This week is the second anniversary for OpenPOWER, and it has made huge strides, based on the presentation it just shared with us. We often drop down and talk about technology in terms of speeds and feeds, with the inherent assumption that the best part wins. But that is rarely true. It isn’t the part that fights the fight, it is the business model. IBM was dominant when Apple came to market and Apple took its business model of vertical integration to market, but both IBM and Apple discovered the weakness in that model when Microsoft slapped both and took over the computer market with Intel in the 1990s. Linux entered the market in the 1990s using a very unique model and successfully moved against Microsoft the following decade. Apple didn’t give up and Steve Jobs once again used the IBM model to open up the then-new mobile market. Then Google, with a blend of the Microsoft and Linux models, surprised Apple, and now it is blending that model further with Nexus, which adds Apple elements.
This showcases that the business model may be more powerful than the technology, and we have an interesting battle brewing between Intel and IBM.
Here we have Intel and IBM going after each other, Intel using a business model similar to Microsoft’s and IBM using a model more like Linux than Android. In a weird way, IBM is using the exact opposite of the model it used to once become dominant, and it appears to be working. Let’s talk about the advantages and disadvantages of each.
Intel: Control, Consistency, and Economies of Scale
The economic advantage of Intel and Microsoft’s model against the more traditional Apple and IBM model has been dramatic. IBM and Apple were based on vertical integration and customer ownership. This model works really well in an emerging market, like computers was in the 1970s and mobile devices were in the 2000s, because most of the parts aren’t mature and the common suppliers don’t really know what to build. We saw this play out between Microsoft’s Plays for Sure and the iPod. Apple tightly controlled the experience using largely proprietary technology, and the fear of litigation kept Microsoft from building a clone. As a result, both the Plays for Sure effort based on its own model, and the Zune effort based on Apple’s failed. However, Google basically copied iOS with Android, accepted the litigation as a cost (much like Microsoft did in the 1980s), and was successful with phones where Microsoft failed.
The path to winning using the Microsoft model is that the firm has to become the standard and get a critical mass of partners so that, with strong margins, the economies of scale allow it to provide a better value. The downside is that the end user may get lost in the process and, because the supplier becomes more powerful than any one customer, the firm may drift far from what the user wants and/or become unable to anticipate their needs. Drivers can become more about control and assuring margins than building and sustaining customer satisfaction, partially because it becomes increasingly unclear who the customer actually is. Is it the OEMs or is it the end user? And, finally, the more dominant the firm becomes, the less profitable its customers become, largely because it becomes harder and harder to differentiate on anything but price, which drives down margins and further erodes end-user satisfaction.
So, for long-term success, the central vendor has to drive aggressive R&D, force some connection to the end user, and maintain a massive focus on a strong blend of end user and OEM customer satisfaction (which includes a strong marketing element). Currently, Intel is lagging in several areas, which is why it is vulnerable.
OpenPOWER: Collaboration, Customization and Customer Focus
What made Linux different wasn’t just the economics (which, in software actually reduced its competiveness because it largely removed profit), it was the collaborative nature of the effort. Those who used it or built on it participated directly in its creation. This forced a relationship between the critical parties that tends to erode if a supplier becomes too powerful. In effect, the OpenPOWER ecosystem forces cooperation and collaboration, and it puts much of the direction in the hands of the member companies who, by nature, are closer to their buyers than a chip supplier can be. To function, the core member has to give up power and become a peer; if it retains absolute control, the competition with the member companies will drive it out of the consortium because it will see IBM’s advantage as unfair. IBM is in the process of transitioning out of the leadership role.
You get far more differentiation, the products should better align with what the end buyers want given it is those needs that drive the ecosystem, and customer satisfaction is built in; it isn’t an add-on process. If folks aren’t satisfied, they’ll exit the consortium.
However, this model has its own set of problems. The products can fragment as the need to differentiate can overcome the economies of scale necessary for a competitive parts price. The lack of one clear leader can result in disagreements that significantly slow advancement, requiring some type of external threat to keep the parties focused; and an excessive tendency to focus on current problems and completely lose track of long-term development.
The OpenPOWER Foundation has a branding requirement that keeps fragmentation down, IBM remains the clear leader and is experienced in driving consortiums forward, and Intel remains the common threat that keeps the parties focused and aren’t going away any time soon.
Wrapping Up: Choices and End Game
In theory, the more generic the problem, the better the Intel model will work. The more unique, the better the OpenPOWER model will work. Intel will have sustaining advantages in economies of scale, OpenPOWER in differentiation and unique market/business focus. Intel’s market is larger, OpenPOWER’s should be able to hold higher individual margins but lower overall revenue. If the bid is mostly on price, Intel should have the advantage. If it is mostly on being able to do something unique, OpenPOWER should prevail. Intel could better challenge OpenPOWER by creating an AMD-like semi-custom consortium (which doesn’t exist at the moment). OpenPOWER could challenge by creating a part that could address the mobile market generically, giving it far better economies of scale (likely a bridge too far at the moment).
Intel’s Apple relationship is very close to semi-custom with Apple, but it seems unlikely it will want to expand that experience given how contentious that has traditionally been. OpenPOWER could bring AMD into the fold to gain scale and AMD is run by ex-IBM executives, but AMD is already spread very thinly and likely can’t make that step.
For now, neither entity is willing/able to remove the other and the competition should continue, with each winning in the server space where their business model best fits.
For some reason, this brings up the old saying, “If you love something, let it go. If it comes back, it is yours. If it doesn’t, it never was.” Maybe this should replace “Think” as IBM’s OpenPOWER tag line.
Rob Enderle is President and Principal Analyst of the Enderle Group, a forward-looking emerging technology advisory firm. With over 30 years’ experience in emerging technologies, he has provided regional and global companies with guidance in how to better target customer needs; create new business opportunities; anticipate technology changes; select vendors and products; and present their products in the best possible light. Rob covers the technology industry broadly. Before founding the Enderle Group, Rob was the Senior Research Fellow for Forrester Research and the Giga Information Group, and held senior positions at IBM and ROLM. Follow Rob on Twitter @enderle, on Facebook and on Google+