HSBC, General Atlantic CEOs flag AI capex-revenue mismatch, 'irrational exuberance'

As the artificial intelligence sector attracts staggering investments, HSBC's CEO Georges Elhedery raised concerns about a growing disparity between capital expenditure and revenue generation. During the Global Financial Leaders' Investment Summit in Hong Kong, Elhedery highlighted a significant challenge for businesses: although investment in AI's computational infrastructure is critical, the current revenue streams may not adequately support such extensive spending. A report from Morgan Stanley in July projected that global data center capacity will expand sixfold over the next five years, with associated costs reaching $3 trillion by the end of 2028. Meanwhile, a McKinsey analysis released in April anticipated that by 2030, data centers built to manage AI workloads will require a staggering $5.2 trillion in capital expenditures, in contrast to the $1.5 trillion needed for traditional IT applications. Elhedery emphasized that consumer readiness to absorb these costs is lacking, suggesting that businesses must adopt a cautious approach, as the anticipated productivity gains from AI will not materialize in the short term. "These trends span five years, meaning significant revenue and willingness to invest may take longer than what investors expect," he explained. William Ford, CEO of General Atlantic, echoed Elhedery's sentiments during the same discussion, noting that while AI will foster new industries and applications, the tangible benefits will emerge over a longer timeframe, potentially spanning a decade or two. Big Tech companies—including Alphabet, Meta, Microsoft, and Amazon—have recently revised their capital expenditure forecasts, collectively estimating to spend over $380 billion this year. OpenAI, the pioneer of the recent AI boom following the launch of ChatGPT, has also announced infrastructure deals valued at approximately $1 trillion with partners like Nvidia, Oracle, and Broadcom. Ford remarked that the substantial investments in the AI sector reflect a recognition of its long-term potential. However, he cautioned that the industry will initially be capital-intensive, requiring significant upfront investment to unlock future opportunities. He also warned of potential pitfalls, such as capital misallocation and excessive valuations, particularly in the early stages of development. "Identifying winners and losers is challenging right now. It’s akin to betting on a technology that could transform the economy over time, much like the railroads or electricity, though its immediate impacts are unpredictable," he concluded.