Synopsys and Amazon Web Services have entered into a multi-year deal worth more than $1 billion to further give Amazon access to Synopsys silicon IP and engineering technologies for custom chip development.
The deal is Amazon becomes the anchor customer for a larger Synopsys portfolio of application-optimized silicon IP, rather than the more traditional chip components the company licenses. The deal aims to enable “increasingly complex custom silicon and AI infrastructure,” according to Synopsys said.
Synopsys offers electronic design automation software and semiconductor IP that are applied in the design of advanced semiconductors. It made $1.75 billion in its most recent fiscal year from its IP business, which it is in competition with companies like Arm.
The new deal is a move toward more specialized blueprints for specific workloads on Synopsys’ part. The two companies have not said which Amazon chips will be using the newly licensed IP.
Over the years, Amazon has been working on its own processors for cloud and AI workloads. It features cloud infrastructure and security with Nitro, general-purpose processors with Graviton and AI training and inference with Trainium. The companies have been collaborating for over 15 years. Synopsys outlined the expanded relationship.
The deal also expands Amazon’s use of Synopsys engineering technologies beyond silicon IP. The companies aim to expand the collaboration to electronic design automation, simulation and analysis, and agentic AI tools, such as enhancing engineering workflows from chip design to system-level validation.
One of the interesting financial shifts is the switch to a license-plus-royalty model for the expanded IP business at Synopsys. In that arrangement, future payments may be raised as production volumes grow, which would peg a portion of Synopsys’ revenue to the commercialization of Amazon’s custom silicon. According to Investing.com’s coverage, the deal is a “growth in Synopsys’ silicon IP business into more application-specific designs.
The partnership also operates in the other direction. Synopsys will use Amazon Elastic Compute Cloud, AWS storage services and Amazon Bedrock for product development and AI applications.
As part of the collaboration, the companies plan to optimize Synopsys software for Amazon’s Trainium and Graviton platforms.
The agreement comes as cloud firms are increasingly creating custom processors rather than using off-the-shelf chips.
Cloud providers can optimize performance, power consumption and infrastructure costs by custom silicon, which can be tailored to specific workloads.
That’s the strategy that is being evident in Amazon’s chip portfolio. Graviton is intended for general-purpose cloud computing, while Trainium will be used for AI workloads.
The company’s ongoing commitment to purpose-built processors has also made it more important than ever to have tools to deal with the complexity of modern chip and system design.
Synopsys is framing the deal as part of a broader transition to AI-powered engineering. The companies intend to use AI throughout the silicon-to-system design and analysis, optimization and validation workflows.
Synopsys will also be creating more agentic AI features for Amazon’s engineering teams, the company announced.
The partnership bolsters AWS’ ability to access to specialized expertise in chip design, adding to its custom semiconductor portfolio. Amazon will be a significant customer for Synopsys’ newer business of application-optimized IP and a larger user of the company’s engineering software.
The terms of the deal don’t detail which specific AWS processors will use Synopsys’ new blueprints, so the commercial implications for specific chips are not yet known.
What is obvious is that the partnership strengthens the collaboration between two companies that are more and more concentrated on AI infrastructure and custom silicon.
