Qualcomm, Amazon Expand AI Chip Partnership

Qualcomm and Amazon are expanding their collaboration on customized AI data center silicon and optical connectivity, with Amazon receiving warrants to acquire about US$4 billion of Qualcomm stock as the companies target AI inference infrastructure.

Qualcomm is deepening its push into the data center market through a multi-generational collaboration with Amazon that will cover customized silicon for AI inference, optical connectivity, and the use of Amazon Web Services (AWS) infrastructure for chip design. The agreement gives Amazon warrants to acquire 25 million Qualcomm shares at US$161.26 per share, representing an investment of about US$4 billion. The warrants vest in tranches tied to commercial arrangements and purchases of up to US$60 billion of Qualcomm server chips and related technology, according to a regulatory filing.

The partnership comes as Qualcomm seeks to establish a larger position in AI infrastructure beyond its traditional smartphone processor business. It also gives the company another hyperscaler customer as cloud providers expand their use of customized chips to manage the cost and performance requirements of AI workloads.

Qualcomm Targets AI Inference and Connectivity

Qualcomm Technologies and AWS will collaborate across multiple generations of customized silicon designed to support Amazon's growing AI infrastructure, with an initial focus on AI inference. Inference refers to the process of running trained AI models to generate outputs, rather than training the models themselves. As AI applications move into production, inference has become a growing source of demand for computing capacity and energy-efficient infrastructure.

The companies will also work on optical connectivity solutions designed for high-bandwidth data center networks. Qualcomm said the collaboration will extend to solutions supporting connectivity of up to 1.6TB per second, using its SerDes and optical digital signal processor technologies.

“As AI demand accelerates, data center infrastructure will require advances in both computing and connectivity to deliver greater performance with more efficiency,” says Cristiano Amon, President and CEO, Qualcomm Incorporated. Amon says the collaboration combines Qualcomm's experience in advanced processing, power-efficient computing, silicon design, and system integration with AWS infrastructure to address the next generation of AI data center requirements.

The agreement also expands Qualcomm's use of AWS AI infrastructure, including Amazon Bedrock, for electronic design automation workloads. The companies intend to use those services to reduce chip design cycles. “This collaboration with Qualcomm Technologies builds on a strong foundation of partnership and reflects our shared commitment to pushing the boundaries of what’s possible,” says Prasad Kalyanaraman, Vice President, AWS.

Kalyanaraman says the companies are combining customized silicon and connectivity to develop infrastructure focused on performance, efficiency and cost.

A Broader Push Beyond Smartphones

For Qualcomm, the Amazon agreement extends a strategy to diversify its business as the company prepares for changes in its smartphone market. Qualcomm has traditionally been associated with processors and connectivity technologies for mobile devices. The company has increasingly targeted data centers as AI workloads create demand for alternatives to established accelerator and processor suppliers.

In June, Qualcomm introduced the Dragonwing C1000 data center CPU and said Meta plans to use the processor when production begins in 2028. Qualcomm has also outlined a broader data center roadmap that includes AI chips and technologies designed to connect multiple chips.

The company has set a target of US$15 billion in data center sales for fiscal 2029. The Amazon agreement provides another indication that hyperscalers are evaluating Qualcomm's technology for large-scale infrastructure.

The deal also places Qualcomm in a market where Nvidia remains dominant in AI computing. Nvidia's position has been built largely around GPUs, which are widely used for training and running demanding AI workloads, while CPUs remain important for general-purpose computing and increasingly complex AI and agentic workloads.

Bank of America estimates that the CPU market could grow from US$27 billion in 2025 to US$60 billion by 2030, according to CNBC.

Financing Tied to Infrastructure Demand

The structure of the Qualcomm-Amazon agreement links part of the financial relationship between the companies to commercial execution. Amazon's warrants expire on Sept. 3, 2036. Qualcomm said the warrants will vest in tranches based on the execution of specified commercial arrangements, including purchases of Qualcomm server chips and other technology.

The arrangement reflects the scale of investment required to build AI infrastructure while giving Amazon a potential equity position in Qualcomm. For Qualcomm, the commercial commitments offer a path to expand its presence in data center computing while broadening its relationship with one of the world's largest cloud providers.