TLDR:
- OpenAI holds over $1.4T in infrastructure commitments against just $13B in annual revenue today.
- Microsoft, Amazon, Oracle, and Nvidia fund OpenAI while receiving its spending commitments back
- Morgan Stanley warns AI capital spending may exceed 50% of all large-cap capex industry-wide.
- Off-balance-sheet deals hide true leverage, leaving investors blind to the full financial exposure.
OpenAI sits at the heart of what Morgan Stanley has described as the most dangerous money loop in modern financial history.
The artificial intelligence company has accumulated over $1.4 trillion in infrastructure spending commitments. These commitments are tied to circular deals involving Microsoft, Amazon, Oracle, Nvidia, and CoreWeave.
Against this, OpenAI’s annual revenue stands at roughly $13 billion. The gap between its spending obligations and current earnings has drawn serious concern from Wall Street analysts.
How the Circular Deals Work
Microsoft gave OpenAI $13 billion in funding. In return, OpenAI committed to spending $250 billion on Microsoft’s cloud services. The flow extends further to Oracle, which signed a $300 billion cloud deal with OpenAI.
Oracle uses that revenue to purchase Nvidia chips for building data centers. The money moves in a tight circle among the same group of companies.
NVIDIA added another layer to this structure. The chipmaker invested $100 billion into OpenAI, and OpenAI uses those funds to buy Nvidia’s chips.
In effect, Nvidia funded one of its own primary customers. CoreWeave, in which Nvidia holds a 7% stake, also committed $6.3 billion to use Nvidia’s own cloud infrastructure.
CoreWeave then invested $350 million into OpenAI and expanded its service contracts to $22.4 billion. Amazon is the newest addition to this cycle.
The tech giant committed $50 billion to OpenAI. OpenAI, in turn, agreed to spend $100 billion on Amazon Web Services over the next eight years.
As Milk Road AI observed, “Every player is funding the next player in the chain.” The capital does not originate from outside customers or independent revenue streams.
Instead, it circulates within a closed ecosystem of major technology firms. Morgan Stanley refers to this pattern as “capital inner circulation.”
The Risk Behind the Numbers
Morgan Stanley warned that OpenAI’s spending commitments far exceed its current revenue base. The firm has already cut its projected total in half.
However, new funds raised still cover only part of what OpenAI has already promised to spend. The remainder depends on future revenue that does not yet exist.
The bank also flagged that many of these deals rely on off-balance-sheet tools. These include guarantees, warrants, and revenue-share arrangements.
Investors cannot see the full exposure on any single company’s balance sheet. The true leverage remains largely hidden from public view.
AI capital spending is now on track to exceed 50% of all large-cap capital expenditures. That level surpasses the spending intensity seen during the dot-com bubble, according to Morgan Stanley. Microsoft, Amazon, Oracle, and Nvidia carry a combined market cap running into the tens of trillions.
If OpenAI cannot generate the revenue needed to honor its commitments, all these firms would take financial hits at the same time. That simultaneous exposure is the core of the systemic risk Morgan Stanley has identified.



