The AI sector is experiencing a fundamental misalignment between capital deployment and earnings generation. Companies are raising and spending at rates that dramatically exceed the sector's ability to produce proportionate returns, recreating the financing dynamics that preceded the dot-com collapse.
The infrastructure requirements are real: AI development demands petabyte-scale data centers, custom silicon, and billions in upfront capex. But the financing mechanism now sustaining this build is circular. Nvidia, valued at five trillion dollars, recently secured five hundred billion dollars from Apollo, BlackRock and Goldman Sachs—ostensibly to fund customer orders for its chips. The structure is explicit: the financiers lend capital, the customer uses it to purchase Nvidia products, and Nvidia services the debt through chip sales. This closes a loop that can inflate revenue metrics while obscuring the underlying demand signal.
Max Gokhman, head of AI and digital asset solutions at Franklin Templeton, framed the risk plainly. "This will end badly," he said of circular financing structures. "We're living on not just borrowed time, but levered time." He emphasized that such intertwined financial arrangements amplify risk across the system.
The parallel to the late 1990s is structural, not superficial. Telecommunications equipment makers extended credit to carriers specifically to purchase their own gear. Revenue surged. Valuations soared. When the cycle broke—when customers could not service debt or demand contracted—the entire supply chain collapsed simultaneously.
Jensen Huang, Nvidia's CEO, has positioned AI compute as an "investable class," signaling industry conviction in sustained demand. But conviction is not the same as cash flow. Gokhman acknowledged AI's technological capabilities while maintaining that even significant technologies can fuel speculative bubbles. He drew a distinction: he does not yet view current leverage levels as immediately alarming, but he flagged the structural risk as material.
The boom has produced both outsized gains and sudden losses across investor portfolios, typical of volatile, speculative expansions where capital allocation is driven by narrative rather than earnings visibility. Until the sector demonstrates that chip deployment translates to measurable AI revenue—not just customer debt service—the circular financing mechanism remains a multiplier on downside risk.
