NVIDIA shares rise after Amazon outlines higher AI spending and continued reliance on its chips

NVIDIA’s stock moved higher on Friday after Amazon said it has raised its planned capital expenditure to $220 billion and, in doing so, underscored the role NVIDIA’s graphics processing units and AI software play in building large-scale artificial intelligence systems. The development comes as cloud providers race to expand capacity for training and deploying AI models across data centers, a spending category that has become a key driver for semiconductor demand.

According to the report, Amazon also reaffirmed NVIDIA’s importance despite continuing efforts to reduce costs and improve performance by developing its own AI chips. The company’s Trainium processors are designed to accelerate machine-learning workloads on AWS, and their progress is intended to give Amazon more flexibility over its supply chain and economics.

The market reaction suggests investors viewed Amazon’s larger spending plan as net positive for the broader AI hardware ecosystem, including suppliers of accelerators and the supporting tools needed to run AI at scale. While Amazon’s in-house chips can affect long-term mix, the report indicates Amazon did not portray that shift as a substitute for NVIDIA’s role in the near term.

For NVIDIA, the immediate takeaway is not just that AI budgets are rising, but that a major customer continues to pair its own silicon strategy with NVIDIA’s platform. NVIDIA has long built an AI “stack,” combining accelerated compute with software layers that help developers and data center operators run workloads efficiently. In cloud contexts, that stack can matter as much as chip performance because it can reduce integration risk when deploying new models and scaling to larger clusters.

The report’s emphasis on Amazon’s capex plan also highlights how cloud operators are funding AI expansion. Capital expenditures typically translate into power, cooling, networking, and racks of specialized compute. When a cloud provider increases that envelope, it can increase total demand for accelerators, whether the accelerators come from a mix of third-party suppliers or the customer’s own custom designs.

Sector context matters because the AI supply chain has been unusually sensitive to forward spending indicates. In recent years, semiconductor investors have often treated cloud capex commentary as an early indicator of how quickly data center buildouts will translate into actual purchases of AI hardware and related components. In that environment, even a reaffirmation of existing dependencies can move sentiment.

Still, investors will be watching for how Amazon’s balance between NVIDIA GPUs and its Trainium chips evolves. The report does not provide detailed disclosure on how much of Amazon’s incremental AI capacity is expected to be served by NVIDIA versus internal accelerators, nor does it specify any schedule for shifting workload mixes.

What to watch next is whether additional guidance clarifies the expected allocation of new AI infrastructure spending. That includes whether Amazon outlines capacity timelines, workload allocation principles for Trainium versus third-party systems, and how those choices flow through to procurement trends for NVIDIA and other component suppliers.