CoreWeave’s pull on profitable AI workloads raises questions for Amazon Web Services

Amazon Web Services’ latest growth figures are coming out strong, but a new market report suggests AWS may be facing a more selective competitive threat in artificial intelligence infrastructure, where the economics of compute and data handling can differ from broad cloud demand.

The report frames the backdrop as an AWS period of unusually fast expansion, describing it as the fastest growth in 18 quarters. In the same breath, it highlights a separate development: a competitor, CoreWeave, is described as routing a large share of what the report characterizes as the most profitable AI workloads away from Amazon’s cloud.

CoreWeave, in this telling, is positioned less as a generic alternative to “big cloud” and more as a specialized destination for AI workloads that can generate higher margins for providers offering the right combination of accelerated hardware, capacity management, and operational tailoring. The report’s core claim is that the companies winning the highest-value AI demand may not line up neatly with the companies showing the biggest gains in overall cloud spending.

The piece also draws a comparison between CoreWeave and “cloud giants,” implying a gap between traditional hyperscaler competition and the way leading AI developers buy and deploy compute. While hyperscalers sell broadly across workloads, the argument here is that AI execution can concentrate quickly in environments optimized for training and inference at scale, with supplier relationships that may become “sticky” once established.

For Amazon, the report’s message is less about whether AWS is growing and more about what kinds of workloads are growing. If AI demand shifts toward a provider that emphasizes dedicated capacity for accelerated compute, AWS’s headline growth rate could still look good while its share of the highest-value segments tightens or becomes harder to expand.

CoreWeave has been part of the AI infrastructure conversation for some time, particularly as investors and customers have sought alternatives to standard, multi-tenant cloud approaches for large-scale models. In the report, that focus is used to argue that not all cloud growth is equal, because the most profitable AI workloads may not follow the same allocation patterns as lower-margin, generalized workloads.

Amazon, through AWS, competes in a market that increasingly treats artificial intelligence as a foundational workload category rather than an add-on. AWS has published extensive company and business updates through its newsroom, including items related to AWS operations and product developments, underscoring that AI is central to the company’s broader cloud strategy.

Still, the market report offers limited detail on the specific routing mechanics or contract structures behind the “around Amazon entirely” framing, and it does not, in the material available here, disclose customer names, workload volumes, or pricing terms that would allow outsiders to quantify the competitive impact.

What to watch next is whether AWS continues to translate strong overall growth into broader AI wins, including new customer commitments and expansion of capacity for accelerated computing. Separately, investors will likely look for evidence that CoreWeave can sustain its position as AI workloads scale, and whether those workloads remain concentrated there or begin to diversify across multiple cloud suppliers.