Felix Prehn of Goat Academy explains a quiet force behind modern semiconductors. It is not a chip maker. It is software. This software category is called EDA, which stands for electronic design automation. In simple terms, EDA is the brain that helps engineers design chips with billions of tiny parts. Without EDA, creating today’s AI chips would be almost impossible.
Think of building a Lego castle with 50 billion pieces. If one piece is wrong, the castle breaks. EDA software helps arrange every piece, checks if the design works, and confirms it can be manufactured. That is why every major chip project—from Nvidia to Google, Amazon, Apple, and others—relies on EDA tools.
The EDA market is about $15 billion today and is growing around 10% per year. Growth comes from AI chips, custom silicon at big tech companies, 5G, electric vehicles, and self‑driving systems. More complexity means more need for smarter design tools.
Only three companies dominate this field: Synopsys, Cadence, and Siemens EDA. This concentration gives them a “moat.” A moat is a durable advantage that protects a business. In EDA, the moat comes from huge switching costs. If a company is designing a $10 billion chip, changing core software midstream risks delays, retraining costs, and mistakes. The risk is too high.
Synopsys (ticker: SNPS) is the leader. Its business has three parts:
- EDA software (about 70% of revenue): tools that design and verify chips.
- IP licensing: building blocks like USB controllers, memory interfaces, and processor cores that customers can license instead of building from scratch. This saves years.
- A divested software security segment (in the process of being sold and no longer core).
Synopsys has three strengths:
- A data flywheel: It uses its own tools to design its IP. More IP creates more design data. That data trains AI‑powered tools. Better tools create better IP. The loop strengthens over time and is hard to copy.
- Deep partnerships with leading foundries: TSMC, Samsung, and Intel work closely with Synopsys to certify tools for the newest manufacturing processes (such as 3nm and 2nm). Certification takes years and creates trust and lock‑in.
- Recurring revenue: About 92% of revenue comes from multi‑year subscriptions, which are predictable and sticky. Gross margins near 75% and net margins near 30% show strong pricing power and high value to customers.
Synopsys also expanded its scope by acquiring a leader in physics simulation software used in autos, aerospace, and semiconductors. This move grows its total market and enables “full‑stack” digital engineering: design the chip, simulate it, design the surrounding system, and simulate the entire system together. That is attractive for both AI data centers and automotive firms.
Nvidia’s recent $2 billion investment in Synopsys underscores the importance of this stack. The collaboration moves key design flows from CPUs to GPUs, cutting design time from weeks to hours in some cases. It also layers Nvidia’s AI into Synopsys tools to automate more of the chip design process—like an “autopilot” for chip layouts and error checking. Digital twins—virtual replicas of chips and systems—let teams test data centers or vehicles before building anything. Putting these capabilities in the cloud will also broaden access, especially for startups.
Why would Nvidia do this? Faster internal design cycles, earlier access to cutting‑edge tools, and more designers choosing Nvidia‑optimized workflows. Software choices often drive hardware demand. If teams adopt GPU‑accelerated design tools, they are more likely to buy GPUs to run them.
From a financial view, Synopsys has grown revenue from about $4 billion to roughly $6 billion in three years, with strong margins. The stock is off recent highs due to China restrictions, the integration risk of a large acquisition, and general AI market swings. Still, analysts see potential upside as chip design demand rises and cloud delivery improves access and profitability.
Key risks include:
- Geopolitical limits on selling to certain regions.
- Large acquisition integration and execution risk.
- Market volatility in AI spending cycles.
- Technical chart risks if momentum weakens near resistance levels.
The takeaway: EDA is the invisible engine of AI hardware. Synopsys is the category leader with durable advantages, deep industry ties, and a growing platform that spans from chip design to full system simulation. Nvidia’s $2B stake validates its role at the center of AI infrastructure.
For more background on Felix Prehn and his work, see the Goat Academy overview here: About Goat Academy and Felix Prehn.