user demand.
55.3% of investment circulates within the AI ecosystem
The link between investment and revenue is at the core of circular finance. From 2021 to 2025, 55.3% of all investment raised by AI companies came from other AI companies. Investments between companies connected through supply chains accounted for 46.4% of the total, based on disclosed amounts.
According to an analysis by the International Monetary Fund (IMF), $300 billion in equity investments led to $1.4 trillion in future sales commitments, which in turn helped raise at least $650 billion in follow-on debt. In effect, small equity investments led to long-term purchase agreements and borrowing, generating financial transactions far larger than the initial investment.
A prime example is the relationship between NVIDIA Corporation and neocloud companies. When NVIDIA invests in a cloud company, that company uses the funds it secures to buy NVIDIA GPUs.
For NVIDIA, investment and product sales happen simultaneously, but it is necessary to examine separately how much of the revenue generated in this process reflects independent demand from end customers.
$1.1 trillion in debt piling up off balance sheets
The bigger problem is that related risks are accumulating off balance sheets. Big Tech and AI companies are expanding infrastructure investment by using special-purpose vehicles (SPVs) to build data centers and provide long-term lease commitments or payment guarantees. By using structures in which obligations are not recognized as direct debt, they may make the actual financial risks they must bear difficult for outsiders to assess.
The $27.3 billion Hyperion project, in which Meta participated, secured credit support through long-term lease commitments but was not included in Meta’s consolidated financial statements. NVIDIA also has contingent obligations of up to $105 billion under the Ohio project, triggered if the customer is unable to pay rent.
According to the Federal Reserve System (the Fed), the outstanding balance of publicly issued debt related to advanced computing infrastructure is estimated to exceed $1.1 trillion. This means that while reported revenues and investment are growing rapidly, borrowing and payment guarantees behind them must also be examined.
The vendor-financing trap that resurfaced during the dot-com bubble
The market is recalling “vendor financing” from the collapse of the dot-com bubble in the early 2000s. At the time, telecommunications equipment makers such as Lucent and Nortel directly provided financing to boost their customers’ purchasing power. The customers used the funds to buy equipment, allowing suppliers to increase sales.
But when telecommunications demand declined, customer credit risks and equipment makers’ collection risks surfaced at the same time, leading to a cascade of financial distress. The AI industry could face similar risks if actual demand falls short of expectations in a structure where investment and purchases are intertwined. In particular, the rapid obsolescence of AI equipment heightens the risk of falling collateral values and financial institutions’ difficulty recovering their funds.
Kang Jin-hyuk, a researcher at Shinhan Investment & Securities, noted, “When suppliers provide financing to customers, some of the reported revenue may be generated endogenously.”
He continued, “The standard for assessing the sustainability of circular finance is not accounting profit or revenue growth, but profitability and cash flow secured from end customers,” and stressed, “This is the time to jointly assess the actual monetization of future AI service revenue, free cash flow (FCF), and whether investment and debt repayment can continue without external financing.”

[email protected] Kang Jung-mo Reporter