World economy: AI development brings growing macroeconomic and financial risks

Event

The US S&P 500 plunged on Thursday 5 November, following the US midterm elections. The sell-off partly reflected concerns about the potential implications of Democratic policies regarding AI-related data centres. During the election campaign, such projects had become an important political issue, notably due to their rapidly growing requirements for electricity, water, land and supporting infrastructure.

Impact

Although AI investment has become an important source of global economic resilience, its large and growing importance represents a new source of macroeconomic and financial vulnerability.

First, the macroeconomic channel. AI development has generated a major investment cycle, notably in the US, boosting demand for semiconductors, data centres, cooling systems, fibre and electrical equipment, electricity generation, grid infrastructure and copper, among others. This boom has supported global economic activity and trade, contributing to the resilience of the world economy in a context of large geopolitical and economic uncertainty. Technology-exporting economies such as Taiwan, South Korea, Singapore and Malaysia have particularly benefited from this boom. However, this success has also created new vulnerabilities. A significant reduction in AI-related investment would weigh on investment, particularly in the US, and reduce demand across the technology, energy and commodity supply chains. Economies and sectors that have benefited most from the AI boom would consequently be particularly exposed.

Second, the financial channel. The enormous capital requirements for AI infrastructure are leading to greater reliance on debt and increasingly complex and opaque financing structures. The combination of elevated asset valuations, concentration among a relatively small number of companies and growing financial interconnectedness increases the risk of financial spillovers in case of shock. Weaker-than-expected earnings growth, lower expectations for future AI profitability or a reduction of investment could trigger a sharp repricing of AI-related assets. Such a correction could have spillover effects well beyond AI-related sectors, leading to a worsening of global financial conditions at a time when sovereign bond yields are already high and on the rise in several major economies, including the US, which undermines governments’ ability to absorb shocks.

Beyond macroeconomic and financial vulnerabilities, the rapid development of AI is generating wider cybersecurity, geopolitical, labour-market, environmental and societal risks, as well as potentially existential risks. These risks are amplified by increasing digitalisation and interconnectedness, as well as by intensifying US-China strategic competition for leadership in AI and other advanced technologies (such as space technologies), as they are becoming increasingly important determinants of economic and military power.

Analyst: Pascaline della Faille (P.dellaFaille@credendo.com)