Maestro PR Blog: The Integrated Device Manufacturing Model: Use It or Lose It

Back in the late-1980s, integrated device manufacturers (IDMs)were so sure of the soundness of their business model that they scoffed at the idea of only offering manufacturing or only doing semiconductor chip design.  Those of us old enough to remember recall the maxim that ‘real men have fabs’.  Unfortunately, it turned out that a lot of real men also wanted to design chips and let someone else deal with the financial and production hassles of getting designs to silicon and into the customer’s hands.   We know the result:   the meteoric rise of foundries and fabless semiconductor companies at the expense of IDM revenues and the allure of their business model.

We can breeze by the 1990s, the dot-com boom-and-bust, and the internet age.  For those past 20-plus years, the models of fab, fabless, and internet companies became accepted, refined, and successful. And the IDM model? It’s muddled along by streamlining, going ‘fab-lite’, and moving into markets like MEMS, where older process technologies are still viable.  Will IDMs never learn???

Fast forward to today in 2014. Business models as we know them are in for changes that will slam many companies into further oblivion.  And it’s all happening with the force of a tsunami moving with a force that is deceptively lazy but unrelenting in  its speed and the force with which it will hit the global shoreline. 


The Tsunami named “Google”.

The name of the tsunami is Google.  No longer an internet/social-media company, it is moving inexorably to a semiconductor model that will rock fabless and IDM companies in the next few years.  How did it begin?

In December, while Wall Street was busy watching retail sales and consumers were checking off their holiday hsopping lists, a Google ‘insider’ hinted to a Bloomberg journalist that the web-search giant was beginning to design its own server processors. It’s unclear how the rumor started or its substance. Maybe it was because some well-known computer architects (Luiz Andre Barroso, Urs Hotzle, James Laudon) already work there.  Or, maybe it’s because Google posted a job opening for a “digital design engineer” with expertise in ASICs.

Word of the shift moved from Bloomberg and spread to multi-national wire services quickly, and then faded away like smoke from a cigarette. After all, the Google spokeswoman, Liz Markham, dismissed the rumor by saying Google is “actively engaged in designing the world’s best infrastructure.” And that, she added, “includes both hardware design (at all levels) and software design.”  So much for a definitive answer.  But, where there’s smoke, there’s always some fire somewhere.

Currently, Google designs its own server systems and buys its server processors from Intel.  If the December rumors prove true, Google would begin designing its own computer chips using ARM Holding’s technology. ARM’s modular architecture lets designers create custom chips using only those blocks necessary to a customer’s design requirements.  And that’s the key concept here: custom design.


What IDMs Know.

As any IDM knows, custom chips provide the best performance. Custom design will allow Google to tailor chips to their architecture, to their specific applications, to the increasing demand for video, and to other services only imagined on the horizon.  It should make an IDM’s head reel at the possibilities.

The major benefit is that custom-designed chips can help speed search services where new apps are popping up almost daily.

So, the benefits of customization are performance, speed, architecture efficiency,  and – oh – cost. Google can potentially save millions by designing its own chips.

Add to this news, the January 17 announcement that Google is developing a glucose monitor in the form of a contact lens.  Our tears contain glucose, and this lens will be able to track glucose levels without the pain and problems of pricking a finger. 

According to an article in the San Jose Mercury News by Associate Press journalist, Martha Mendoza, the lens has miniature transistors – the stuff of semiconductor chips – with a fine antenna.  Along with the lens, Mendoza reported that Google has also developed a driverless car. And, of course, we all know about the web-surfing eyeglasses and its proposed network of balloons that will deliver Internet signals to the remotest places in the world.

Google’s exploits are the momentous tsunami I referred to earlier.  The impact will be felt in a variety of ways.  First, Google now moves into a competitive space. Did you really ever think of a web-search company as a fabless semiconductor company? 

Secondly, a world-class company (yes, we’re still talking about Google) is taking major steps toward aggregation of  services, systems, architecture and the chips that are the foundation for all three. 

If you’re in the IDM business, you know what aggregation is. It’s the basis of the IDM model; ie, a company that does it all: design, build, package and assemble chips and then sell them as neat one-size-does-it-all devices that can be purchased off the shelf. 

Third, it is an implicit recognition that market leadership – even in software – is based on hardware.

The rhetorical questions for semiconductor companies are these: Is Google an anomaly? A company beyond all companies in the world?  A company with more imagination, more energy, and more determination to be a problem solver as well as a brilliant success?  Or, are we witnessing another tsunami of change in our industry?

