In the past year and a half, some of the biggest checks in tech history have flowed into artificial intelligence companies — and then quickly flowed right back out again. OpenAI, Nvidia, Microsoft, Oracle, CoreWeave, AMD and SoftBank have all inked multibillion-dollar deals that look less like traditional investment rounds and more like a high-stakes game of hot potato being passed between the same handful of players. One company injects capital into another, who then turns around and funnels that capital into the original investor’s chips, cloud contracts or data center buildouts.
Circular Financing Definition
Circular financing is a funding arrangement in which companies supplying a product or service are also major investors. This creates a loop where the invested capital flows back to the supplier through purchases, equity swaps or long-term contracts.
Known as circular financing, these “round-trip” transactions are shaping the entire AI industry — and, by extension, the global economy. They accelerate fast-moving innovation. They guarantee revenue. But they also blur the line between genuine market demand and artificially engineered growth that’s fed by the same companies supplying the core infrastructure.
Some analysts say it’s simply how modern hyperscale computing gets funded. Others see early warning signs of a bubble about to burst. This piece explains what circular financing is, how it works, what’s driving it in AI and why many people are suddenly so worried.
What Is Circular Financing?
The term “circular financing” refers to a deal structure where money flows between companies in a closed loop. It involves at least two parties, but can extend to multiple participants within an interconnected network that exchanges different forms of capital, including equity swaps, service contracts or infrastructure commitments. Typically, one company provides funding to another, who then uses that money to purchase the investor’s products or services, creating a guaranteed revenue stream. These cycles can repeat again and again, to reinforce growth and strategic alignment.
“It's a way for one company to ensure their own success and growth by financing it on their own terms,” Omar Rajjoub, vice president of private client group at HudsonPoint Capital, told Built In. “Companies do it because it fits their narrative and keeps them at the top of the market, where cash and speed are paramount.”
Circular financing is not illegal, and it’s not new. Wall Street has seen versions of this for decades, including vendor financing during the telecom boom in the late 1990s, when equipment makers lent money to buyers so those buyers could afford to buy more equipment. What is new, however, is the unprecedented scale of these deals in the AI sector. These aren’t just modest credit lines anymore — they’ve grown into $20 billion cloud deals, $100 billion chip purchases and $300 billion data center investments, often involving companies whose valuations exclusively hinge on ongoing AI investment.
How Does Circular Financing Work?
In general, circular financing tends to follow a fairly structured loop: Company A invests in Company B, providing capital that Company B can use for growth or operational needs. Company B then spends that capital on products or services from Company A, creating guaranteed demand for Company A’s business. This loop can expand to include an interweaving web of additional players as well. For example, Company B might use some of that funding to buy infrastructure or cloud services from Company C, which in turn could hold equity in Company B or Company A, creating another cycle of investment and guaranteed demand.
In the end, these multi-layered loops allow all participants to accelerate growth while locking in long-term financial and strategic benefits as long as they rotate within the roster.
These days, perhaps the boldest real-world ringleader of this method is OpenAI — the maker of ChatGPT— and its seemingly endless menagerie of deals. The structure works like this: Microsoft invests $13 billion in OpenAI (Company A to Company B). OpenAI then spends most of that money on Azure, creating guaranteed demand for Microsoft’s cloud business. From there, more players are looped in to build out that ecosystem. Oracle is constructing $300 billion in OpenAI’s Stargate Project data centers, which OpenAI will pay for over long-term contracts. This adds another cycle of capital and services. Equity can also be leveraged within the interconnected circuit. OpenAI received $350 million in CoreWeave stock, which it can later sell to pay CoreWeave for compute, while Nvidia and AMD participate through GPU purchases and stock options.
Circular financing creates a self-reinforcing loop of investment, infrastructure and product usage — and in the AI industry, that loop delivers some pretty powerful advantages:
- Faster Innovation: Companies get immediate access to compute, compressing months (or even years) of fundraising and procurement into weeks so they can train models and iterate in real time.
- Aligned Incentives: For better or worse, everyone is tied together. Companies scale faster while suppliers lock in predictable revenue, creating an ecosystem where everyone grows together.
- Physical Infrastructure Growth: The billions of dollars flowing through these deals translate into tangible facilities, accelerating data center construction and ensuring the physical backbone of AI keeps pace with development.
At the same time though, the circular financing model also introduces some pretty glaring vulnerabilities:
- Distorted Demand Signals: Because revenue is often generated from investor-funded spending, demand can appear stronger than it truly is, muddying the distinction between true market traction and money simply being recycled back and forth.
- Overdependance: If a chipmaker slows production, for example, or construction on a data center is delayed, it can affect the entire industry, delaying projects and straining cashflow for everyone else involved — especially dependent startups, who often take on debt they might not be financially ready for just to participate.
- Overbuilding: Circular financing encourages everyone to go full steam ahead, pushing companies to expand capacity based on optimistic forecasts rather than confirmed demand — a recipe for wasted resources if growth doesn’t materialize as predicted.
- Potential Market Collapse: At a system-wide level, any mismatch between expectations and reality can trigger broader instability, with one failure cascading through interconnected contracts, valuations and credit markets — the kind of chain reaction that fuels concerns about a trillion-dollar AI bubble.
Why Is Circular Financing Suddenly Everywhere in AI?
Circular financing has become the norm within the AI industry because there are few other ways for startups to survive and scale in this sector. The main driving force is extreme compute demand, which involves training cutting-edge AI models nonstop, running thousands of high-end GPUs for weeks at a time — amounting to a cost that most startups simply can’t cover with cash alone. That gives a small group of big-name suppliers — NVIDIA, Microsoft, Amazon and Google — outsized control, since they dominate both the chips and cloud capacity the market relies on.
