[ADI – Analog Devices] A Silent Hero

Almost every electronic device you use on a daily basis is infested with analog chips that monitor, amplify, and transform real world phenomena (weight, light, temperature, amplitude, power), into digital signals that electrical systems can understand.  These chips are the silent and unappreciated heroes that enable safety and infotainment features in cars, wireless infrastructure equipment, consumer devices, and a broad array of industrial equipment and medical devices.  They can be found in systems used for airbags, factory machine monitoring, dialysis, X-rays, and battery cell testing, among other applications.

Relative to digital chips [microprocessors and memory chips that process on/off signals] and excluding the consumer market, analog chips are less subject to price competition, relatively insulated from the punctuated boom/bust cycles for which the broader industry is notorious, and sold to a far wider and fragmented set of customers whose products have significantly longer life cycles.  A well-designed general purpose part can last for generations, with some parts introduced in the 1970s still selling in high volume today; a third of ADI’s revenue comes from products that were released over 10 years ago.

Given the long-lived nature of analog product, the equipment used to manufacture these chips is less vulnerable to obsolescence.  Compared to digital chip process technology – which optimizes chip speed and minimizes size through ever-increasing circuit density – the analog integrated circuit manufacturing process, which is more focused on the precise placement of IC elements, has far less onerous capital requirements…ADI’s and LLTC’s capex as a percent of revenue has averaged ~5% over the last decade vs. 16% for INTC [and vs. the 15%-20% of revenue that ADI and LLTC consistently spend on R&D].

Analog integrated circuit design is a multi-variable optimization where various non-linear performance trade-offs have to be balanced during each design phase.  Digital chips, with their clean, logic-based underpinnings, can generally be modeled reliably and predictably with simulation tools.  The analog design process, on the other hand, is comparatively much more an art requiring creativity and experience to manage the uncertain outcomes that result from imperfectly and unevenly capturing properties of the physical world.  For instance, applying stress on a chip – like the kind that comes from packaging or from real world phenomena like temperature and humidity that the chip is meant to capture – can broaden circuit parameters in unexpected ways and de-bugging circuits is still a hands-on, experimental exercise that some circuit engineers describe as “black magic”.

As Bob Dobkin, CTO and legendary co-founder of Linear Technology [founded by Bob Dobkin in 1981 after his defection from National Semiconductor and well-known for its strong engineering talent] once described analog design as follows:

“I look at doing analog design like learning another language. When you start out learning a language, you start working with the dictionary and a set of rules or grammar. And as you’re learning the language and you’re, say, translating a piece of information from one language to another, you keep looking back at the dictionary to learn what things do. You should look at the rules of grammar to figure out how it’s put together. “

Given the steep learning curve and lengthy development cycle of many products, having a corporate culture that challenges, appreciates, and retains designers is more than just feckless lip service.  At both Linear and Analog, under a kind of apprenticeship, a fresh PhD is assigned a senior engineer who reviews and critiques his designs; only after around 5-10 years of experience can he start intuiting output from varying inputs and competently designing his own circuits.  The average engineer at Linear Technology has more than a decade of experience at the company.  Ultimately, because technology and performance win the socket and whoever occupies the socket can do so near-indefinitely, an experienced engineering corp can pay dividends for decades.

The upshot of long product cycles, low capital requirements, sticky and diversified customers, loyal workforces, strong engineering cultures, and fragmented product sub-markets is most saliently reflected in the historical financial performances of Analog Devices and Linear Technologies, who have been persistent value-generative outliers compared to their peers:

The exhibit above actually understates the relative financial strength of ADI and LLTC (Linear Technology).  If you stretched the time series back to 1993 for LLTC and 1996 for ADI, you’d see that each company has been solidly profitable on a clean EBIT basis (no add-backs) each and every single year, including the harrowing periods 2001-2002 and 2008-2009, a claim that almost no other semiconductor company can make.  Over the last 20 years, ADI’s unlevered after-tax return on gross capital has averaged 13%; Linear’s is more like 22%.  These are both stable, very well run companies.

In July 2016, Analog announced its $16bn cash/stock acquisition of Linear Technologies, offering a generous 11x revenue and 23x EBITDA (pre-synergy), levering its balance sheet to 3.8x net debt/EBITDA in the process (this has since been whittled down to 2.9x).  At the current stock price of ~$86, the combined company sports a ~$34bn market cap and a ~$41bn enterprise value.  The acquisition closed in March of this year.