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

HSBC CEO Georges Elhedery cautioned on Tuesday that corporate spending on artificial intelligence risks outpacing near-term revenue gains, creating a potential disconnect between heavy capital outlays and companies’ ability to monetise them quickly. Speaking at the Global Financial Leaders’ Investment Summit in Hong Kong, Elhedery said the magnitude of required computing resources is clear, but consumers and businesses are unlikely to pay enough soon enough to justify the scale of investment many firms are committing to today.

Elhedery noted that the productivity benefits from AI are unlikely to materialise within a single year or two, observing that the transformation is better understood as a multi-year trend. He argued companies should expect a longer lead time before they begin to see convincing revenue upside that matches the pace of current capital expenditure.

Several industry studies cited at the forum underscore the scale of the buildout. Morgan Stanley estimated in July that global data-centre capacity will expand sixfold over the next five years, with data-centre construction and hardware costs reaching $3 trillion by the end of 2028. An April McKinsey report projected that by 2030 AI-capable data centres alone will require $5.2 trillion in capital investment, compared with $1.5 trillion for traditional IT-focused capacity.

General Atlantic chairman and CEO William Ford echoed Elhedery’s caution, agreeing that while AI will create substantial new industries and productivity gains over time, those payoffs are likely to unfold over a decade or more. Ford said the sector’s early phase will be capital-intensive and warned of the risk of poorly allocated investments, mispriced assets and speculative excess as market participants rush to back AI opportunities.

The current wave of spending by major technology companies is already sizeable. Alphabet, Meta, Microsoft and Amazon have all raised their capital-expenditure forecasts, and together these firms expect to invest more than $380 billion this year. OpenAI, whose release of ChatGPT in late 2022 triggered the recent surge in AI activity, has announced infrastructure commitments with partners that amount to roughly $1 trillion.

Ford described the scale of investment as a sign that the market recognises AI’s long-term significance, but he also cautioned that early years tend to produce both winners and many disappointments. He compared the technological shift to foundational innovations such as railroads or electricity: transformative over the long run, yet difficult to predict in terms of which companies will ultimately capture the most value in the short term.

Both Elhedery and Ford therefore urged patience and discipline. Their message to corporate leaders and investors was consistent: expect substantial upfront costs, recognise that revenue and productivity benefits may lag materially, and be mindful of the potential for overvaluation and capital misallocation during the initial surge of activity.