The EDA Industry: An Excellent Through-the-Cycle Investable Niche

The insurance industry is cursed with a set of dismal economic characteristics that make for a poor long-term outlook … only a very low-cost operator or someone operating in a protected, and usually small, niche can sustain high profitability levels.
— Warren Buffett

Please refer to Important Disclosure Information at the end of this research note.

Sometimes the best investment opportunities lie in high quality niche businesses residing within more commoditized or cyclical industries.  Such an opportunity has existed for at least the last five or so years in the Electronic Design Automation (EDA) industry. 

For example, over the last five years, Cadence (CDNS) shares have appreciated 186% compared to the technology-heavy Nasdaq which is up 117%, before dividends.  While Apple (AAPL) shares have done better up 210%, CDNS shares have been far less volatile.

Here is one perspective on this remarkable and overlooked industry, in which I happened to work for more than 7 years before pursuing a career in finance.

EDA companies

EDA companies like Cadence, Synopsys, and Mentor Graphics, sell software that enable companies such as Intel, Samsung, and Texas Instruments to design and manufacture complex computer chips.

To better understand how critical EDA software is, it is worth noting that hardware chip design companies (like Intel, Samsung, and even Apple) live and die by two key challenges:

  1. Time-to-market: In an industry where the shelf-life is measured in months, any product delay could translate into an inescapable death spiral for a competitor.  Remember 3dfx, the innovative graphics chip company that lost out to ATI and NVIDIA? Having the right set of software tools to effectively and reliably design complex chips is a “must-have” not a “nice-to-have.”
  2. “Bad logic” bugs:  If there is a bug in the latest version of iOS software, all Apple has to do is broadcast a software patch to all affected devices. On the other hand, if there is a fatal (“bad logic”) bug in the A9X processor chip, Apple may have no option but to do a costly and brand-impairing recall of all devices.  This illustrates the difference between the consequences of poorly designed software and poorly designed hardware.  EDA tools enable hardware designers to identify and eliminate hardware bugs.

Some Historical Context

The EDA industry is about 30 years old and was effectively invented by Cadence, Synopsys, and Mentor.  Before EDA tools, chip designers “drew” a chip on a white board as one would sketch a painting.  As chip complexity increased, this manual design process was unscalable. 

Synopsys invented "logic synthesis", born out of research done by its CEO Aart De Geus’ PhD work – i.e., a set of software algorithms that allowed chip designers to use a software language (C, C++, or Java) to abstract out the chip design.  Synopsys’ tools would then apply advanced algorithms (with fancy names like “genetic programming”, “dynamic programming”, “simulated annealing”, “conjugate gradient”, “binary decision diagrams”, etc.) to convert the software program into chips optimized for speed, area, and power.

The EDA industry is protected by high IP content which has only gotten stronger with time. 

Back when I was at Synopsys, I worked with two of the smartest people I’ve ever known.  One was a brilliant engineer/scientist out of Stanford and now works in a senior role at Cadence.  The other owned the fastest SAT solver algorithm in the world.

Industry Evolution

The industry has evolved over the years, mostly for the better.

EDA used to be a best-of-breed “point” tools business stitched together by the customer.  Synopsys had the best “synthesis” tools and Cadence dominated “place & route.” When Synopsys met with a prospective client, they had an incentive to lead with the “synthesis” product and give away their “place & route” tool for free. Cadence did the opposite, resulting in irrational pricing behavior to the detriment of these companies.  This era, right up to around 2004, was a fertile ground for startups with new algorithms developed in academia.

Things began to change when customers began to value “flows” more than “best-of-breed” tools.  The reason for this shift was the relentless force of Moore’s Law.  Post 45 nanometers (nm), transistors became very "leaky" (i.e., chips got very hot) which led to high complexity (retooling, clock gating, etc.) which required EDA tools to be tightly integrated so that deep visibility into final chips was possible.  This trend led to consolidation in the industry, as the bigger players with complete product suites edged out the smaller players.

