Wednesday, April 28, 2010

Apple Acquires Intrinsity

According to the Wall Street Journal, Apple has acquired start-up chip technology firm Intrinsity. The article explains that, "Some analysts believe that Intrinsity's technology is used in the A4, the Apple-designed chip used in the new iPad." As my students have learned, such an acquisition represents an example of backward integration, i.e. a firm choosing to produce its own components.

In general, I'm skeptical of vertical integration because I believe firms should pursue focused strategies and allow other companies to specialize at points of the value chain in which they have unique and superior capabilities. Moreover, I believe the transaction costs of using the market often are low enough that contracting with outside parties for component manufacturing makes more sense than in-house production.

However, this particular acquisition may have great benefits for Apple. Let's step back and look at several valid reasons to backward integrate. First, firms may do so to enhance the differentiation of their finished product. Apple certainly may want to own a chip technology firm for this reason. Second, firms may backward integrate so as to foreclose access to key inputs for key rivals. Apple may choose to keep certain innovations in chip technology away from competitors. Third, firms can benefit from backward integration if they find themselves in a situation in which asset specificity exists. What does that mean? Well, suppose that Intrinsity has to invest in assets and capabilities completely specific to Apple in order to supply parts for the iPad. Moreover, suppose Apple cannot secure that technology from any other supplier. Both firms would be beholden to each other because of assets specific to each other. That mutual dependence creates the potential for contracting difficulties if they remain independent. A solution to such challenges would be to bring the two firms together, i.e. an acquisition or merger. Thus, we have a potentially very strong rationale for backward integration in this instance.


Jagadeesh Venugopal said...

The chip business is capital intensive and cut-throat; it is dominated by the likes of Intel and AMD. Apple itself recognized this dominance when they went with Intel CPUs for the Mac some time ago.

I'm hard-pressed to see how Apple can compete with the likes of Intel and AMD who probably outsell Apple a hundred or a thousand to one in the chip (CPU) business. Because they sell so much, it is possible for Intel and AMD to ramp up the learning curve and go to newer generation processors much faster than Apple can.

Would it not make much more sense for Apple to focus on the "platform" that includes the OS, the applications on top of it, and the design of the devices, rather than venture into a low-quantity CPU business in which it has little or no experience and strengths?

Tim Berglund said...

The chip fabrication business is cut-throat and capital-intensive, but chip design is not. It is more akin to software design, where the value created takes the form of abstract, non-physical intellectual property. A fab-less semiconductor company differs from a software company in that its intellectual property output must be run through a capital-intensive manufacturing process, but there is—subject to market conditions—ample excess fab capacity in the world for a small chip design firm to be able to sell physical chips to customers.

As the proud new owners of Intrinsity, Apple must now contract with a fab to produce iPad processors, but it was almost certainly in this situation before the acquisition. It probably had one or two two nontrivial application-specific integrated circuits in its each of its devices, each of which would require duly sourced fab capacity to produce reliably and in volume.

Note that I'm speaking in terms of generalities here, and could be wrong on any number of specifics. However, the distinction between Intel's manufacturing business and a start-up's CPU design business remains a strong one.

Jagadeesh Venugopal said...

The point though is that a chip is a chip is a chip. Whether we're talking about memory chips, CPUs, disk controllers, graphic chips, etc. they can all be controlled via software and provide no discernible competitive advantage.

Apple could just as easily recompile its software to run on garden-variety CPUs. Unless, Apple has some strong chipmaking IP that we're not aware of. In which case, are they a media company? a software company? a hardware company? a fabless chipmaker?

Tim Berglund said...

But memory chips and disk controllers are not the same as CPUs and graphics processors. The former are commodity items, but the latter can exhibit significant differentiation that can and does confer a competitive advantage. Processors selected for mobile devices must run as fast as possible while using as little power as possible—priorities that are at odds with each other during design. The way the chip design trades off between those competing priorities (and various others) is the essence of the chip's differentiation in the marketplace, in a manner directly akin to the way one operating system or database server might be different from another. While some might argue that OSes and relational databases are being commoditized, there is still a flowering of diverse ideas in system management and information retrieval in general, just as there is in mobile CPUs.

You may be right that Apple's core competencies do and ought to lie at different points in the value chain than at the semiconductor level, and I would be willing to concede that point fairly easily. However, given the non-commodity nature of CPU designs and the considerations Professor Roberto outlined in his blog post, Apple could still be on to something.