Synopsys and Amazon Web Services have signed a chip design licensing agreement worth more than $1 billion, one of the largest commercial commitments ever disclosed in the electronic design automation (EDA) sector.

The deal, reported by The Mighty 790 KFGO, gives AWS access to Synopsys's portfolio of chip design software and intellectual property as the cloud provider continues to expand its custom silicon programs. AWS designs its own processors in-house, including the Graviton server chips based on Arm architecture and the Trainium and Inferentia families used for machine learning workloads in its data centers.

The agreement signals how deeply the largest cloud operators now depend on EDA tooling and licensing to sustain their internal chip roadmaps. For Synopsys, the contract anchors a revenue stream from a hyperscale customer that has moved from buying merchant silicon to designing its own — a shift that has reshaped demand across the semiconductor supply chain since AWS shipped its first Graviton processor.

A licensing commitment of this size is unusual for the EDA industry, where tool contracts are typically structured as multi-year subscriptions but rarely disclosed with a headline value exceeding $1 billion. The figure underscores both the breadth of Synopsys's portfolio — spanning design, verification and IP — and the scale of AWS's silicon ambitions across its compute, networking and AI accelerator lines.

The deal also lands at a consequential moment for Synopsys. The company completed its acquisition of Ansys earlier this year, a roughly $35 billion transaction that expanded its portfolio into simulation software, and it continues to navigate tightening U.S. export controls on chip design tool sales into China. A billion-dollar licensing agreement with an American hyperscaler provides a stable, geopolitically uncomplicated revenue base as sales into restricted markets come under scrutiny.

For AWS, the arrangement secures long-term access to the design infrastructure needed to keep pace in custom silicon, where cloud rivals Microsoft and Google also build their own accelerators — Google with its Tensor Processing Units and Microsoft with its Maia chips. Custom processors have become a lever for controlling cost and power efficiency at data center scale, and each new generation demands sustained EDA and IP investment.

Neither company has publicly broken down the contract's term length, which product lines it covers, or how the value splits between software licensing and IP. The disclosed figure represents the total commitment of the agreement as reported.

The contract is expected to support AWS's next generations of custom compute and AI silicon, and it reinforces the position of the top EDA vendors — Synopsys, Cadence and Siemens EDA — as suppliers whose licensing terms now sit alongside wafer capacity and advanced packaging as gating factors in the custom chip race.