In the race to dominate artificial intelligence, leading tech firms are quietly piling on debt in ways that many investors never see.
Big names like Meta and Oracle are looking far beyond traditional borrowing channels.
They are tapping private lenders, creating off-screen financial entities, and structuring obligations so cleverly they barely register on corporate balance sheets.
AI Ambitions Deepen Risk
Dario Perkins, managing director of global macro at TS Lombard, sounds the alarm. He points out that the drive to pour billions into AI infrastructure is creating “an acknowledgement that this is getting out of hand.”
Meta is reportedly seeking a staggering thirty billion dollars in private capital for its expanding network of data centers. Other giants like Oracle have issued huge public debt — Oracle’s latest round totaled eighteen billion dollars — all to keep their AI ambitions fueled.
Critics worry that these loans are increasingly handled by special purpose vehicles that keep liabilities in the shadows. Perkins likened these modern tactics to Wall Street’s risky maneuvers before the subprime meltdown. “SPVs mean companies like Meta do not need to show the debt as their debt,” Perkins wrote, referencing the kinds of off-balance tricks that once reassured investors while hiding the real risks.
Paul Kedrosky, an investor and author who spoke recently on the Plain English podcast, sees it as a warning signal. “That for me is a reflection of not wanting the credit rating agencies to look at what they’re spending,” he said.
Despite these red flags, markets keep rewarding companies for their aggressive AI bets. The euphoria is reminiscent of other market bubbles, but Perkins is skeptical about how long it can last. “I wouldn’t touch this stuff now,” he remarked, adding that today’s environment looks much like the final days of the dotcom boom.
Insider selling and debt kept off the books are symptoms of an overheated market cycle, he says. Investors blinded by soaring gains might miss that the fundamentals do not add up. The market frenzy could continue for some time, they admit, yet cracks are forming beneath the surface.
For now, tech’s extraordinary profits may enable the spending spree. But if these companies are moving new debt out of sight, it suggests they know returns may not arrive soon enough to justify it all. The streak of record earnings and AI optimism may be more fragile than it seems, as even major developments in advanced AI systems like next-generation Claude model and related discussions in advanced AI offerings highlight both opportunity and risk in the sector.
