Cadence Design Systems and the price of a durable moat
What the business is
Cadence sells picks and shovels for a gold rush in which the miners must redesign their shovels every few years.
The company provides electronic design automation software, design IP, and hardware for emulation and prototyping. These tools help customers design and verify semiconductors and broader electronic systems. Annual subscription licences form the economic core, supplemented by hardware, individual IP licences, and services.
FY2025 revenue was $5.30 billion, comprising $4.82 billion of product and maintenance revenue and $0.48 billion of services. The attraction is not merely growth but continuity: once an engineering team has built its workflow around a particular tool chain, changing vendors during a complex chip project introduces cost, delay, and error risk. Cadence also benefits from decades of accumulated software and IP, certification with foundry processes, and an industry structure dominated by Cadence and Synopsys, with Siemens EDA as the other major participant.
The moat remains substantial, though some of its familiar shorthand needs tidying. Cadence once described approximately 85% of annual revenue as recurring. The more recent figures are 79% on a trailing basis and approximately 80% in FY2026 guidance. That decline reflects faster growth in up-front hardware and system-design software rather than demonstrated customer departures. Cadence does not publish a renewal-rate figure, so I would not treat the often-repeated near-perfect renewal claim as audited scripture.
More complex chips, chiplets, 3D-IC designs, and custom silicon create scope for greater tool usage. AI-assisted products such as Cerebrus may reinforce the incumbents’ data and workflow advantages, but that is evidence of a product direction, not yet proof that the moat has widened. The Hexagon D&E acquisition similarly expands Cadence into adjacent simulation and system-design work. The transaction has occurred; successful integration and stronger competitive economics remain to be demonstrated.
What the numbers say
FY2025 owner earnings were approximately $1.195 billion. The internal thesis also considers normalized owner earnings of $1.5–1.7 billion after allowing for the $140.6 million one-time penalty and FY2026 growth guidance. Applying a disciplined multiple to that normalized range produced an intrinsic-value estimate of $22.5–25.5 billion.
Against that earning power, today’s valuation arithmetic is severe:
- Market cap (price-adjusted, 2026-10-09 close): $97.32 billion
- Owner-earnings yield (price-adjusted): 1.23%
- Share price (2026-10-09 close): $352.84
- 10-year Treasury: 5.24%
The 1.23% owner-earnings yield sits 4.01 percentage points below the 5.24% Treasury. Meanwhile, the price-adjusted market cap stands at a 333% premium to the low end of the intrinsic range and a 282% premium to its high end. A fine business can survive a great deal; a fine valuation cannot survive arithmetic.
Capital allocation is mixed. Cadence repurchased approximately $928 million of shares in FY2025 and pays no dividend. Those repurchases have largely countered stock-based compensation, which has run at 4–6% of revenue, rather than meaningfully shrinking the share count. The Hexagon transaction also issued 3.2 million shares, while 69.24% of its $3.101 billion consideration was assigned to goodwill. That does not prove the acquisition was poor, but it puts a sizeable burden on integration results.
There is also a governance blemish. Cadence reached a $140.6 million settlement with the DOJ and BIS over exports made from 2015 through 2021 and pleaded guilty to one conspiracy count. No recurrence was identified in the subsequent reviewed filings, but another violation would materially alter my assessment of management.
What Wall Street believes vs what I see
The sampled headlines offer the usual three-course menu: a “buy the dip” technical case, an argument that broader growth can drive another earnings leg, and a warning that the share price is years ahead. The mood is therefore constructive about the business, with at least some recognition that valuation has become the awkward guest at dinner.
I see ample support for the constructive business view. Reported backlog moved from $7.8 billion to $8.0 billion and then $8.1 billion, while demand for verification hardware and adjacent system design has broadened the revenue base. None of that, however, makes a 1.23% owner-earnings yield competitive with a 5.24% Treasury. The market is not merely charging for durability; it is charging in advance for a long stretch of successful growth and integration.
My monitoring is specific. If backlog declines for two consecutive quarters from the latest $8.1 billion level, I would revisit pricing power and the earning-power assumptions. If trailing recurring revenue falls further from 79%, I would examine whether the cause remains product mix or has become customer erosion. Material deterioration from the FY2026 non-GAAP operating-margin guidance of 43.75–44.75% would prompt a fresh review of the Hexagon capital-allocation decision. A renewed export-compliance violation would force a reassessment of management quality rather than merely the valuation.
Verdict
Verdict: pass.
The 1.23% owner-earnings yield is below the 5.24% Treasury, so the shares fail both the watch threshold and the more demanding circle test. The $97.32 billion market cap also carries a 333% premium to the $22.5 billion low end of the intrinsic range, leaving no discernible margin of safety. I would change the verdict if price and owner earnings moved enough to put the yield above the Treasury, with a circle requiring both the additional yield spread and the prescribed discount to intrinsic value.
This analysis is AI-generated, educational, and not investment advice. Figures may contain errors or be delayed. Disclaimer