NEW YORK / SINGAPORE — October 14, 2026 — The global artificial intelligence revolution is being powered by a vast financial network in which some of the world’s largest technology companies are simultaneously investors, suppliers, customers and lenders to one another.
The arrangements are helping finance extraordinary spending on AI chips, data centers, cloud computing and electricity infrastructure.
But a growing number of economists, investors and financial regulators are warning that the interconnected system could amplify losses if the industry’s future profits fail to justify today’s investment commitments.
Bloomberg’s October 12 analysis, Inside the Giant Global Circular Financing Loop Driving AI, examines the money flows connecting companies such as Nvidia, Microsoft, OpenAI, Amazon, Google, Oracle and emerging cloud infrastructure providers.
The basic structure is deceptively simple.
A large technology company invests in an AI developer. The AI developer uses financing to purchase computing services or specialized chips. Those purchases then generate business for the technology company or its commercial partners.
As investment and revenue commitments grow, the companies become more financially dependent on one another.
The pattern is now substantial enough to attract attention from the Bank for International Settlements and the International Monetary Fund.
A BIS study published on October 1 found that nearly half the value of AI-to-AI investment deals over the 2021–2025 period involved companies that also had commercial supply-chain relationships.
On October 13, the IMF warned that highly concentrated exposure to AI assets, rising debt and potentially excessive growth expectations could increase the danger of a broad financial-market correction.
The AI boom is attracting hundreds of billions of dollars because its potential is enormous. But the bigger question is whether the industry’s increasingly circular financing is supporting genuine customer demand — or creating a web of promises that could unravel if AI profits arrive more slowly than expected.
What Is Circular Financing in Artificial Intelligence?
Circular financing describes a relationship in which a company invests in another business that also buys products or services from the investor.
Consider a simplified example.
A semiconductor manufacturer provides funding to an AI company.
The AI company uses its new capital to purchase processors or computing infrastructure that uses those processors.
The semiconductor manufacturer benefits in two ways.
It gains an investment in the AI company and generates demand for its products.
If the arrangement works, the AI company expands, the supplier sells more equipment, and both companies may become more valuable.
This cycle can be commercially legitimate.
Similar strategic investments exist in other industries.
The concern arises when financing flows become so interconnected that reported revenue, valuations and future purchase commitments depend heavily on continued investment from the same small group of companies.
Circular financing is not automatically fraudulent accounting.
It should also not be confused with sham transactions designed to create fictitious revenue.
The critical questions are whether the products and services are genuinely needed, whether the contracts are sustainable and whether the participating companies can ultimately generate sufficient income from independent customers.
BIS Study Reveals How Interconnected AI Investments Have Become
The Bank for International Settlements examined investment relationships across the AI sector between 2021 and 2025.
Its October 1 bulletin identified several significant patterns.
About 28.7% of the total value of AI firms’ investments involved AI companies investing in other AI firms.
Meanwhile, 55.2% of incoming investment into AI companies originated from other AI companies.
Most strikingly, 46.4% of the value of AI-to-AI investment deals involved firms that also maintained commercial supplier-customer relationships.
By the number of transactions, those relationships represented 16.1%.
The difference between transaction count and transaction value suggests that the largest deals are especially likely to involve overlapping investment and commercial relationships.
These statistics do not prove that nearly half of every dollar spent on AI is fraudulent or artificial.
They demonstrate that strategic investment and commercial purchasing are closely connected in an unusually capital-intensive industry.
The BIS concluded that these structures create legitimate economic benefits but also generate opacity and potential financial vulnerabilities.
Microsoft and OpenAI Helped Establish the Model
One of the most influential examples involves Microsoft and OpenAI.
Microsoft has invested more than $13 billion in OpenAI over several years.
The partnership gave OpenAI access to substantial computing capacity required to train and operate increasingly advanced AI systems.
At the same time, OpenAI became a major customer of Microsoft’s Azure cloud infrastructure.
Microsoft also integrated OpenAI technology into some of its products and services.
The relationship created advantages for both companies.
OpenAI obtained capital and computing resources.
Microsoft gained exposure to AI innovation, commercial demand for Azure and opportunities to expand AI-powered software.
However, the arrangement also illustrates the basic structure of circular financing.
Investment capital and commercial payments flow between closely connected companies.
The critical issue is whether the eventual demand for AI products is large enough to sustain the economic commitments made by both sides.
Nvidia Becomes Both Chip Supplier and AI Investor
Nvidia occupies an unusually important position in the AI economy.
Its graphics processors are widely used to train and operate advanced AI models.
The company benefits from demand generated by cloud providers, AI laboratories and infrastructure developers.
But Nvidia is also investing in businesses that expand demand for AI computing capacity.
That can help developers secure financing for new facilities and services.
