Synopsys and Amazon announced a multi-year agreement worth more than $1 billion on September 30, 2026, putting Amazon at the center of the chip-design giant’s push into a new business model built on silicon intellectual property and agentic AI. The deal, confirmed by Synopsys and covered by Converge Digest, Pulse2 and other outlets, makes Amazon the lead customer for a new category of Synopsys products called application-optimized silicon IP — purpose-built design blocks tailored to specific workloads rather than the generic, reusable interface IP Synopsys has historically sold.
The announcement lands at a moment when every major cloud provider is racing to control more of its own silicon supply chain. Amazon has spent over a decade building custom chips for AWS, including the Graviton CPU line, Trainium AI accelerators and Nitro infrastructure silicon. This agreement expands how much of that design pipeline runs through Synopsys tools, and it reshapes how Synopsys gets paid for the IP underneath it. Shares of Synopsys (NASDAQ: SNPS) rose more than 3% in trading after the news broke, according to contemporaneous market coverage.
Don't miss new tech stories on Google
Add Tech Insider once in the Google app and our stories appear in your news suggestions.
What the Synopsys-Amazon Agreement Actually Covers
According to the companies’ joint announcement, the agreement spans four areas: application-optimized silicon IP, electronic design automation (EDA) software, simulation and analysis tools, and agentic AI technologies for engineering workflows. Engineering.com, reporting in October 2026, framed the software side of the deal as combining three AI-enabled engineering capabilities — EDA, physics simulation and agentic AI — rather than a single licensing line. Rather than licensing a single chip design or one-off IP block, Amazon is folding Synopsys technology into its broader custom-silicon pipeline across multiple chip families at once.
Reports describe the IP as reusable, workload-specific design blocks — building blocks tuned for AI training, inference, networking or security tasks rather than generic components meant to serve any chip. That distinction matters because it signals Synopsys is moving up the value chain, selling pre-optimized subsystems instead of raw interface IP that customers have to tune themselves. The companies also said Synopsys will expand its own use of AWS infrastructure — specifically Amazon EC2 and cloud storage, per Reuters‘ September 2026 report on the agreement — and will adopt Amazon Bedrock starting in September 2026 to build and deploy its own AI applications and engineering agents, a reciprocal cloud commitment layered on top of the IP licensing.
Neither company has disclosed a line-item breakdown separating IP licensing revenue, EDA software fees, royalty payments and cloud consumption. The $1 billion-plus figure is the only confirmed headline number tied to the agreement, and outlets including Converge Digest and Pulse2 have been careful to note that it represents the deal’s total disclosed value, not a guaranteed annual run rate.
Amazon Becomes Synopsys’ Lead Application-Optimized IP Customer
Synopsys has sold IP for decades, but mostly in the form of standards-based building blocks — interfaces, memory controllers and interconnects meant to be dropped into almost any chip design. Application-optimized IP is a different product category: IP shaped around a specific workload and a specific customer’s system architecture. Amazon is now the flagship customer for that new line of business, which gives Synopsys a high-profile proof point as it tries to convince other hyperscalers and chipmakers to adopt the same model.
That positioning matters because AWS is not a typical IP licensee. It is one of the largest in-house chip design operations in the world, competing directly with the merchant silicon that companies like Nvidia sell. By becoming Synopsys’ lead customer for a new IP tier, Amazon gets a say in how that product category evolves — and Synopsys gets validation that its application-optimized IP can handle the scale and complexity of hyperscaler workloads, not just smaller custom-chip programs.
Inside the New License-Plus-Royalty Business Model
The most structurally significant part of the deal is the shift to what Synopsys describes as a license-plus-royalty commercial model. Historically, EDA and IP vendors like Synopsys and Cadence Design Systems have earned revenue primarily through upfront licensing and subscription fees, largely disconnected from how many chips a customer eventually ships. The new structure ties part of Synopsys’ revenue to production volume — meaning Synopsys earns more as Amazon manufactures and deploys more chips built with its IP.
This is a meaningful bet on AWS’s custom-silicon trajectory. If Amazon keeps scaling Trainium and Graviton deployments across its data centers, Synopsys participates in that growth directly rather than collecting a flat license fee regardless of volume. It also gives Synopsys exposure to the broader AI infrastructure buildout without having to manufacture or sell chips itself — a lower-risk way to ride the same wave that has pushed Nvidia, AMD and Broadcom’s data-center revenue sharply higher this year, as detailed in AMD’s own 107% jump in data-center revenue.