IDM companies laughed at the fab and fabless models until they were passed by and had to scramble to stay viable on Main Street as well as Wall Street.  IDMs then shrugged their shoulders and looked for new revenues in markets  that didn’t require advanced processes and technologies.  Now, Google is forging a new business model.  IDMs can strengthen their aggregated model and begin innovating in different ways that serve current and new markets.  OR, IDMs can once again laugh at, ignore, or watch the future roll over them. 

What’s an IDM to Do?

IDMs are in a perfect position to leverage news that chip design has gained allure amongst companies outside the traditional semiconductor sector.  Why? Because IDMs know the value of owning chip design and manufacturing better than any other business model in the  industry.  (For several reasons – too lengthy for this piece – it seems that only IBM, Intel, and a very few other companies have been able to make the model a true success since the start of the fabless model.)

Aside from the technology innovation standpoint, the ownership of design and fabrication is really about money.  You just get the best performance – that is, speed-- design, product quality, and speed to market when you own the entire chain from idea to finished IC (integrated circuit). If you’re an IDM, you own most of the infrastructure needed to gain time to market.

Your Action List.

  1.  Get out of the office. IDMs need to take a fresh look at the trends shaping the industry. Google did not think up the idea of a contact lens for measuring glucose levels in their lab. The University of Washington did, and then reported on their invention at ISSCC some years ago.  Google then developed it for market adoption.


  1. Attend technology events that feature technology innovations and connect the dots between their areas of expertise and the new ideas being presented. ISSCC is still the best venue for this, in my opinion.  But so are the SPIE Photonics, Opto, and Lithography conferences.


  1. Speak up!  IDM’ers also need to be active participants in technology events, presenting papers that describe their design and manufacturing expertise. Through these two activities, IDMs hear what processes are required for innovative technologies, and what the challenges are to getting them.  Fab managers are in the best position to link their company’s expertise to what’s needed in the future.  Remember, fabless companies do not understand all of the nuances of manufacturing. They must rely on the foundry. And why shouldn’t that foundry be yours?  So, turn off your Facebook page for a day, because face-to-face meetings are still the way to form business friendships fast.


  1. Fab managers need to know process technology backwards and forwards. Not just your own, but other processes, as well.  Once you understand process technology, you can figure out the equipment you need. And I don’t mean upgrades to your current equipment.  You’ve got equipment vendor partners in place with the know-how to get you what you want. Utilize them.


  1. Advocate the need for R&D. Yes, this is painful, but it’s the only way to survive and thrive financially.  I’m not talking about building an R&D ivory tower.  I’m talking about collaboration. Work with universities, think-tanks like Imec, or others. Leverage grants.  Do more than just manufacture and sell.


  1. Talk the talk of CEOs and CFOs. They care about money, not another piece of equipment. Fab managers need to talk their language. Describe compatible products in new market sectors. Link the manufacturing technology needed to support new device types. Describing how your equipment is failing and why your production line was down for hours is not their problem. It’s yours. Even if it is their problem at the end of the financial quarter.


  1. Keep customers happy.  This sounds so easy and is the hardest thing to do. If you doubt this statement, try being your own contractor for a kitchen remodel.  Regardless of how beautiful a plan you have, the torture is in getting all the vendors to cooperate and be on time. Google is designing chips and end-market products because aggregation has many more benefits than coordinating vendors and matching time schedules like puzzle pieces.  When you know your customer, you know their current problems and future opportunities.  Yes, you may be in a fab and not see customers often, but you do see them. Maximize the opportunity.


  1. Leverage what you’ve learned by being on the front lines. But keep an open mind. Innovators didn’t succeed by reshaping the mold. They broke it.


A note to materials and packaging companies:  You really have an open door in terms of spinning ideas off the diabetic contact lens concept.  Innovative companies already make flexible endoscopes that fit inside a ball-point pen. The need for similar products made from silicone and polymers will be even greater as the home medical market continues to heat up.

Innovation and trends don’t spring from a hole in the universe. They are the children of need, desire or, in the case of competition, the drive for greater revenue and market share.  There’s nothing stopping you or your company to innovate, grab a trend or leverage one.  You can start anywhere.  Just start.


~Barbara Kalkis, Maestro Marketing & Public Relations 

The opinions expressed in the blog are those of the author.