Big Tech is hiding 300 billion dollars in AI investments through guarantees

Major technology companies are using financial guarantees to keep over $300 billion in AI investments off their balance sheets, allowing corporations to shift risks associated with capital-intensive AI projects to third-party organizations and partners. According to the Financial Times, the scheme involves using complex contracts that are not formally accounted for as direct company obligations, resulting in investors and regulators receiving a distorted view of the actual level of debt and financial risks that industry giants bear when scaling infrastructure for model training. The use of such mechanisms has become a response to the rapid growth of costs for GPU clusters and data center construction, with companies entering into agreements where financial obligations for renting or maintaining facilities fall on subsidiaries or specialized investment funds backed by guarantees from parent corporations. Experts note that this strategy allows for artificially lowering CAPEX figures in quarterly reports, creating an illusion of higher operational efficiency, but this practice raises concerns among analysts, as in the event of a market correction or decrease in demand for AI services, hidden obligations may require emergency asset write-offs, negatively impacting company capitalization. This situation highlights the growing dependence of Big Tech's financial stability on opaque infrastructure financing schemes, with discussions of the risks associated with these methods being actively conducted in the professional community, including discussions on Hacker News, where market participants point to the potential threat to the stability of the technology sector.