Amazon Web Services (AWS) has signed a multiyear agreement with Synopsys worth more than $1 billion to license chip design intellectual property, deepening the cloud provider’s push to develop increasingly sophisticated custom processors for AI and cloud infrastructure.
The agreement expands a 15-year relationship between the companies and makes Amazon the lead customer for Synopsys’ expanding portfolio of application-optimized silicon IP. AWS has multiple custom chips, including Graviton CPUs for general-purpose cloud workloads, Trainium processors for AI training and inference, and Nitro chips for cloud security, networking and storage, but the two companies did not disclose which specific AWS processors will incorporate the licensed designs.
The Synopsys agreement gives AWS access to reusable chip design building blocks optimized for particular applications rather than strictly using conventional, general-purpose IP. Synopsys has historically supplied IP for relatively standardized components, but the company has increasingly expanded into more sophisticated designs, including computing cores and application-specific technologies.
For AWS, the agreement comes as Amazon continues to invest heavily in custom silicon as a way to control more of the technology underlying its cloud and AI infrastructure. Purpose-built processors can be tailored to specific workloads, allowing AWS to optimize performance, power consumption and system-level costs.
The companies are also expanding their collaboration beyond conventional chip IP. Amazon will increase its use of Synopsys electronic design automation, simulation and analysis technologies, while the two companies plan to work on multiphysics solutions for Graviton and Trainium.
Synopsys said the companies will collaborate on applying AI throughout the silicon-to-system engineering process, including the design, analysis, optimization and validation of chips and systems. The work will include Synopsys AI-powered EDA and agentic AI technologies, as well as customized AI capabilities for Amazon’s engineering teams.
The relationship is mutual and bidirectional. Synopsys plans to use Amazon Elastic Compute Cloud (EC2), cloud storage (S3) and Amazon Bedrock to accelerate development of its own software and AI applications.
The agreement reflects how hyperscalers are becoming as much chip designers as semiconductor customers. Companies such as Amazon, Google and Microsoft have invested in internally developed processors to tailor computing infrastructure to their particular workloads, particularly as AI has increased the cost and complexity of data-center computing.
