Synopsys Unveils $1 Billion Amazon Silicon Deal and OpenAI Partnership While Raising Long-Term Growth Targets to Mid-Teens
Investor Day, September 30, 2026: Company reveals new IP business model, custom GPT with OpenAI, and pulls forward FY2027 guidance by two and a half months
Synopsys used its first Investor Day since closing the Ansys acquisition to disclose two significant commercial developments that reshape how investors should think about the company's growth trajectory: a $1 billion multi-generation licensing agreement with Amazon covering custom silicon programs, and a new partnership with OpenAI to build a semiconductor-specific AI model. Together with an upward revision to long-term growth targets and an early release of fiscal 2027 guidance, the event amounted to one of the more consequential updates from the company in recent years.
Amazon Deal Validates New "Factory 2" IP Model
The most concrete new disclosure was a $1 billion license agreement with Amazon covering multiple generations of its Graviton, Trainium, and Nitro silicon programs. CEO Sassine Ghazi described this as proof of concept for a business model shift the company has been developing for roughly a year, splitting its IP unit into what it calls Factory 1 and Factory 2. Factory 1 is the traditional standards-based IP business, licensed once and sold broadly across many customers. Factory 2, formally called Application Optimized IP, involves Synopsys engineers embedding directly with a customer's system design team to build customized interface IP for specific workloads, monetized through a license fee, a customization fee, and a royalty once the chip ships in volume.
Ghazi was direct about the skepticism this model faced internally and externally. "I remember as well when we talked about it, many doubters, how will you ever change a business model that's been established for 3 decades. We don't see it. We don't get it. There is no way the customer will pay for this," he said, before noting the Amazon deal is accompanied by parallel agreements with ASIC and connectivity partners in the same ecosystem. Management is guiding to $1 billion in Application Optimized IP revenue by 2030 based on current contracts, with an explicit ambition for royalty revenue to eventually exceed license revenue, a dynamic CFO Shelagh Glaser flagged as effectively 100% margin revenue the company has never had in its model before. Notably, no royalty revenue is assumed in the fiscal 2027 guide, with Ghazi pointing to roughly 14 months between design start and production before meaningful royalty flow begins.
OpenAI Partnership Introduces a Third AI Monetization Lane
Synopsys also announced a multiyear partnership with OpenAI to build a specialized model, referred to internally as GPT-Synopsys, post-trained on the company's proprietary tools, agents, and workflows. OpenAI's Greg Brockman appeared to describe the arrangement: "Our progress in AI depends fundamentally on the chips that we run on, and now we have an opportunity to use AI to help design those chips... this is something we can build on itself. Better AI helps engineers design better chips. Better chips makes AI more capable, more efficient, more accessible to everyone."
Management framed this as the third of three customer AI adoption paths, alongside customers using the full Synopsys agentic stack (branded Autopilot) or building their own agent platforms while licensing specific Synopsys agents and tools. The OpenAI arrangement is structured as a service, with customers accessing the model, compute, and Synopsys tooling through OpenAI, monetized via a revenue share tied to design outcomes such as improved power, performance, and area. Ghazi was careful to note that Synopsys IP does not bleed into the broader foundation model, calling that protection "a nonstarter for Synopsys" otherwise. He also indicated early engagements already exist where OpenAI is acting as a chip buyer directing its suppliers' design choices using the model, a revenue stream Synopsys says it would not have previously captured.
Guidance Raised Across Every Segment, FY2027 Pulled Forward
Beyond the two marquee deals, Synopsys raised its long-term growth algorithm meaningfully. The company now guides to mid-teens overall revenue growth through 2030, up from a prior double-digit target, built on a 13% floor for EDA (versus a prior low-double-digit expectation), a 17% floor for design IP (up from mid-teens), and a 10% floor for simulation and analysis. Glaser also disclosed the company pulled forward its fiscal 2027 guidance by two and a half months specifically to share it at this event, projecting revenue of $11.15 billion, representing 15% growth, with non-GAAP operating margin expanding 250 basis points to 44% and EPS growing to $19.08.
Longer-dated, the operating margin target was raised to approximately 50% by fiscal 2030, up from a prior mid-40s target, driven largely by the economics of the new IP model and AI monetization streams layering onto the existing cost base. Glaser also confirmed that Ansys-related cost synergies, originally targeted at $400 million run-rate by fiscal 2028, will now be complete a year early in fiscal 2027, while revenue synergies remain on track for a $400 million run-rate by 2029, with more than $100 million already expected to be realized by the end of fiscal 2027. Stock-based compensation is targeted to fall to 8% of revenue by 2030, down from a peak of roughly 13% in 2025. The company separately announced its intent to repurchase $1 billion of shares in the coming months, funded by free cash flow management expects to reach approximately $3.1 billion in fiscal 2027, up roughly $500 million year-over-year.
Multiphysics Fusion and Hardware Roadmap Underpin the Core EDA Business
Addressing a sharper line of analyst questioning around whether classic core EDA software growth, which has run in the low-to-mid single digits recently, can plausibly support a mid-teens blended target, Ghazi pointed to the first wave of combined Synopsys-Ansys multiphysics products, which the company says has already attracted five customers in early deployment with benefits including tenfold faster photonic simulation and threefold higher fidelity 3D geometry modeling. Revenue synergy from this integration is expected to begin showing up in results in fiscal 2027. Ghazi also disclosed, without a formal product launch, a next-generation hardware-assisted verification system internally codenamed Artemis, expected in the first half of calendar 2027, succeeding the company's ZeBu emulation and prototyping line with greater capacity and performance.
On the structural debate around whether AI model capability eventually displaces EDA tools altogether, a question raised repeatedly by analysts given recent chip-design announcements using large language models, Ghazi was unambiguous: "I don't see how a model can bypass EDA generation, generating of data and signing off and checking the data." He summarized customer sentiment on the issue with a line he said reflects what Synopsys hears consistently from its base: "In AI, we believe, but in physics, we trust."
Segment Reporting Changes Coming in Fiscal 2027
Glaser also flagged a reporting change that investors will need to track closely starting next fiscal year. Synopsys will move the semiconductor business unit, approximately 10% of Ansys revenue through the third quarter of fiscal 2026, out of the standalone Ansys disclosure and into the EDA segment, with the remainder reported within simulation and analysis. Management said the change reflects how intertwined the multiphysics fusion products have become and promised like-for-like comparisons throughout fiscal 2027 to preserve transparency following a year in which the company provided full standalone Ansys disclosure as promised at the time of the deal close.
China and Competitive Dynamics
On a question about whether the Application Optimized IP opportunity extends into China, Ghazi was candid that demand has slowed for the most advanced IP tier specifically because Chinese customers lack access to gate-all-around process technology and advanced 3D IC packaging, limiting what can actually be designed there regardless of Synopsys's willingness to engage. He characterized this as a function of export restrictions rather than commercial strategy.