Qualcomm and Amazon partner in $60 billion multi-generational AI data center deal

Marijan Hassan - Tech Journalist

11 minutes ago

2 min read

Qualcomm Technologies has finalized a landmark multi-generational collaboration with Amazon Web Services (AWS) to co-develop customized artificial intelligence inference silicon and next-generation optical connectivity for cloud-scale data centers. The agreement marks Qualcomm’s largest data center win to date as the mobile chip giant aggressively expands beyond smartphone processors into enterprise AI infrastructure.

Following the announcement, Qualcomm shares jumped 10%, driven by investor optimism over its growing presence in cloud compute.

Custom Inference Silicon and $60 Billion Contingent Warrant

Under the multi-year agreement, Qualcomm and AWS will co-engineer custom accelerators specifically optimized for AI inference - the operational phase where trained models process user queries and generate outputs at scale.

To align long-term commercial incentives, Qualcomm issued a stock warrant allowing an Amazon affiliate to purchase up to 25 million shares of Qualcomm common stock at $161.26 per share (a potential $4 billion equity stake). Warrant vesting is directly tied to Amazon procuring up to $60 billion in Qualcomm server chips, system hardware, and optical networking technologies over the next decade, with an initial tranche of 3.75 million shares vesting immediately upon contract execution.

Key technical aspects of the collaboration include:

Targeting High-Efficiency Inference: The partnership leverages Qualcomm’s low-power compute architecture to deliver cost-effective, energy-efficient inference capacity across AWS data centers, providing an alternative to power-dense accelerators from market leader Nvidia.
1.6 Terabit Optical Interconnects: The companies will co-develop high-bandwidth optical interconnect solutions capable of speeds up to 1.6 terabits per second (1.6T). Using Qualcomm’s SerDes and optical Digital Signal Processor (DSP) technologies, the hardware aims to eliminate network latency bottlenecks across massive AI clusters.
Reciprocal Cloud & EDA Integration: Qualcomm will expand its internal use of AWS cloud infrastructure, including Amazon Bedrock, to run electronic design automation (EDA) workloads, leveraging generative AI tools to speed up its future silicon design cycles.

Diversifying the AI Hardware Ecosystem

The partnership provides AWS with another customized silicon pipeline alongside its in-house Trainium and Inferentia chips, helping the cloud provider manage rising infrastructure costs and meet surging global compute demand.

For Qualcomm, the deal follows a recent contract to supply data center CPUs to Meta, cementing the chipmaker's transition into a primary hardware provider for top-tier hyperscalers. Executives from both companies confirmed that co-engineered custom inference silicon and optical connectivity hardware will begin deploying into AWS availability zones in phases starting next year.