Jul. 2 at 4:25 PM
The scale of AI infrastructure spending from the major hyperscalers continues to expand at a rapid pace, with
$MSFT,
$GOOGL,
$AMZN, and
$META collectively guiding roughly ~
$725B in AI-related capex this year, up significantly from prior years.
Over a multi-year lens, this represents a sharp acceleration in capital deployment, driven by sustained demand for compute, storage, networking, and power infrastructure.
What stands out in the current cycle is not just the growth rate, but the visibility of future demand - with reported backlog levels across cloud platforms still indicating large unmet capacity, particularly in Azure and Google Cloud.
Importantly, this spending is not evenly distributed - a meaningful portion continues to flow into chips, memory, and high-performance networking, which are becoming increasingly critical to AI workload scaling.
From a market perspective, what looks like “capex pressure” is increasingly also “future revenue visibility” across the semiconductor and infrastructure ecosystem.
The key question is not whether spending is high - but how long supply constraints keep it structurally elevated.