The tradeoff is timing. Royalty revenue depends on Amazon’s chip shipment volumes, which neither company has disclosed, and on production timelines that typically run years behind initial IP licensing. Investors reading the 3% stock pop as an immediate earnings catalyst should note that the companies themselves have not provided a revenue-recognition schedule or royalty rate.
Why AWS Needs Synopsys Now: Graviton, Trainium and Nitro
AWS’s custom-silicon strategy rests on three named chip families that Synopsys itself cites as covered by the September 2026 agreement: Graviton, Trainium and Nitro. Graviton, Arm-based processors, handle general-purpose cloud compute. Trainium targets AI model training and inference, and Synopsys has specifically flagged Trainium and Graviton as the two processor platforms slated for multiphysics optimization under the new deal. Nitro offloads security, networking and storage functions from host servers, and Amazon’s broader chip portfolio referenced in the agreement also extends to a fourth category — dedicated storage and networking silicon — alongside those three headline families. Each of those product lines requires the same underlying engineering: architecture definition, RTL design, verification, physical implementation, timing and power analysis, thermal modeling, and system-level validation — all areas where Synopsys and Ansys (which Synopsys acquired in 2025) sell tools.
Designing chips at this scale and pace is not something AWS can fully internalize. Even hyperscalers with large silicon teams lean on EDA vendors for the tooling that turns an architecture into a manufacturable chip. By deepening its Synopsys relationship instead of building that infrastructure from scratch, AWS can iterate faster on new Trainium and Graviton generations while directing its own engineering headcount toward chip architecture rather than tool development.
The strategic subtext is also about reducing dependence on general-purpose accelerators. AWS has already been buying Nvidia GPUs at enormous scale — the company placed an order for 2 million additional Nvidia GPUs earlier this year — while simultaneously investing in its own silicon to offer workload-specific alternatives. A deeper Synopsys relationship supports both tracks: AWS keeps buying merchant silicon where it makes sense, while building out proprietary chips where custom design delivers better performance per dollar for its own cloud services.
What Synopsys and Amazon Executives Are Saying
Sassine Ghazi, president and CEO of Synopsys, framed the deal as validation of the company’s shift toward application-specific IP. “As we expand into application-optimized IP solutions, we’re proud to have Amazon as our lead customer for this next phase of growth,” Ghazi said, according to the companies’ joint statement reported by StockTitan. He added that combining Synopsys and Amazon’s engineering capabilities could “dramatically increase the productivity of our teams.”
Peter DeSantis, Amazon’s senior vice president for foundational AI, custom silicon and quantum computing, tied the agreement back to AWS’s long-running silicon strategy. “Amazon has invested in custom silicon for more than a decade, and it has fundamentally changed what’s possible for AWS and our customers,” DeSantis said. “From Graviton to Trainium, purpose-built chips deliver better performance at lower cost because they’re designed for exactly what customers need.”
Neither executive disclosed specific revenue targets, royalty rates or which upcoming Trainium or Graviton generations will carry the new application-optimized IP first. The public statements focus on strategic direction rather than financial detail — consistent with how both companies have historically communicated custom-silicon partnerships.
Synopsys Stock Reaction: SNPS Climbs on Deal News
Synopsys shares traded up more than 3% on September 30, the day of the announcement, according to market coverage of the news. As of October 1, 2026, SNPS trades around $440 a share with a market capitalization near $83 billion, according to stock-data tracker StockAnalysis.com. That valuation reflects a company still absorbing its largest-ever acquisition: the roughly $35 billion purchase of simulation software maker Ansys, which closed in July 2025.
The market reaction appears tied to two factors investors flagged in the days after the announcement: the size and multi-year nature of the Amazon commitment, and the royalty structure’s potential to compound as AWS chip volumes grow. Reports have not identified a specific analyst price-target revision tied directly to the Amazon news, so that reaction should be read as an initial trading response rather than a confirmed change in Wall Street’s long-term SNPS outlook.
The Ansys Acquisition Backdrop: How a $35 Billion Deal Set the Stage
The Amazon agreement does not exist in isolation. Synopsys closed its acquisition of Ansys in July 2025, combining its EDA and chip-IP business with Ansys’s multiphysics simulation software — tools used to model how chips behave thermally, mechanically and electromagnetically once they leave the design phase and enter real systems. That combination directly supports the kind of “silicon-to-system” engineering the Amazon deal references: designing a chip is only part of the job when that chip has to function inside a data-center rack running at full AI training load.
