A recent warning from Morgan Stanley highlights that the AI industry's capital intensity is reaching unprecedented levels as investment cycles for artificial intelligence computing accelerate. However, capital remains constrained across the broader ecosystem beyond the hyperscale cloud providers. To support massive infrastructure expansion, the US AI sector is evolving a highly complex, interwoven system of "new financing structures" and internal capital recycling.
According to a report from Morgan Stanley, AI-related investments are projected to account for approximately 50% of total capital expenditure for large-cap stocks in the coming years, surpassing the capital intensity seen during the previous dot-com bubble. The sheer scale and highly front-loaded nature of these investments have created a significant mismatch between recent capital demands and the realization of AI revenues. It is precisely this mismatch that has spurred the emergence of various novel financing structures throughout the ecosystem. The report uses OpenAI as a central case study, mapping capital flows involving multiple parties such as NVIDIA, Microsoft, Oracle, CoreWeave, Amazon.com, Advanced Micro Devices, and Walt Disney.
Morgan Stanley points out that the current AI boom is built on a foundation of mutual funding, warrants, and off-balance-sheet guarantees between suppliers and customers. While this model greatly accelerates infrastructure scaling, it may also inflate superficial contract pricing and obscure true economic leverage and risk outside of balance sheets, potentially leading to an overestimation of the "massive contracts" reported in headlines.
The core finding of the report is that the AI ecosystem has formed a highly interconnected structure where suppliers fund customers, and customers, in turn, support suppliers. These arrangements include supplier financing with preferential terms, long-term purchase commitments (Take-or-Pay contracts), revenue-sharing agreements, supplier buyback agreements, high customer concentration, third-party guarantees and endorsements, IP licensing in exchange for model access, and equity investments swapped for compute commitments.
Morgan Stanley believes these arrangements are essentially financing mechanisms that enable numerous ecosystem participants to expand their infrastructure scale to levels unsustainable by their own cash flows. The report explicitly warns that while these financing agreements accelerate data center construction, they may also pull forward future demand and reallocate risk among counterparties.
Using the relatively well-documented OpenAI ecosystem as an example, Morgan Stanley illustrates the staggering scale of this internal capital circulation. The analysis reveals deep entanglements with key players: Compute leader NVIDIA is deeply involved through a planned $300 billion investment in OpenAI, holdings in cloud provider CoreWeave, and complex buyback and leasing arrangements for computing capacity. Microsoft's extensive penetration includes a $130 billion investment in OpenAI, a commitment from OpenAI to purchase $2.5 trillion in Azure services, revenue-sharing agreements, and massive hardware procurement plans. Oracle and Advanced Micro Devices are linked via enormous orders, with OpenAI agreeing to purchase $300 trillion in compute from Oracle over roughly five years. Amazon.com and Walt Disney represent cross-sector involvement, with Amazon committing $500 billion to OpenAI and Disney investing $1 billion for model access, a deal that effectively uses IP as a form of non-cash financing.
Morgan Stanley indicates that within this AI capital network woven from trillions of dollars, new capital often covers only a portion of the total compute commitments. The fulfillment of the remaining contracts heavily relies on future revenue growth or additional rounds of financing. Investors pricing the AI fervor must remain vigilant about the systemic fragility that this circular capital flow might introduce.
Concurrently, the report clearly outlines seven potential risks inherent to this circular structure. Warrant grants can distort true pricing, as customers exchange long-term purchase commitments for supplier warrants, meaning headline contract values may not reflect repeatable pricing levels. Off-balance-sheet guarantees, often provided by cloud suppliers to support data center builds but not reported on their balance sheets, can hide true economic leverage, creating a significant divergence from reported leverage. IP licensing agreements, where content creators grant rights under favorable or non-cash terms for AI model access, can mask the true operational costs and cash needs of AI labs. Supplier equity investments amplify debt risk by providing cash flow backing for other suppliers selling to the same client, enabling them to take on more debt and creating a chain reaction that further fuels capacity expansion. High customer concentration significantly increases counterparty risk, as revenue growth becomes increasingly dependent on the success of a few major AI labs. Revenue-sharing arrangements among multiple parties can obscure true demand, as they may allow several entities to recognize the same revenue stream under US GAAP, making it difficult to gauge actual AI demand. Finally, supplier buyback agreements, which transfer risk back to the supplier and provide clients with downside protection, might artificially inflate apparent demand figures.
Comments