The AI Boom Has A New Problem: The Cost Of Capital

Oracle exemplifies rising capital intensity and credit risk in the AI infrastructure buildout, with capex surging and free cash flow turning sharply negative. AI infrastructure funding is shifting from internal cash to debt, raising the cost of capital and pressuring returns, as seen in discounted Project Jupiter loans and wider Oracle CDS spreads. Credit markets are signaling caution: Oracle’s credit rating is now BBB-, CDS spreads have tripled, and loans for major AI projects are trading below par. While AI demand is real, if infrastructure returns lag cost of capital, overbuilding and concentrated financing could trigger investor losses despite ongoing technological progress. The Real Risk In recent years, investors have discussed almost every major aspect of artificial intelligence: valuation, productivity, employment, security, and regulation. As an investor, I also ponder the question of whether it will be financially feasible to continue funding the infrastructure This article was written by Agar Capital 5.66K Followers Follow I’m a Portfolio manager (flexible equity funds and private clients), fundamental equity research, macro and geopolitical strategy.Over 10 years across global markets, managing multi-asset strategies and equity portfolios at a European asset manager.I combine top-down macro, bottom-up stock selection and real-time positioning (Bloomberg, models, data).I focus on earnings, tech disruption, policy shifts and capital flows — to identify mispriced opportunities before the market.On Seeking Alpha I share high-conviction ideas, contrarian views and deep breakdowns of both growth and value names.For more insights: follow me on X @AgarCapital