It can also strengthen Nvidia’s relationships with important customers.
In January 2026, Reuters reported that Nvidia invested an additional $2 billion in CoreWeave.
The investment increased Nvidia’s exposure to a company building large-scale AI infrastructure.
CoreWeave has become an important cloud provider for businesses requiring specialized computing power.
However, the January investment was designated for infrastructure expansion, research and related development, not simply a direct requirement to purchase more Nvidia chips.
Such distinctions are important.
A commercial relationship may create incentives to expand purchases without every invested dollar being contractually earmarked for the investor’s products.
CoreWeave Shows How the AI Money Loop Works
CoreWeave has become a prominent example of the new AI infrastructure financing model.
The company builds and operates computing capacity for AI customers.
Its business requires large investments in processors, data centers and related technology.
Nvidia has invested in CoreWeave.
OpenAI has also entered into major computing agreements with the company and obtained an equity interest.
These relationships connect AI development, chip manufacturing and cloud infrastructure.
The economic logic is straightforward.
AI developers need computing power.
Cloud infrastructure companies provide that power.
Chip manufacturers supply the essential hardware.
Investors finance expansion in anticipation of future demand.
But the relationships can also create overlapping financial exposure.
If a major AI customer reduces spending, the cloud provider may suffer.
That could affect the cloud provider’s ability to purchase equipment or service debt.
The semiconductor company may then lose both sales and investment value.
The risks can spread through several companies even when the original shortfall originates with one customer.
Amazon, Google and Anthropic Form Another Financing Network
Anthropic has attracted significant investment from large technology companies.
Amazon and Alphabet’s Google have invested billions of dollars in the AI developer.
Anthropic also uses computing and infrastructure services associated with those companies.
For technology groups, these investments provide exposure to advanced AI development while encouraging demand for their cloud platforms.
For Anthropic, the relationships provide financial and computing resources.
The arrangement can help secure capacity that might otherwise be difficult or expensive to obtain.
However, the same financial interdependence raises questions about concentration.
If AI developers are reliant on a small number of infrastructure suppliers and investors, commercial problems may affect several businesses simultaneously.
The risk is especially important when long-term contracts assume sustained increases in AI adoption.
Oracle and OpenAI Raise the Stakes With Enormous Compute Commitments
Oracle has also become deeply involved in the AI infrastructure expansion.
The company is building data-center capacity and offering computing services to major AI customers.
OpenAI has been linked to substantial long-term purchasing commitments involving Oracle infrastructure.
Earlier reporting placed one major arrangement at approximately $300 billion over five years.
Such figures represent contractual commitments or anticipated future spending, not revenue already collected.
The commercial value depends on delivery schedules, computing demand and the terms of the agreements.
Oracle, in turn, must finance equipment, data-center construction and electricity requirements.
Some of that infrastructure relies on advanced Nvidia hardware.
The result is another chain in which future customer spending helps justify present-day borrowing and investment.
If demand develops as expected, the arrangement could support significant growth.
If it does not, infrastructure providers may be left with expensive assets and financial obligations.
SoftBank Raises $11 Billion as Its AI Exposure Grows
Japanese investment group SoftBank has made some of the industry’s largest financial commitments.
Reuters reported on October 13 that the company issued approximately $11 billion in high-yield bonds.
The borrowing increases the financial stakes behind CEO Masayoshi Son’s commitment to artificial intelligence.
SoftBank has invested heavily in OpenAI and other businesses associated with AI infrastructure.
Its strategy reflects confidence that the technology will create extraordinary long-term value.
However, concentrating a portfolio in a rapidly developing sector can increase risk.
Higher-yield debt also involves greater financing costs.
If investments disappoint, SoftBank could face pressure on valuations, liquidity and its borrowing profile.
The company has experienced major investment successes and failures in the past.
Its AI strategy may produce substantial gains, but there is no guarantee that every investment will deliver the projected return.
AI Infrastructure Debt Is Approaching Half a Trillion Dollars
The rise of circular investment has been accompanied by a sharp increase in debt financing.
According to Financial Times reporting on October 11, companies involved in AI-related activity had raised approximately $466 billion in debt during 2026.
That compares with roughly $101 billion in 2025.
The reported increase reflects demand for financing to build data centers, purchase chips and support infrastructure expansion.
However, the figure includes a wider ecosystem of companies, not just a handful of technology giants.
It also measures debt financing, not total AI investment or completed capital expenditure.
Some financing supports assets that may not become operational for several years.
Investors must therefore consider whether future revenues will be sufficient to cover interest payments and eventual repayment obligations.
IMF Warns of $950 Billion in AI Financing Linked to Future Revenue Expectations
The International Monetary Fund has added another layer of concern.
