
The AI buildout just met the bond market
Jamie Dimon’s latest message is basically: the AI gold rush is real, but the tab is starting to look like it needs its own spreadsheet. He told Bloomberg that rising rates, government borrowing, and a mountain of AI infrastructure spending are all competing for the same pool of capital.
The headline number making the rounds? JPMorgan estimates AI capex could jump from about $700 billion this year to $1 trillion next year. That’s not a typo — that’s “build the future” money. But as more debt gets piled onto data centers, chips, and power hungry infrastructure, investors have to ask a very unsexy question: who pays for all this, and at what interest rate?
The bill is not landing evenly
Not every AI giant is feeling the squeeze the same way:
- Alphabet reportedly posted its first ever quarter of negative free cash flow, even while lifting its 2026 capex forecast.
- Amazon raised its 2026 spending plan too, and its trailing free cash flow slipped below zero.
- Microsoft, meanwhile, is still throwing off plenty of cash even after a huge capex bill.
So yes, the AI story is still alive. But it’s starting to look less like “easy money printing machine” and more like “very expensive group project.”
Rates up, enthusiasm still up
Dimon’s not exactly pounding the table against AI. He basically said the long term payoff can still be worth it, the same way the internet era eventually justified a ton of spending that looked wild at the time.
The catch is that higher rates change the math. Government borrowing for consumption is one thing; financing productive infrastructure is another. And when the 10 year Treasury is hanging around levels we haven’t seen in ages, the capital markets start acting less like a buffet and more like a bouncer with a clipboard.
Big picture: AI may still be a monster opportunity, but the cost of feeding that monster is getting bigger. That matters because when capital gets pricey, the companies with the strongest balance sheets get to keep dancing longest.