Synopsys’ revenue has already reflected the Ansys integration. The company’s fiscal second-quarter 2026 revenue rose 42% year-over-year to $2.28 billion, a jump driven largely by the newly combined Ansys business rather than organic EDA growth alone. The Amazon agreement gives Synopsys a marquee customer to demonstrate that the Ansys integration produces real engineering value, not just combined financial statements.
Synopsys vs Cadence: Battling for the AI Chip Design Market
Synopsys does not operate in a vacuum. Cadence Design Systems remains its closest rival, and the two companies together control the overwhelming majority of the global EDA market — estimates place their combined share anywhere from roughly 74% to 85%, depending on how the market is segmented. Cadence has pursued its own simulation expansion, agreeing to acquire multiphysics software maker Hexagon’s manufacturing intelligence division in a deal reported at roughly $3.16 billion, a clear echo of Synopsys’s Ansys strategy.
The Amazon deal strengthens Synopsys’s hand in a market where both vendors are chasing the same hyperscaler customers for AI-chip design work. It does not make Synopsys AWS’s exclusive EDA or IP supplier — the public announcement does not claim exclusivity, and AWS has historically worked with multiple design-tool vendors across its chip programs. But landing Amazon as the named lead customer for a brand-new IP category is a public relations and competitive win Cadence will have to answer, likely with its own hyperscaler-focused announcements in the coming quarters.
15 Years in the Making: The Synopsys-Amazon Relationship
Synopsys and Amazon describe their relationship as stretching back more than 15 years, predating AWS’s current custom-silicon ambitions entirely — a timeline Yahoo Finance reiterated in its October 2026 coverage of the agreement. That history matters for context: this is not a new vendor relationship being forged under AI-boom pressure, but an existing partnership being substantially expanded in scope. Amazon has used EDA and IP tools from multiple vendors across its chip programs since it began custom-silicon development over a decade ago, and Synopsys tools have reportedly been part of that mix for most of it.
What changed is the depth of the commitment and the introduction of a volume-linked royalty structure — a sign that both companies expect AWS’s custom-chip shipments to keep climbing for years, not just through the current AI infrastructure cycle.
Market Impact: Custom Silicon Spending Keeps Accelerating
The Synopsys-Amazon agreement is the latest in a string of billion-dollar-plus custom-silicon commitments across the cloud and chip industry in 2026. Qualcomm and AWS signed their own multibillion-dollar AI chip arrangement earlier this year, and competing cloud providers have pursued parallel paths with their own custom CPU lines, as seen in the ongoing race between AWS Graviton5, Azure Cobalt 200 and Google Axion. The common thread across all of these deals is the same: hyperscalers want tighter control over the silicon running their AI workloads, and EDA vendors like Synopsys are positioning themselves as indispensable partners rather than simple tool suppliers.
For Synopsys specifically, the deal reinforces a broader pattern: traditional EDA licensing revenue is increasingly supplemented by IP royalties and simulation-software fees tied directly to AI infrastructure spending. That diversification could make Synopsys’s revenue less cyclical and more correlated with the long-term AI buildout than with the historically boom-bust semiconductor design cycle.
Competitive Landscape: Where Nvidia, Broadcom and Marvell Stand
The deal also sits inside a larger shift the industry has been tracking all year: AWS’s biggest GPU supplier, Nvidia, is increasingly competing with its own largest customers as they build custom silicon in-house. Broadcom and Marvell, which both design custom AI accelerators for hyperscalers, occupy a similar space to Synopsys in the sense that all three companies profit from the same trend — cloud providers wanting workload-specific chips instead of off-the-shelf GPUs — even though their roles in the supply chain differ. Synopsys and Cadence sell the tools and IP that make custom chips possible; Broadcom and Marvell often co-design and help manufacture the finished accelerators.
That dynamic has already reshaped how Nvidia talks about its own business, acknowledging that its biggest cloud customers are becoming silicon rivals even as they continue buying Nvidia GPUs at record volumes. The Synopsys-Amazon deal is a reminder that this shift depends on an entire ecosystem of design tools, IP licensing and simulation software that rarely gets the same attention as the chips themselves, despite being just as essential to the custom-silicon buildout chronicled in reporting on Nvidia’s growing list of custom-chip rivals.
Predictions: What Comes Next for Synopsys and AWS
Based on the scope of the agreement and the broader industry trajectory, several outcomes look likely in the coming quarters:
- Synopsys will likely cite the Amazon deal as a template when courting other hyperscalers for its application-optimized IP line, using Amazon’s scale as proof the model works at production volume.
