Qualcomm’s push beyond smartphones gathered pace with its June unveiling of the Dragonfly C1000 data-center CPU, built for agentic AI workloads and designed to deliver performance without excessive power use. Meta is slated to use the processor when production begins in 2028, giving Qualcomm an early foothold in a market it has said could generate $15 billion in data-center sales by fiscal 2029.¹

On Tuesday, that ambition received a far larger endorsement. Qualcomm and Amazon announced a multigenerational collaboration on customized silicon for AWS AI infrastructure, with a particular focus on inference—the stage at which trained models answer queries and perform tasks. The companies also plan to develop “high-performance optical connectivity solutions,” widening the agreement beyond processors alone.²

The commercial terms underscore the stakes. Qualcomm issued Amazon warrants for 25 million shares at $161.26 apiece, potentially worth $4 billion. The warrants vest in tranches tied to commercial arrangements, including purchases of up to $60 billion in Qualcomm server chips and other technology, and expire in September 2036.³ Qualcomm shares rose 4% after the announcement.³

For AWS, the partnership adds another supplier as demand for compute, storage, networking, memory bandwidth and energy-efficient infrastructure rises. Qualcomm will also use Amazon’s AI servers to accelerate its own chip-design work—a reciprocal arrangement that makes AWS both customer and development partner.²

The deal does not dislodge Nvidia, whose GPUs remain central to training sophisticated models and handling the biggest AI workloads. But Qualcomm’s case rests on a changing bottleneck: CPUs are increasingly important for AI systems, especially agentic workflows. AWS’s commitment suggests the contest is no longer only about who makes the fastest accelerator, but who can assemble the most efficient full data-center stack.⁴