Put simply: Circular deals allow these giants to lock in future demand, Rajjoub explained. “In essence, they are engaging in the opposite of mutually assured destruction, creating a mutually assured security.”
“It’s a growth industry and the companies that are relevant and active in this practice are highly profitable,” he added. The money they put in comes back as guaranteed revenue through compute-based contracts. “They have the dollars to do so and are going to try their best to make sure they are relevant with the next tech cycle.”
To be fair, the industry’s sheer compute requirements leave few alternative options. Training frontier models demands colossal, continuous GPU power — an unreachable expense far beyond what most startups can finance outright — and the ecosystem revolves around a tiny group of suppliers who control scarce quantities of chips and cloud capacity. That concentration of resources, combined with runaway demand, makes circular financing the fastest (and often only) viable way to secure the infrastructure required to compete.
Everyone is racing the same, accelerated timeline, too. Falling behind on model quality or data-center scale can be existential. And circular deals compress that timeline by pre-baking the spending into the investment itself rather than requiring companies to raise funds and then deploy them later. The structure also appeals to investors, giving them equity exposure to high-value AI leaders that are staying private longer.
In short, circular financing has offered a way for the AI industry to fuse operational needs with financial strategy, thus binding the interests of startups, suppliers and investors into one, big blob. For now, it seems to be the only strategy fast and flexible enough to keep pace with the dizzying speed of AI.
What Makes the AI Industry Deals Different?
There are several key differences between today’s circular financing deals in AI and those from previous tech eras. For example, the hyperscalers driving the largest infrastructure spending are using their own prodigious free cash flow rather than relying heavily on outside, vendor-funded capital, as telecom and dot-com companies did in the 1990s.
“Back in the day, the big telecom companies loaned money to its customers, and those customers used the money to buy the telecom equipment,” Jason Tassie, a growth strategy expert and founder of Know Your Business, told Built In. “But that meant that the demand wasn’t real, it was artificially created.”
This ability to self-finance allows these giants to build a sturdier, more resilient foundation for their AI operations. Meanwhile, many of the most promising AI startups are still being fueled by massive venture capital rounds, but they are often generating revenue through subscription models as they scale, a key differentiator from the “build first, monetize later" approach of the past.
So far, taking advantage of bullish investor interest is working in their favor. AI startups captured more than half of all venture capital dollars invested globally in 2025, marking the first time a single technology sector has claimed a majority share.
Even with all of this money swirling around, the most advanced AI chips — the actual hardware that keeps the industry going — are still incredibly hard to get. While general data center capacity is relatively stable, these high-tech GPUs are running almost nonstop, showing how focused the market is on securing essential compute power. And while rapid buildouts typically increase the risk of overbuilding, today’s companies are expanding their capacity in more strategic stages based on customer contracts, which keeps any equipment from sitting idle as it did in previous market bubbles.
“AI is on a much stronger footing today,” Tassie said, “but there is still the risk that companies start to build more AI capacity than the world is ready for.”
Is Circular Financing a Sign of a Bubble?
Skeptics say that circular financing in and of itself isn’t the problem — the real risk is what the structure tends to conceal underneath. When suppliers invest heavily in startups that turn around and spend that money on the suppliers’ own cloud or hardware, the resulting growth metrics can blur the line between organic demand and engineered momentum.
“When you start to notice that a company's customers are also its investors or its lenders, that's the most obvious sign [of a bubble],” Tassie said — and that exact arrangement is simply how circular financing works in practice. Another indicator is when revenue grows faster than real usage. “If no one is using a product yet, but a company announces huge future investment deals, then perhaps that demand is being created on paper,” he added.
This echoes warnings from multiple economists and market watchers. As described in a recent piece from CNBC, some experts now argue the current AI boom shows classic signs of a bubble: steep valuations, waves of speculative infrastructure spending and a disconnect between investor optimism and actual consumer or enterprise uptake. Adding further weight, a paper from Cowles Foundation at Yale University uses data from 2023 to 2025 to show that many AI‑related firms are trading at valuations well above what their real performance supports — a classic bubble pattern.
It’s worth noting, however, that bubbles are notoriously hard to recognize until they pop. Cutting-edge technologies like artificial intelligence frequently generate hype that may look irrational in the moment but later proves to be foundational, making it difficult to distinguish true excess from legitimate investment.
Why You Should Care About Circular Financing
Circular financing matters even if you’re not building AI models. It shapes everything from product pricing and availability — including ChatGPT subscriptions and the AI features in apps you use every day — to the valuations of the largest companies in the S&P 500. It dictates where data centers are built and how much capital flows into chips versus other sectors of the economy. And if circular financing is inflating a bubble, the fallout won’t be isolated to AI companies; it’s likely to affect nearly every company and the world’s economic stability at large.
Frequently Asked Questions
What is circular financing?
Circular financing is a looped funding arrangement where investors provide capital to a company, and that company then spends the money on that same investor’s products or services, effectively recycling the funds within a closed network. In AI, this often involves a startup using investment cash to buy GPUs, cloud services or data-center space from the very companies that funded them.
Why is circular financing becoming so common in the AI industry?
AI demands huge upfront spending on hardware, cloud services and data centers — costs that most startups can’t cover alone. Investors and tech giants step in with both the money and the resources, creating a cycle that speeds up growth but also piles on risk for everyone involved.
