Synopsys and Cadence lose $12 billion as AI fears rattle EDA sector

Synopsys Inc. and Cadence Design Systems Inc. lost a combined $12 billion in market value on July 25 as investors worried that generative AI tools could upend the electronic design automation (EDA) software that dominates chip development.

"The market is pricing in a disruption scenario that assumes AI can replace decades of accumulated design know-how embedded in EDA tools," said Rachel Kim, semiconductor analyst at Edgen. "That's a much harder problem than training a large language model."

Synopsys shares fell 8.2% while Cadence dropped 7.6%, their steepest single-day declines in more than 12 months. The sell-off erased roughly $7 billion from Synopsys's market capitalization and $5 billion from Cadence's, according to exchange data. Trading volume for both stocks exceeded their 30-day averages by more than 2x.

The two companies control more than 70% of the EDA market, a duopoly that has proven resilient through multiple technology shifts. Their tools are deeply embedded in the chip design workflows of every major semiconductor company — from Nvidia Corp. to Advanced Micro Devices Inc. to Intel Corp. — creating switching costs that analysts estimate at 12 to 18 months of engineering time per design team.

The fear driving the sell-off centers on whether large language models could eventually automate portions of chip design, reducing the need for Synopsys's and Cadence's per-seat software licenses. EDA tools generate roughly $16 billion in annual revenue industry-wide, with Synopsys and Cadence capturing the majority. Any disruption to that model would threaten a highly profitable recurring revenue stream that has produced gross margins above 75% for both companies over the past five fiscal years.

But the comparison overlooks a fundamental difference between EDA and other software categories. Chip design involves verifying billions of transistors against physical constraints — power, heat, timing — where a single error can cost $50 million or more in mask costs and months of delay. EDA tools have been incorporating machine learning for years. Synopsys's DSO.ai and Cadence's Cerebrus already use AI to optimize floor planning and timing closure, two of the most compute-intensive steps in the design flow. The question is whether generative AI can leapfrog these specialized tools, not whether AI is new to the field.

"The EDA incumbents have been integrating AI into their products since 2020," Kim said. "They're not being disrupted by AI — they're the ones selling the AI tools that chip designers use."

Synopsys trades at roughly 32x forward earnings, while Cadence trades at 34x, both below their five-year averages of 38x and 40x, respectively. The sell-off has pushed both stocks into territory that some analysts view as a buying opportunity for long-term investors who believe the EDA duopoly will hold. Nvidia, the largest customer of EDA tools and the biggest beneficiary of AI infrastructure spending, saw its shares fall 1.3% on the same day, suggesting the sell-off was concentrated in the EDA names rather than a broader semiconductor rout.

This article is for informational purposes only and does not constitute investment advice.