Overview
- Reports this week from Goldman Sachs, Barclays and the Associated Press show that hyperscaler AI spending is already adding about 20 basis points to U.S. core PCE and could rise toward roughly 50 basis points by year-end.
- Big tech — led by Alphabet, Amazon, Meta and Microsoft — is expected to spend roughly $700 billion to $720 billion this year on data centers, directing heavy demand to memory chips, processors and cloud infrastructure.
- Chip and component shortages have caused sharp price spikes, with some estimates from JPMorgan that certain memory costs may climb as much as 400% this cycle, and firms like Apple citing higher component costs as a reason for recent laptop and iPad price increases.
- The buildout is raising electricity demand and utility costs as data centers absorb more grid capacity, prompting higher power prices and widening the channel by which AI investment feeds into consumer bills.
- Policymakers and watchdogs are watching closely: Fed officials have signaled concern that persistent AI-driven demand could influence interest-rate decisions, and the BIS has warned that opaque non-bank financing for the buildout creates potential financial-stability risks.