Its October 2026 financial stability analysis identifies risks from large AI-related debt and equity commitments and the possibility of disappointing profitability.
Contemporary reporting cited approximately $950 billion in AI-linked equity and debt commitments associated with around $1.4 trillion in anticipated future revenue.
These measures describe different categories of financial exposure.
They should not be added together as though they represent an amount already spent.
Nor should future revenue commitments be treated as profits or guaranteed cash receipts.
The IMF warned that a reassessment of expected AI returns could lead to falling asset prices.
Because investment and lending relationships overlap, a correction in one part of the industry might affect other companies and financial institutions.
The concern is not simply that AI stocks could decline.
It is that leverage and concentration could amplify those losses.
Why Circular Financing Can Make AI Growth Look Stronger Than It Is
A central criticism involves how investors interpret revenue growth.
When an AI company obtains funding and uses it to buy computing capacity from one of its investors, that spending generates genuine revenue for the supplier.
The sale is not inherently improper.
However, investors must understand where the purchasing power originated.
If demand depends on continued fundraising rather than revenue earned from independent customers, the arrangement may be vulnerable to changes in financing conditions.
The critical distinction is between investment-supported expansion and self-sustaining demand.
Early-stage industries often rely on external capital while customers and business models develop.
That can be economically rational.
But if projections become unrealistic, companies may invest in more infrastructure than the market can ultimately support.
Circular relationships can make these dependencies harder for outsiders to assess.
The Risk of Double Exposure
Circular relationships can concentrate financial losses.
Imagine that a chip company invests in an AI infrastructure provider.
The infrastructure provider buys chips from the manufacturer.
The arrangement helps both businesses expand.
But suppose the infrastructure provider’s customers later reduce spending.
The provider may struggle to repay debt and cancel future chip orders.
The chip manufacturer then suffers from weaker sales.
It may also record a loss on its investment in the infrastructure provider.
The same downturn has therefore affected the company through two channels.
Financial institutions that financed the infrastructure could experience additional losses.
This is the main reason interconnected relationships concern regulators.
The problem is not necessarily one company’s failure.
It is the possibility that many related companies face financial pressure at the same time.
Expensive AI Chips Could Become Obsolete Before Loans Are Repaid
Another major concern involves technology depreciation.
Advanced AI hardware is expensive.
Companies may borrow money to purchase processors and build computing facilities.
But semiconductor technology changes rapidly.
New generations of chips can offer better performance or lower operating costs.
Older equipment may become less economically competitive.
The Bank for International Settlements has highlighted the risks associated with long-lived financing arrangements backed by assets whose value may change quickly.
If a data center borrows against equipment expected to generate revenue for years, its economics depend on sustained utilization.
A rapid change in technology or customer demand could reduce that equipment’s earning power.
Investors would then need to reassess the value of the assets supporting their loans.
AI Energy Demand Creates Another Financial Constraint
Data centers need substantial electricity.
Their requirements include computing processors, cooling equipment, networking and supporting infrastructure.
As companies expand AI capacity, access to power becomes increasingly important.
Developers may need to arrange new generation capacity, transmission connections or other infrastructure.
These projects can be expensive and take years to complete.
That creates another mismatch between immediate financial commitments and future commercial returns.
A data center may secure billions of dollars in financing before all required power infrastructure is available.
Delays can increase financing costs and postpone revenue.
Higher energy prices can also reduce profitability.
The AI investment cycle is therefore linked not only to technology markets but also to energy, construction and infrastructure finance.
Investors Are Already Becoming More Selective
Signs of investor caution have begun appearing.
The Financial Times reported that AI-related debt issuance declined to approximately $23 billion in September.
That was less than half the level recorded in August.
The slowdown followed a much stronger period of borrowing earlier in 2026.
Investors have become more concerned about high borrowing costs, uncertain project returns and delays in infrastructure development.
However, reduced debt issuance does not mean the AI investment boom has ended.
Large projects and financing arrangements remain under development.
Instead, the slowdown suggests that lenders are demanding more convincing financial structures and clearer paths to repayment.
That could ultimately make the industry more resilient.
But it may also slow projects that were based on unusually inexpensive or readily available capital.
DBS CEO Says the Market Is Repricing AI Financing Risk
The debate has become particularly relevant to Asian financial institutions.
In an October 12 Bloomberg Television interview, DBS Group Chief Executive Tan Su Shan said investors were reassessing the financial risks of AI infrastructure.
She indicated that markets were demanding higher returns and examining whether financing structures could generate sufficient income to repay bondholders, lenders and shareholders.
Her comments reflect a broader change in investor attitudes.
The market is increasingly focused on financial returns rather than simply the scale of AI spending.
For banks, that means evaluating borrowers’ underlying business models, customer contracts and exposure to technological changes.
