Amazon and Synopsys sign billion-dollar deal to accelerate AI chip design

Multi-year agreement sees Amazon license Synopsys IP while chip tools maker adopts AWS cloud services

By LineZotpaper
Published
Read Time2 min
Amazon and chip design software company Synopsys have entered a multi-year agreement reportedly worth over a billion dollars, deepening their cooperation on AI chip design, with Amazon licensing Synopsys intellectual property and expanding its use of electronic design automation (EDA) tools, while Synopsys will optimise its multiphysics solutions for Amazon's Trainium and Graviton chips and adopt AWS cloud services.

The deal covers both companies providing services to each other. Amazon will license Synopsys' silicon IP designs and expand its use of Synopsys' EDA software tools, including AI-powered engineering software, to accelerate development of custom AI chips and AWS infrastructure hardware. From its side, Synopsys will work to optimise its multiphysics solutions for Amazon Trainium and Graviton chips, will adopt AWS cloud computing and storage services, and will begin using Amazon Bedrock to build and deploy AI applications and agents for its own development work.

The partnership comes as Synopsys and its competitors move towards automating greater portions of the chip design process using AI. Amazon ensuring that process is optimised for its own hardware places it in a more favourable position as chip design becomes easier and faster. This is particularly relevant as major AI companies look to develop their own inferencing hardware to reduce reliance on expensive Nvidia GPUs.

The core of the deal centres on chip design collaboration. Amazon will incorporate Synopsys' application-optimised IP blueprints into its own chips and use Synopsys' AI-powered engineering software. Many major AI developments in 2026 have involved AI agents, leading to hardware shortages and a race to fill that gap with optimised hardware. The Synopsys/Amazon deal could position both companies to take advantage of that demand.

§

Analysis

Why This Matters

  • The deal signals a strategic push by Amazon to reduce dependence on Nvidia for AI hardware by deepening its own chip design capabilities through Synopsys' tools and IP.
  • For Synopsys, the agreement secures a major customer and promotes its EDA and AI software on AWS, potentially influencing other chip designers to adopt the same stack.
  • The collaboration could accelerate the development of custom AI accelerators, affecting the broader AI infrastructure market and cloud pricing.

Background

Amazon has been developing its own server chips for several years, including Graviton CPUs and Trainium AI accelerators, aiming to offer more cost-effective alternatives to Nvidia and Intel. Synopsys is a leading supplier of electronic design automation software and semiconductor IP, used by most major chip companies. The trend toward AI-powered chip design tools has been growing, with Synopsys and rivals like Cadence introducing platforms that automate parts of the design process.

Key Perspectives

Amazon: Seeks to optimise its chip design pipeline and reduce external dependencies, while making its cloud services more attractive to chip designers. Synopsys: Gains a long-term revenue stream and a reference customer for its AI-enhanced EDA tools, as well as a commitment to use AWS internally. Critics: The deal's success depends on whether the optimisations actually deliver performance gains that translate into cheaper or more powerful AI hardware. There is also the risk that such exclusive partnerships could limit competition in chip design tools.

What to Watch

  • Whether Amazon announces new Trainium or Graviton chips specifically enabled by Synopsys IP or tools.
  • Competitor responses: Nvidia's reaction to Amazon's push and any similar deals between other cloud providers and EDA vendors.
  • Adoption of Synopsys' AI agent-driven chip design platforms by other customers, as a signal of industry shift toward automated design.

Sources

Zotpaper

Written by software from the reporting listed above, scored by an automated standards desk, and published without a person reading it first. If something here is wrong, tell the editor and it will be put right.