How I learned to stop worrying and love hyperscaler capex
The New Stack Alex Wilhelm
AI cloud spending looks scary, but the numbers may not back up the panic. Amazon, Google, and Microsoft’s cloud units seem to be getting better at turning more capex into more growth.
Based on reporting by The New Stack, Alex Wilhelm — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
The loudest complaint around AI has shifted again. After the doomsayers missed on “no data,” then “no real use case,” then “too expensive to use,” the new target is hyperscaler capex: Amazon, Alphabet, and Microsoft are supposedly pouring too much money into AI infrastructure for too little return.
The New Stack article pushes back by looking at the cloud businesses behind that spending. The case it makes is simple enough: when you pull together the data from Amazon, Alphabet, and Microsoft, growth appears to be speeding up rather than slowing down. That matters because the whole argument against heavy infrastructure spending depends on demand failing to show up.
And the financial picture, at least in this read, gets better rather than worse at larger scale. The piece says hyperscaler profitability scales with scale, and that capex efficiency is improving. So the worry isn’t that these companies are blindly burning cash. It’s that they may be buying the machinery for a market that’s still expanding under everyone’s feet.
My take — AI-written commentary, not fact-checked reporting
The industry loves a good capex panic because it sounds wiser than admitting demand might actually be real. But when Amazon, Alphabet, and Microsoft are all still pushing harder on infrastructure, the boring explanation is usually the right one: the money is following usage. The bigger mistake is assuming AI only looks expensive before it starts looking inevitable.
Read more about this at: The New Stack
