AMD Dragon vs. NVIDIA ION: Making and Breaking Technology Bundles


This week, AMD launched its Dragon Platform. It brings back the concept of a technology bundle as the company moves against a much larger and entrenched vendor in its space. NVIDIA continues to push with its Ion graphics platform, which it is using to try and break the Intel Atom bundle of processor, chipset and imbedded graphics parts in the hot netbook market. The use of bundles, and the reasons to break them, is an ongoing drama in the tech market.


Microsoft Office: The First Big Tech Bundle


The first really big tech bundle was Microsoft Office. It combined what was then, and is largely now, the core applications that people wanted on a computer. These were a word processor, a spreadsheet, a presentation product, and an e-mail product. It was actually predated by a similar and better-integrated product called Lotus Symphony, which failed in the market. In 1984, or eight years before Office 3.0 released, Symphony had a spreadsheet component, a word-processing component, a graphical-presentation component, a database-management component and a communications component.


Symphony was Lotus 1-2-3 expanded to do the additional functions, and while this led to much better integration, it also made the stand-alone components from competitors like Corel, Microsoft, Ashton Tate and Borland vastly better. Best of breed won out and Symphony failed. The migration from any competitive offering to Symphony was ugly, raising the switching costs to unacceptable levels.


When Microsoft Office 3.0 came out eight years later, it was a bundle of competitive products and even though it wasn't well integrated, it was priced aggressively. This, coupled with the fact that each major component was competitive, let it lead the market. This was when the Microsoft "Embrace and Extend" strategy became evident. Virtually all the components were file compatible with the products the bundle would eventually grow to replace, creating an easy competitive migration path from the stand-alone offerings to Office.


The lesson here is that bundles work if the bundle contains components that are good enough, it has a low migration cost and the price is aggressive. They fail if any one of these components is out of line with requirements.


Dragon: Making a Bundle


One of the nice things about the PC market is you don't worry too much about migration cost because the operating environment can be kept common across all X86 platforms as long as you remain within Windows, Linux or the MacOS. This eliminates the user switching cost to something like Dragon.


For the bundle to work, AMD has to have competitive or better performance at a more aggressive price. This is where their ATI component and unique three-core concept come into play. Intel still lacks strong graphics and the AMD three-core processor is positioned against the Intel two-core offering.


The first provides a performance advantage even at the low end and the last provides performance headroom for free in a market where most applications don't even use two cores effectively making three, not four, the ideal path for those that want to get the most for the least amount of money and still avoid premature obsolescence.


It's a good bundle and wrapping it with the Dragon brand gives AMD an easier path to marketing the related solution. Finally, there are the perceived benefits of being able to better tune the entire platform, ensure that drivers don't conflict, and providing related software to better make use of the result.


NVIDIA Breaking a Bundle


Intel bundles its amazing Atom processor with a chipset and a very limited Intel graphics solution. NVIDIA brought out Ion which, when combined with Atom, creates a result that is competitive with much more expensive laptop computers. Intel's bundle is priced aggressively, and has the advantages associated with the Dragon bundle above, but it misses on one key part and that is graphics.


This limits the related netbook platform and prevents it from being as good for movies (particularly high-definition movies) and games. NVIDIA is attempting to break Intel's bundle on graphics for a product class, netbooks, which have continued to sell well despite the massive economic pressures on the segment.

What is interesting is that once the bundle incentives are removed from the Atom solution, the result should actually be more profitable for Intel, which should make it favor the NVIDIA Ion path. But I haven't seen Intel even endorse it yet, probably for fear the result would cannibalize the more expensive Core 2 Product line by pushing netbook performance into that range.


The Atom + Ion-based netbook will be an interesting test. It does provide a significant performance advantage, which should allow it to break the Intel bundle for those wanting more graphics performance and create a high end of the netbook segment. This high end will likely be hard to differentiate from the low end of the notebook market and may outperform notebooks, with the end result a more pronounced blurring of the two product classes than now exists.


Wrapping Up


Bundles, when done right, are very hard to compete against and provide some of the best values in the market segment. The AMD Dragon platform is an example of a good bundle. Breaking a successful bundle like Intel has created for Atom is much more difficult but, if NVIDIA's Ion is successful, it should move netbooks farther into the mainstream. The result should be a much more interesting notebook market by year end.