Qualcomm, Amazon strike $4 billion custom AI chip deal: What are they building for next-gen data centres?

Qualcomm Technologies and Amazon Web Services (AWS) have announced a multi-generation collaboration to develop customised silicon for large-scale AI data centres, with a focus on AI inference and high-performance connectivity.

Qualcomm shares jumped about 10% on Tuesday (8 September) following the announcement, as investors responded to the chipmaker's expanded push into data centre infrastructure.

Qualcomm said on Tuesday it had issued a warrant to Amazon, giving it the right to buy an about $4 billion stake, as part of a collaboration to develop custom chips for AI data centers. The warrant lets Amazon buy Qualcomm shares at $161.26 apiece and is linked to up to $60 billion of business under the deal, according to a regulatory filing.

Qualcomm has traditionally been associated with processors used in smartphones and other mobile devices. That began to change in June, when the company unveiled the Dragonfly C1000, a central processing unit designed for data centres.

Qualcomm also announced that Meta plans to deploy the chip in production starting in 2028.

Qualcomm, Amazon target custom silicon for AI data centres

Under the expanded partnership, Qualcomm will work with Amazon Web Services (AWS) to develop customised chips for AI data centres. The chips will be designed to handle AI inference, the process through which AI models generate responses and other outputs, while delivering the computing power needed to meet growing AI demand.

The collaboration will allow Amazon to tailor the chips to the specific requirements of its AI infrastructure, rather than relying on a one-size-fits-all approach.

The collaboration brings together AWS's cloud and AI infrastructure capabilities with Qualcomm's expertise in power-efficient processing, silicon design and system-level integration.

As AI workloads expand, data centres are facing growing requirements for computing power, storage, networking, memory bandwidth and energy efficiency. The companies said their collaboration is aimed at addressing those demands with more efficient infrastructure.

Qualcomm and Amazon to develop 1.6T optical connectivity

The partnership will also focus on high-performance optical connectivity for AI data centres, with solutions capable of supporting speeds of up to 1.6T.

Qualcomm and AWS will use Qualcomm’s SerDes and optical digital signal processor (DSP) technologies to help data centres handle the growing amount of data and bandwidth required by AI workloads.

Qualcomm to use AWS AI infrastructure for chip design

As part of the expanded collaboration, Qualcomm also plans to increase its use of AWS AI infrastructure, including Amazon Bedrock, for electronic design automation (EDA) workloads.

The companies expect the move to help Qualcomm reduce chip design cycles by using cloud-based AI infrastructure in its development process.

"As AI demand accelerates, data center infrastructure will require advances in both computing and connectivity to deliver greater performance with more efficiency," said Cristiano Amon, President and CEO, Qualcomm Incorporated.

"Qualcomm is pleased to work with AWS on customized silicon and connectivity solutions, bringing decades of leadership in advanced processing and power-efficient compute, to deliver breakthrough performance and enable the next generation of AI infrastructure."

AWS calls Qualcomm partnership a step towards efficient AI infrastructure

Prasad Kalyanaraman, Vice President, AWS, said the expanded collaboration builds on the companies' existing partnership.

"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," he said.

“By working together on customized silicon and advanced connectivity, we're delivering more performant, efficient, and cost-effective infrastructure for our customers.”