With this transition to product portfolios now complete, the top-3 companies - Synopsys, Cadence, and Mentor - have an almost unassailable advantage over new entrants.

EDA Growth drivers

Growth in this industry is not spectacular with semiconductor R&D spend growing mid-single digits on a normalized basis.  Step function growth rates rarely occur and only when there is a major upgrade cycle due to a process node transitions.

That said, the business of selling software to the R&D department (i.e., the “DNA” of a company), is highly stable and the relationships highly sticky. 

EDA Business model

Prior to 2004, EDA companies used a perpetual license scheme.  i.e., customers paid upfront for a license to buy the software, leading to highly lumpy bookings. 

To make matters worse, the EDA participants were aggressive on pricing, sometimes offering 40-60% discounts, in order to close a deal and meet quarterly numbers.

Then in July 2004, Synopsys (SNPS) changed its model from perpetual to a 90/10 ratable model where instead of the customer buying the software they rented the software on a 3-year contract with cash collections every quarter (say $30m per quarter for 12 quarters).  Readers would recognize this as the classic Software-as-a-Service (SaaS) model with the associated predictability.  Also the lower spend effectively de-risked the purchase decision for the customer leading to a “win-win” outcome.

The most immediate impact of the model transition was a significant drop in revenues, margins, and free cash flow (FCF) and it took 3 years for the model transition to complete.  SNPS completed the transition in July 2007. 

Even though this was a low-uncertainty completely mechanical transition, it was accompanied with a large drop in the stock price.

SNPS:  stock price showing the rapid fall off in 2004 with the model change

SNPS:  stock price showing the rapid fall off in 2004 with the model change

CDNS initiated a similar transition under Lip-Bu Tan in 2008, with similar consequences. 

Since then, market participants have become much more comfortable with these model transitions as can be seen by the muted stock reactions for Adobe (ADBE), Autodesk (ADSK), and Microsoft (MSFT).

CDNS:  stock price showing a similar fall off during the model change

CDNS:  stock price showing a similar fall off during the model change

With the business model transition effectively complete, we are left with highly defensible businesses with 90% gross margins and 28-30% operating margins, capital-light, and tremendous operating leverage and free cash flow.

It is worth noting, that Cadence (CDNS) just announced a $1.2 billion share repurchase program, or 20% of shares outstanding at current prices.

Final Note:  Rational pricing

While pricing in this oligopoly has become more rational as a result of consolidation and the shift to the subscription-based revenue model, these businesses haven’t even begun to tap the pricing power that they command over their captive customers.  If they transition to an inflation-plus pricing we could truly see a "lollapalooza" effect in this attractive niche within the semiconductor industry.

 

Important Disclosure Information:

At the time this report was submitted for publication, the principals and clients of Burr Capital LLC owned securities issued by Cadence (Ticker: CDNS). All stock price charts were created in Google Finance.

Please remember that past performance may not be indicative of future results.  Different types of investments involve varying degrees of risk, and there can be no assurance that the future performance of any specific investment, investment strategy, or product (including the investments and/or investment strategies recommended or undertaken by Burr Capital LLC), or any non-investment related content, made reference to directly or indirectly in this research will be profitable, equal any corresponding indicated historical performance level(s), be suitable for your portfolio or individual situation, or prove successful.  Due to various factors, including changing market conditions and/or applicable laws, the content may no longer be reflective of current opinions or positions.  Moreover, you should not assume that any discussion or information contained in this research serves as the receipt of, or as a substitute for, personalized investment advice from Burr Capital LLC.  To the extent that a reader has any questions regarding the applicability of any specific issue discussed above to his/her individual situation, he/she is encouraged to consult with the professional advisor of his/her choosing.  Burr Capital LLC is neither a law firm nor a certified public accounting firm and no portion of the research content should be construed as legal or accounting advice.  A copy of Burr Capital LLC’s current written disclosure statement discussing our advisory services and fees is available for review upon request.