- Cadence Design Systems will probably respond with its own hyperscaler-focused IP or simulation announcement within the next two to three quarters, continuing the pattern set by its Hexagon acquisition.
- Future Synopsys earnings calls will likely include more detail on royalty-based revenue recognition as the Ansys integration matures and the Amazon agreement moves from signing to active engineering work.
- AWS will likely expand the scope of chips covered by the deal beyond Graviton, Trainium and Nitro as it develops new silicon for networking, storage or specialized AI inference tasks.
- Expect continued scrutiny of exactly how much of Synopsys’s reported revenue growth comes from organic EDA demand versus Ansys-driven simulation revenue, a distinction investors and analysts are likely to press for in coming quarters.
Synopsys-Amazon Deal at a Glance
| Detail | Figure / Description |
|---|---|
| Announcement date | September 30, 2026 |
| Deal value | More than $1 billion (multi-year) |
| Lead customer status | Amazon named lead customer for application-optimized silicon IP |
| Business model | Shift to license-plus-royalty structure |
| Technologies covered | Silicon IP, EDA software, simulation/analysis tools, agentic AI |
| AWS chip families referenced | Graviton (general compute), Trainium (AI training/inference), Nitro (security/networking/storage) |
| Reciprocal cloud commitment | Synopsys to expand use of Amazon EC2, cloud storage and Amazon Bedrock |
| Prior relationship length | More than 15 years |
| SNPS stock reaction | Up more than 3% on announcement day |
| SNPS market cap (Oct. 1, 2026) | Approximately $83 billion |
EDA Duopoly Comparison: Synopsys vs Cadence in 2026
| Metric | Synopsys (SNPS) | Cadence (CDNS) |
|---|---|---|
| Approx. EDA market share | 31-38% (estimates vary by source) | 30-36% (estimates vary by source) |
| Major simulation acquisition | Ansys, ~$35 billion, closed July 2025 | Hexagon manufacturing intelligence unit, ~$3.16 billion |
| Recent quarterly revenue growth | 42% YoY in fiscal Q2 2026 (Ansys-driven) | Not covered in this report’s sources |
| Strength areas | Digital EDA, application-optimized silicon IP, multiphysics | Analog chip design, system simulation |
| Hyperscaler silicon IP play | Amazon named lead customer, Sept. 2026 | No equivalent announcement reported as of Oct. 1, 2026 |
Together, Synopsys and Cadence control the large majority of the global EDA software market, with estimates of their combined share ranging from roughly 74% to 85% depending on how analysts segment the industry. The duopoly structure means any major customer win — like Amazon’s new IP relationship with Synopsys — carries outsized competitive weight, since there are effectively only two vendors capable of serving chip designs at hyperscaler scale.
Frequently Asked Questions
What exactly did Synopsys and Amazon agree to?
A multi-year agreement worth more than $1 billion covering Synopsys silicon IP, EDA software, simulation and analysis tools, and agentic AI technologies, with Amazon named as the lead customer for Synopsys’s new application-optimized IP business.
When was the deal announced?
September 30, 2026, with coverage continuing into October 1 as more outlets picked up the story.
What is a license-plus-royalty model?
It is a commercial structure where Synopsys earns revenue both from upfront licensing fees and from ongoing royalties tied to how many chips Amazon actually produces using the licensed IP, giving Synopsys exposure to AWS’s production volume growth.
Which AWS chips does this deal involve?
The companies’ announcement references Graviton (general-purpose compute), Trainium (AI training and inference) and Nitro (security, networking and storage offload), though the exact IP blocks used in each chip family have not been publicly detailed.
How did Synopsys stock react?
SNPS shares rose more than 3% on the day of the announcement, according to market coverage, and the stock traded around $440 with an approximate $83 billion market cap as of October 1, 2026.
Does this make Synopsys AWS’s exclusive chip design partner?
No. The announcement does not claim exclusivity, and Cadence Design Systems, Synopsys’s chief EDA rival, continues to compete for the same hyperscaler customers.
How does this connect to Synopsys’s Ansys acquisition?
Synopsys closed its roughly $35 billion acquisition of simulation software maker Ansys in July 2025. The combined multiphysics and EDA capabilities directly support the “silicon-to-system” engineering scope described in the Amazon agreement.
Is this Amazon’s first custom-silicon IP deal of this size?
No. AWS has pursued multiple large custom-silicon and chip-design partnerships in 2026, including a multibillion-dollar AI chip arrangement with Qualcomm, as part of its broader strategy to reduce dependence on merchant GPU supply.
