JPMorgan raised its forecast for global AI infrastructure spending to $5.5 trillion（約880兆円） through 2030, reflecting higher debt financing and larger capacity expansion among major cloud providers. The bank argues the economics are holding up so far, with hyperscalers maintaining profitability and strong cash flow, though investors remain uncertain whether demand will grow fast enough to justify the massive buildout.
Summaries like this, in your inbox every morning.
Sign up free →What happened
JPMorgan increased its estimate of total global AI-related capital expenditures to $5.5 trillion（約880兆円） through 2030, up from $5.1 trillion（約820兆円）, with debt financing rising to $4.1 trillion（約660兆円）. The bank projects hyperscalers will spend $650 billion（約100兆円） in 2026 and exceed $1.1 trillion（約180兆円） in 2027, while maintaining operating cash flow projected to surpass $900 billion（約140兆円） by 2027.
Why it matters
The bank's analysis suggests AI investment is not just a speculative bubble—large cloud providers are entering this phase from a position of strength, and financing conditions remain favorable with loan-to-cost ratios above 85%. However, investors remain divided on whether AI demand will grow quickly enough to match the capacity being built, and concentration of spending among a small group of technology companies means any slowdown could have an outsized impact on industry growth.
What to watch
JPMorgan forecasts that high-grade corporate debt will account for more than $2.1 trillion（約340兆円） in data center financing over five years, with $150 billion（約24兆円） in U.S. hyperscaler debt issuance and another $100 billion（約16兆円） equivalent abroad anticipated in 2026. The critical open question is whether adoption will keep pace with the trillions being invested.
No comments yet. Be the first to share your thoughts!
Log in to join the discussion
Get curated AI news from 200+ sources delivered daily to your inbox. Free to use.
Get Started FreeFree · takes 30 seconds · unsubscribe anytime
5 minutes a day. The AI essentials.
200+ sources · Email / LINE / Slack
