Forbes: Semiconductors Rule And Embracing Software Will Allow That To Continue

This article was originally published on Forbes.

At CES last week, a company that most people haven’t heard of took away a coveted award for innovation – not an upstart Silicon Valley firm, but a sleepy old semiconductor company, Skyworks (NASDAQ: SWKS). The remarkable thing about this innovation award is that it wasn’t for a new integrated circuit, but for software that helps to eliminate unwanted noise from an audio signal. This is more evidence that an industry that underpins advances in the technology, automotive, health, and dozens of other industries is itself evolving rapidly from its roots.

I first started working with semiconductor companies over a decade ago. When one former client heard this, he was eager to take me out to lunch and warn me. He wanted me to know that it was a dead industry, “the days of silicon ruling the world are coming to an end; software and algorithms are the future.” His argument was that the industry had reached the limits of “Moore’s Law,” whereby the transistors in an integrated circuit (IC) double every eighteen to twenty-four months. This capability had spurred growth not only for the IC industry, but also the wider technology sector. The firms that could sustain this rate of innovation – Intel and Analog Devices amongst them – commanded superior financial returns because of the quality of their technologies.

Now, my cheerful lunch companion warned, the tech industry would find that lesser firms could provide chips that were “good enough” and then use software capabilities to tune performance to higher levels. Over the next decade they would start to feel the squeeze as lower cost competitors replicated their silicon advantage and more nimble software companies delivered the innovation the tech industry demanded.

Why Yesterday’s Industry Is Still Today’s

The logic rang true. Many semiconductor companies were founded in the late 1960s by small teams of engineers, many of whom were still involved in the companies as CEO or Chairman. I was privileged to meet several of these industry legends. They were still full of entrepreneurial fire, but most bemoaned the slow speed of innovation in the firms they founded. One complained that “the senior managers we have now are just operating the business for short-term returns so that it benefits their retirement plans.” Instead of keeping pace with the technology industry and exploring into new areas, he felt they were relying on cranking the wheel on the machine that he had created. This is classic organizational inertia: the more successful a firm becomes, the more likely it is to hone the capabilities, processes, leadership, and culture associated with that success.

Another thing my concerned client had right was that low-cost providers were coming to eat the lunch of the established IC companies. These firms focus less on technically improving the circuits themselves, and far more on radically reinventing the manufacturing processes that produce them. Monolithic Power Systems is the leader of the pack and has been nipping at the heel of once mighty firms like Texas Instruments, forcing them to defend themselves in core power electronics markets in which they once reigned supreme. As any student of Clay Christensen knows, disruption often emerges from the low end of industries, adding capability over time before it displaces the incumbent- unless they can find a renewed source of differentiation.

That is where we come back to the potential for software, algorithms, and artificial intelligence to enhance the value of the silicon to the end user. What’s interesting is that instead of representing a threat to the semiconductor industry, it has instead presented enormous areas of opportunity, if, you can overcome the gravitational pull from the existing business that promises a continuation of the good old days of Moore’s Law.

How Semiconductor Firms Stay On Top

One firm that embraced such a strategy for differentiation was Analog Devices (NASDAQ: ADI). Since 2013, its visionary CEO, Vincent Roche, has pursued a deliberate strategy of taking the firm “up the technology stack” into new sources of value, while at the same time picking off weaker competitors in a series of acquisitions. ADI’s software defined radio is a critical innovation behind the global proliferation of 5G. This ability to bring world-class software capabilities to transforming silicon performance is precisely what my friend had said was impossible. The net impact is a firm with a market cap over $90B firm, though Roche’s legacy may be even greater given that leaders who grew up under him are now assuming leadership roles across the industry. It’s no coincidence that the Skyworks innovation emerged from a unit led by Yusuf Jamal, a former SVP at ADI. This embrace of innovation extends beyond ADI. The German semiconductor firm Infineon (Frankfurt: IFX) recently launched a new contactless sleep device that combines software and silicon to improve the quality of sleep. There are many more examples that I cannot share. This is an industry on the move.

Why? It turns out that semiconductor companies are far more resilient than my lunch friend understood. Their hidden advantage is the exploratory nature of the electrical engineer. These are people that like to play with new technology and learn new things. This leads to a fair number of side projects and skunk works: engineers learning how to use machine learning and AI in their spare time, developing innovations “off the grid.” In my book, Corporate Explorer, I talk about how exploratory individuals in large corporations confound conventional wisdom to build new ventures form inside existing organizations. These electrical engineers are all potential Corporate Explorers.

Don’t Rely on Serendipity: Discipline your Innovation

The ingredient that separates successful Corporate Explorers from engineers with a secret project is the quality of their passion for solving customer problems. What makes machine learning algorithms powerful is their ability to solve a specific problem for a customer. In our follow-up book, Corporate Explorer Fieldbook, we provide the case study of an engineer that has developed a patch which can predict congestive heart failure three days before an incident occurs. This is “silicon plus algorithms plus deep domain insight plus an understanding of how the device is used by clinicians in practice.” Like the Infineon and Skyworks examples, this is a new created and led by a Corporate Explorer, who has identified the market need, advocated for the solution, and built support internally.

There are still many slow moving semiconductor companies out there, but the message is that this is a sector that is profoundly changing to find new sources of strategic advantage staying ahead of the classic low-cost disruptor. Those that learn how to value end-user and customer insight as much as technology – in the way leaders at Analog Devices, Skyworks, and Infineon have started to do – may find themselves unexpected winners in the world of AI.

It’s probably time I call my friendly former client back for another lunch. I wonder what he is saying now.