Qualcomm’s AWS Deal Is a $4 Billion Bet on Breaking Nvidia’s AI Grip

Qualcomm’s AWS Deal Is a $4 Billion Bet on Breaking Nvidia’s AI Grip
Qualcomm’s long push beyond smartphone processors took a major turn on Tuesday with an agreement to work with Amazon across “multiple generations” of customized silicon for AWS AI infrastructure, with an emphasis on inference—the work of running trained AI models at scale.

The companies also said they will develop high-performance optical connectivity, broadening the pact beyond processors and into the networking systems needed to move huge volumes of AI data. Qualcomm, in return, plans to use Amazon’s AI servers to accelerate its own chip-design work.

The commercial terms underline how consequential both sides believe the relationship could become. Qualcomm issued Amazon warrants for 25 million shares at $161.26 each, representing a potential $4 billion investment. The warrants vest in stages based on commercial milestones and purchases of up to $60 billion in Qualcomm server chips and other technology; they expire on September 3, 2036.

For Qualcomm, the deal is a show of confidence from a hyperscaler as it tries to translate its reputation for efficient mobile silicon into the data center. Its June unveiling of the Dragonfly C1000 CPU, slated for Meta deployment when production begins in 2028, had already marked that shift. The company is targeting $15 billion in data-center revenue in fiscal 2029.

Amazon’s perspective is more pragmatic: AI workloads are multiplying, and the infrastructure needs compute, storage, networking, memory bandwidth and energy efficiency. The companies say the alliance combines AWS’s AI platform with Qualcomm’s processing and system-integration expertise.

That does not dislodge Nvidia overnight. Nvidia remains the dominant GPU supplier for the most demanding AI training and inference tasks, while CPUs handle broader sequential workloads. But Qualcomm’s AWS partnership signals that the AI buildout is becoming a wider silicon contest—one where the bottleneck may no longer be only GPUs. As Nvidia’s AI infrastructure chief Dion Harris put it, “CPUs are becoming the bottleneck” in expanding AI and agentic workflows.

CNBC separately characterized the arrangement as Qualcomm working with Amazon on multiple generations of customized silicon to expand AWS’s AI infrastructure.

Continue reading https://foxvector.com/stories/01a089d7-2864-37f9-72bc-2e4aad659e15