Jim Cramer Argues Big Tech Cannot Skimp on AI Spending

CNBC's Jim Cramer argued on air that cloud computing giants cannot afford to be cheap on AI infrastructure spending, saying the customers for large-scale compute "already are coming," CNBC reports. He cited Amazon and its cloud business, Amazon Web Services, as an example and CNBC reports Amazon has committed about $200 billion in capital expenditures this year, with much of that directed to expanding data-center capacity. Cramer warned, "If you don't build the stadium, they are going elsewhere and you will leave a lot of money on the table," CNBC reports. CNBC's coverage names OpenAI and Anthropic among large customers searching for compute partners.

Companies building for AI workloads are primarily sizing for GPU/accelerator capacity, networking bandwidth, and power/cooling. Industry observers note that demand for GPU-dense racks and low-latency networks can outpace data-center build cycles, producing short-term scarcity even when cloud providers plan large capex programs.

Reporting frames this commentary as part of a broader narrative about an infrastructure spending race among major cloud vendors, where capacity expansion and long lead times for specialized hardware create competitive pressure. Observers tracking the market point to public capex disclosures, large leases, and hardware vendor supply as practical indicators of that competition.

For practitioners: track quarterly capex guidance and data-center expansion announcements from major cloud providers, large enterprise or AI vendor procurement agreements, and GPU/accelerator shipment reports. These signals indicate whether raw compute availability will tighten or loosen for AI training and inference workloads.