Build1 publisher2 min readPublished
Amazon's $1bn-plus Synopsys deal tunes chip-design simulation for Trainium and Graviton
Amazon will license Synopsys IP and expand its use of Synopsys chip-design tools in a multi-year deal reported at over $1 billion. Synopsys's pledge to tune its simulation software for Trainium and Graviton does more for AWS's position than the license does.
The Engineer · Build desk
Drafted by a language model from the sources cited here and checked against its claim ledger before publication. How we use AISend a correction
What happened
- Amazon will draw on Synopsys blueprints of application-optimized IP, silicon designs it can build into its own chips, and use Synopsys AI engineering software on custom hardware.
- Amazon already markets its Graviton 5 processor for CPU-intensive agentic AI workloads.
- Amazon's announcement also names Trainium for AI training and Nitro for cloud security, networking and storage, suggesting the work could reach several chip lines.
- Synopsys will run its own IP and software development on AWS compute and storage and use Amazon Bedrock to build AI agents for its products.
Compiled by The EngineerSomething wrong?How this is made
Why it matters
- decision Engineering teams running Synopsys multiphysics jobs get a vendor-tuned option on AWS's own chips, so choosing where simulation runs becomes a toolchain question as well as a cloud one.
- exposure Synopsys's IP income will rise with Amazon's chip volume, giving the toolmaker a direct financial stake in Trainium and Graviton shipping in quantity.
- constraint Cadence and Siemens are automating chip design too, and with no exclusivity reported, any design-side edge for Amazon depends on how quickly it puts the tools to use.
The agreement moves work in two directions. Amazon's half makes it a bigger Synopsys customer. It licenses Synopsys IP and widens its use of Synopsys electronic design automation (EDA) software for AI chip design and agentic AI [2]. Amazon will also give its own engineering teams Synopsys's AI-powered EDA, physics-based simulation and agentic tools [8]. The companies claim this will help them design, analyze, optimize and validate chips more efficiently [8].
Synopsys's half puts its software on Amazon's hardware. Synopsys will work with Amazon to optimize its multiphysics solutions for Trainium and Graviton [4]. Tom's Hardware argues this could make Amazon's hardware more attractive to customers running those engineering workloads [10]. That only holds for a given customer if their own simulation jobs run at least as well on the tuned build as on the hardware they use now, at a price that justifies moving. The report does not include performance figures for these ports, a breakdown of the reported $1 billion-plus value, or any exclusivity terms [1].
Tom's Hardware also makes a wider claim. As toolmakers automate more of chip design, it wrote, "Amazon ensuring that process is optimized for Amazon hardware places it in a much more favorable position in a world where chip design is easier and faster" [11]. The reported commitment is narrower than that. Synopsys tunes simulation software to run on Trainium and Graviton [4]. That helps AWS sell compute to people who design chips. The design-side benefit is that Amazon buys more Synopsys tooling and IP [2]. The demand for custom silicon comes from many large AI companies wanting their own inference hardware to cut what they spend on Nvidia GPUs [12]. In my view this deal changes where that design work runs more than it changes who can do it.
Tom's Hardware concedes "some element of cooperative back-scratching" in the arrangement [14]. Most supplier contracts have some. This one writes it into the payment terms, with a license-plus-royalty model on the IP that the report ties to production volume [6].
Amazon describes its side in terms of speed. "As our chip designs grow more ambitious and AI reshapes the engineering process itself, Synopsys helps us move faster across the design cycle, helping us deliver more capable, efficient computing for customers worldwide," said Amazon SVP Peter DeSantis in a joint press release [13].
What to watch
- Published performance numbers for Synopsys multiphysics on Graviton or Trainium instances against the hardware customers run those jobs on today.
- Any disclosure of exclusivity terms, or of how the reported $1 billion-plus splits between license fees and royalties.
- Whether Cadence or Siemens sign comparable tuning agreements with AWS or a rival cloud.