Imagine a chipmaker that hands a few billion dollars to one of its biggest customers, and that customer turns around and spends the money buying the chipmaker's chips. The chipmaker records a big new sale. But the cash that paid for it came from the chipmaker itself. The same dollar left as an investment and came home as revenue.

This is circular financing, and it sits at the heart of the artificial-intelligence boom. The companies selling the picks and shovels, the chips and the cloud computing, are also investing in the companies digging the mine. Analysts estimate the web of these deals now runs well past $800 billion.

The machine we are drawing is that loop. Done honestly, it is a normal way to finance a young, capital-hungry industry. But it also means the same money can be counted as revenue, as funding, and as future backlog at different stops along the chain, making demand look far larger than it may truly be.

The arrangement has grown enormous and tangled. The leading chipmaker's total web of AI deals is estimated to have crossed roughly $750 billion. It agreed to invest up to $100 billion in the maker of a famous chatbot, which in turn has committed well over a trillion dollars to buy computing power, much of it built on that same chipmaker's hardware.

Regulators have taken notice. In October 2026, the International Monetary Fund warned of a wrong-way risk in these deals: the investments lose value at exactly the moment the revenue they were meant to generate dries up. The loop that lifts everyone together can also pull everyone down together.

Why this reaches the reader: the companies at the center of this web, the big chipmaker and the cloud giants, are among the largest holdings in his index funds and retirement accounts. If the loop is inflating how much real demand exists, the headline growth propping up those stocks may be partly an echo of money chasing its own tail.

According to Wall Street legend Dr. Martin D. Weiss, there’s a $3 trillion financial crisis happening right now.

And almost nobody but him is warning people.

In fact, according to WSJ, this $3 trillion market “is under growing strain, despite the industry's upbeat tone.”

Weiss says the biggest danger is NOT just another stock market crash.

The biggest danger is NOT only a major bank failure.

Rather, it’s a chain reaction of crashes and failures across the globe.

>Go here now to see his shocking warning.

Start with the real problem it solves. Building artificial intelligence requires staggering amounts of expensive chips and data centers, far more than the young AI companies can pay for out of their own revenue. They need capital now and promise to pay it back with profits later. Someone has to bridge that gap.

The chipmakers and cloud providers step in to bridge it, because they are desperate to lock in buyers for their scarce, pricey hardware. So a chipmaker invests in an AI lab, or guarantees its debts, or agrees to buy capacity the lab cannot sell. In exchange, the lab commits to spend on that chipmaker's products. Supply and financing are bundled into one deal.

The load-bearing part is what happens to a single dollar inside this loop. The chipmaker's investment becomes the lab's funding. The lab's spending becomes a cloud provider's revenue and backlog. The cloud provider's purchases become the chipmaker's revenue again. The same underlying dollar is booked, in different forms, at three or four companies, each of which reports it as a sign of booming demand.

That is why outsiders cannot easily judge the truth. When invested cash returns as revenue, it is nearly impossible to tell, from the headline numbers, how much of that revenue is real demand from outside customers and how much is just money the companies handed to each other. The reported growth is genuine on paper and partly illusory in substance.

Here is how it can endanger giants. As long as AI keeps generating enough real, outside revenue to justify the spending, the loop is a sensible way to finance a build-out. But if that outside demand falls short, the whole chain is exposed at once: the chipmaker loses both its sales and the value of its investments, the lab cannot meet its commitments, and the cloud provider is left with debt and half-empty data centers. Everyone is tied to everyone.

The first joint is the inflated demand signal. Because the same money circulates and is counted at each stop, reported revenue and backlog across the AI chain can overstate how many genuine, paying, outside customers actually exist. Investors pricing these stocks on that growth may be paying for demand that is partly the companies financing one another, the exact trap that sank the telecom-equipment makers who lent their customers the money to buy their gear in 2000.

The second joint is the concentration of risk in a single point. The chipmaker now sits at the center as supplier, shareholder, and lender to its own customers. A stumble at one big AI lab would hit it three ways at once: fewer chip sales, a markdown on its investment, and a soured loan. The company that looks strongest because it touches every deal is also the one most exposed if any deal fails.

The third joint is the leverage hidden in the weakest links. The newer cloud players financing this build-out have piled on debt, in one prominent case tens of billions of it, with most of their revenue coming from a single customer. If that customer slows down, the debt remains. These firms are the canaries: their bond prices and single-customer dependence are where the strain in the whole loop shows up first.

What persists: the legitimate core, if the demand is real. Vendor financing is a normal, even rational, feature of capital-intensive industries, locking in capacity and sharing risk when the payoff is years away. The leading chipmaker still generates tens of billions of dollars in genuine free cash flow each quarter, real money, not accounting fiction. If artificial intelligence delivers the value its backers expect, the loop simply financed the future early.

What can break: the whole structure, together, if outside demand disappoints. The flaw is not that any single deal is fake; it is that the deals are interlocked, so a shortfall anywhere ricochets everywhere, and the circularity makes the shortfall hard to see coming. The machine looks most impressive, record revenue, soaring commitments, right when it is hardest to tell how much of the growth is real.

This machine connects to two others the reader's money touches. It drives the earnings and share prices of the handful of giant companies that now dominate his index funds, the same concentrated weights we have traced, and it ultimately rests on the mountain of debt raised to build the physical data centers that house the chips. The investment, the revenue, and the backlog are three faces of one circulating dollar.

The next reading arrives with each quarter's results from the chipmaker, the clouds, and the labs, and with the progress of real, outside AI revenue against the trillion-dollar commitments. Those will show whether the loop is financing a genuine new industry, or whether the same dollars are chasing each other in a circle that only looks like growth.