Alphabet and Amazon Commit $420B to AI Infrastructure, Boosting Nvidia, Broadcom, Micron and Sandisk

Alphabet and Amazon are on track to spend a combined $420 billion on capital projects in 2026, with GOOGL guiding to roughly $195 billion to $205 billion and AMZN to about $220 billion . These are two separate company budgets rather than a joint venture, but taken together they amount to one of the largest single-year infrastructure commitments in corporate history, and the bulk of it lands with the same short list of hardware suppliers.

Four names sit directly in the path of that spending: NVDA, still the default accelerator for large-scale training and inference; AVGO, which co-develops the custom silicon behind Google's TPU program; and the memory suppliers MU and SNDK. Memory is the detail worth pausing on. Rising memory prices are what pushed Amazon's own 2026 plan up from about $200 billion to $220 billion , which means memory vendors are capturing incremental budget rather than simply riding a fixed one.

The two companies are not outliers. The five largest hyperscalers are collectively projected to spend in the region of $660 billion to $725 billion on capex this year, with roughly three quarters of it tied to AI infrastructure, and Nvidia has told investors it expects hyperscaler AI spending to top $1 trillion next year. Supply remains the binding constraint: memory tightness is expected to persist into 2028, which supports pricing for MU and SNDK but also raises the bill for everyone building data centers.

The question for investors is conversion, not commitment. Guided capex is a plan, and the useful signal over the next few quarters is how much of it shows up as booked revenue and backlog at the four suppliers, how memory contract pricing settles as new capacity arrives, and whether the depreciation load from this build begins to visibly compress free cash flow at GOOGL and AMZN. A spending cycle this large tends to reward the suppliers before it rewards the spenders.

Related Stocks

• GOOGL
• AMZN
• FLEX