The important question is whether infrastructure will generate enough cash to justify the financing required to build it.
Why Asian Economies Are Exposed
The AI financing cycle has major implications for Asia.
Taiwan, South Korea and other economies supply important semiconductor components and advanced technology.
Cloud providers and data-center operators are investing across several Asian markets.
Financial institutions also participate in lending, underwriting and investment arrangements connected to AI infrastructure.
Strong AI demand can support manufacturing, exports and technology investment.
However, a sudden reduction in global spending could affect chip orders, construction projects and financial markets.
Companies with concentrated exposure to a small number of major AI customers may be particularly vulnerable.
The effects would depend on contractual arrangements and the extent to which demand comes from sustainable end-user activity.
The entire Asian technology sector should not be treated as having identical risk.
What This Means for the Philippines
The Philippines may benefit from AI adoption through software services, business-process outsourcing, digital infrastructure and workforce productivity.
However, the global financing model also offers lessons for domestic investors and businesses.
Large AI projects should be assessed according to realistic demand forecasts and operating costs.
Electricity availability, financing terms and customer concentration are especially important for data-center investments.
Businesses adopting AI should examine whether expected productivity improvements justify the costs.
For Philippine investors, overseas technology stocks and funds may offer exposure to the global AI boom.
But concentration risk matters.
A portfolio with holdings in several different technology companies may still be heavily exposed to the same underlying demand for AI computing.
The existence of several company names does not necessarily provide true diversification if their fortunes depend on one interconnected investment cycle.
Is This Another Dot-Com Bubble?
The circular financing debate has revived comparisons with the technology boom of the late 1990s.
During that period, companies invested heavily in telecommunications networks and internet infrastructure.
The underlying technologies transformed the economy.
But some businesses borrowed or raised capital based on overly optimistic demand forecasts.
When expectations were revised, companies failed and investors suffered large losses.
The comparison provides a warning, not a prediction.
AI may prove economically transformative while some projects remain unprofitable.
The relevant question is whether today’s valuations and financing commitments adequately reflect commercial and technological uncertainty.
There is no established finding that the entire AI industry is a bubble.
But the combination of high valuations, heavy borrowing and circular financial relationships warrants scrutiny.
What Financial Regulators Want to Understand
Regulators are examining several areas of potential risk.
They want greater clarity about which companies are financing one another.
They also need to understand whether assets are being used as collateral for multiple financing arrangements.
Private agreements can make these relationships difficult to observe.
Banks and investment funds may hold exposures through loans, bonds, equity investments and structured products.
Those exposures may become correlated during market stress.
The BIS has warned that complex and poorly disclosed financing arrangements can obscure risks.
The IMF has similarly emphasized the importance of monitoring leverage and concentrated investments.
Neither institution has declared that every circular AI deal is illegal.
The central concern is whether investors and regulators have enough information to assess what happens when economic assumptions change.
Five Developments to Watch
AI company earnings: Revenue generated from customers outside the financing network will be a critical measure of sustainable demand.
Data-center utilization: Investors will examine whether new facilities are being used enough to generate appropriate returns.
Corporate debt issuance: Financing costs and investor demand will influence future infrastructure expansion.
Semiconductor orders: Any sharp change in purchasing plans could affect both suppliers and infrastructure providers.
Regulatory disclosure: Greater transparency may reveal risks not fully visible in current private contracts.
Technology depreciation: Faster advances in hardware efficiency could affect the financial value of existing equipment.
These indicators will help determine whether the industry is moving toward sustainable profitability or continuing to rely heavily on fresh capital.
The Bigger Question: Is the AI Boom Building Real Wealth or Recycling Expectations?
The artificial intelligence boom has created extraordinary opportunities.
New models and software products are changing how people write, analyze information, develop software and conduct business.
The infrastructure needed to deliver those services requires massive investment.
Strategic partnerships between investors, chipmakers, cloud companies and AI developers can help finance that expansion.
But the same relationships can make the industry more fragile.
A company may be both an investor in and supplier to a major customer.
A lender may finance infrastructure whose value depends on contracts with a handful of AI laboratories.
A new round of investment may help fund purchases that support another company’s revenue.
These structures are not inherently improper.
Their success depends on whether real customers ultimately generate sufficient cash to sustain the ecosystem.
That is why the BIS and IMF are urging caution.
The most important test is no longer simply how much capital the AI industry can attract.
It is how much lasting value that capital can produce.
The world’s biggest AI companies are financing an extraordinary technological transformation through increasingly interconnected deals. But if future revenues fail to match their ambitions, the same financial loops accelerating growth today could transmit losses tomorrow — turning an AI investment slowdown into a much broader reckoning for technology companies, banks and global markets.
