Updated
Updated · CNBC · Oct 20
HSBC CEO Urges Selective AI Data Center Bets as Global Spending Targets $7 Trillion by 2030
Updated
Updated · CNBC · Oct 20

HSBC CEO Urges Selective AI Data Center Bets as Global Spending Targets $7 Trillion by 2030

2 articles · Updated · CNBC · Oct 20

Summary

  • Georges Elhedery said investors should not take an all-in approach to AI data centers, arguing some projects are harder to back despite the sector's rapid expansion.
  • Tenancy agreements and contract pricing are the key filters, he said, warning that not all deals have the quality needed to support strong investment.
  • HSBC, with nearly $3 trillion in assets, is exposed on both sides of the boom as a financier and operator of data centers while using AI to re-engineer the bank.
  • McKinsey sees data center spending reaching $7 trillion by 2030, but financing strains are rising as hyperscalers and governments compete for capital in bond markets.
  • JPMorgan estimated earlier this year that AI-related debt issuance will hit $4.1 trillion by decade-end, adding to concerns over higher yields and swelling corporate leverage.

Insights

Why is one of the world's largest banks urging extreme caution in the middle of a $7 trillion AI infrastructure gold rush?
Will severe power grid limits and cooling demands turn today's multi-billion-dollar AI data centers into tomorrow's stranded assets?
Could the trillions in hidden, off-balance-sheet data center debt trigger the next major global financial crisis?