Key facts
- Qualcomm and Amazon will collaborate on custom silicon for AI data centers.
Qualcomm and Amazon have announced a multi-generational collaboration to develop custom chips for AI data centers. The partnership will focus on AI inference chips and high-speed optical connectivity solutions, aiming to enhance performance and efficiency for growing AI workloads.

This collaboration signifies a significant move in the race to build more efficient and powerful AI infrastructure, potentially impacting the competitive landscape for AI chip development and cloud services.
Qualcomm and Amazon have entered into a multi-generational collaboration to develop custom silicon for artificial intelligence data centers, focusing on AI inference and high-speed optical connectivity. The partnership aims to address the rapidly growing demand for compute, storage, and networking bandwidth driven by AI workloads.
Under the agreement, the companies will work together on chips designed for AI inference, which involves running trained AI models. This area has become a critical battleground for semiconductor firms. Additionally, they will develop advanced optical connectivity solutions capable of supporting up to 1.6 terabits per second, crucial for managing the increasing bandwidth requirements within AI data centers.
As part of the collaboration, Qualcomm plans to increase its utilization of Amazon Web Services (AWS) AI infrastructure, including services like Amazon Bedrock, for its electronic design automation (EDA) workloads. This move is intended to reduce chip design cycles and accelerate the development of next-generation AI hardware.
Qualcomm's President and CEO, Cristiano Amon, stated that the collaboration with AWS on customized silicon and connectivity solutions will bring decades of leadership in advanced processing and power-efficient compute to deliver breakthrough performance for AI infrastructure. Prasad Kalyanaraman, Vice President at AWS, highlighted the shared commitment to pushing boundaries and delivering more performant, efficient, and cost-effective infrastructure for customers.