Allen Benn's take on INTC, CSCO and WIND
John & everyone,
I hope the following post will encourage G&K folks to also bookmark the SI-WIND thread. We struggled a bit in his absence, but now that Allen is again posting, I think technically savvy folks will find our thread even more worth their time :)
By the way, I think Dr. Benn understood the 'Gorilla Game' principles well before the FM was published, because QCOM was one of his favorites way back when ...
--alan
Message 13529274
From: Allen Benn
Sunday, Apr 30, 2000 ÿ2:51 PM ET
Reply # of 7709
Over the last year, I have been surprised and disappointed how the Street consistently fails to appreciate the growing importance of WIND's relationships with major companies worldwide. Most of all, I struggle to understand why no one has connected all the dots filling in WIND's strengthening relationships with both Intel and Cisco. Intel's New York analyst conference last week was one more dot that should depict an easily recognizable picture.
In trying to answer James probing question about the continuing importance of I2O, I found myself once again arguably claiming that I2O alone justifies WIND's market cap. What I didn't say much about are the other announced relationships with Intel and Cisco, and what they mean for WIND's future. Imagine what it must feel like for me to be arguing the importance of I2O, while all the time realizing that I2O is just the tip of the Intel/Cisco iceberg. I think this is my pedantic way of saying, "You ain't heard anything yet."
Intel recognized years ago that the PC paradigm was getting long of tooth. Scurrying to the relative safety of high-end servers helps in the short term, but more will be needed as growth in PC's continue to play out and servers fail to make up the slack. In the search for new growth opportunities, Intel found the Internet, particularly anything silicon in the Internet. While Intel's size and economy of scale translate into a powerful competitor for anything silicon, it is not Intel's style simply to dominate a commodity business by brute force. Commodity businesses are never easy to run profitably, as economies of scale bump up against diminishing marginal returns. These are the two counter forces primarily causing global giants on the one hand, while preventing monopolies on the other.
One second after Intel announced Internet eXchange Architecture last September, you should have realized Intel had found its bearing. The next major initiative underway at Intel is not just targeting network equipment, but targeting it the Intel way. Expect Intel to unleash its formidable resources to brand IXA and to make it a de facto standard underlying the lion share of most network equipment. The approach Intel is taking repeats their tried and true success formula, honed in the computer industry: Create a proprietary processor family that energizes a mass market of software and hardware developers that compete successfully with inflexible hardware and expensive proprietary software solutions.
This takes no interpretation on my part. All you have to do is listen to what Intel is saying, and watch what they are doing. What seems to have evaded most, if not all, pundits is what this means for WIND. Not one pundit has connected the obvious dots that the primary winner when this was done in the computer industry was the software provider, Microsoft. If again software is the key, is there a software winner in the wings? According to what Intel said last September, at the February IDF, and at the analyst conference the other day in New York, software is indeed the key. By fixing an API for logical manipulation of packet content processed by multiple microengines, and an API for the StrongARM core, that applies across the family of network processors, Intel sets in motion the development of a huge support industry leading to a mass market and the commoditization of network equipment-just like PCs.
Now think. Will IXA succeed any faster if the software platform implementing the APIs takes more than one form? Of course the answer is no. Anything that increases variety decreases network effects. Indeed, even if a focus on a single software platform limits short-term growth in units, the benefit from not diluting network effects will more than compensate. Further, Microsoft's success with DOS and then Windows proves conclusively that, as a whole, industry only cares that the software platform is adequate. The only thing developers really want beyond adequacy is constancy, universal access to an invariant mass market.
As beneficiary of the world's greatest software monopoly, Intel has to know at least as much about the underlying economic realities as I do, and want to limit IXA development to their reference software platform, provided by WIND. However, being careful to avoid the legal pitfalls that trapped Microsoft, Intel will never act overtly to restrict third-party software variation on IXA platforms. This means companies like Microware are free to invest their limited resources to try to hitch their OS for a ride on the IXA express, as they announced not long ago. But industry knows better than to chase down blind alleys. Expect virtually every serious developer planning to use IXP to choose Intel's reference software design. While this no doubt is true today, this expectation will become ironclad as IXA becomes prevalent and network effects take hold. Since Tornado/VxWorks provides the foundation for Intel's reference IXA design, there can be no other conclusion but that WIND has been anointed for greatness on this basis alone.
As time goes by and networks effects miraculously enhance value-add, WIND will find itself with awesome pricing power. Just as Microsoft makes most of its money on applications, expect WIND to benefit increasingly from high-end support software, like TMS, or total solutions consisting of complete network equipment application software. With the acquisition of EST, it is imaginable that WIND could move into the hardware realm as well.
Surely by now you have connected the I2O dots I posted the other day with the IXA dots in this post. Do you recall the iLAN NIC I described that could contain an IXP chip turning a NIC into a high-end switch currently worth thousands of dollars? This means that Intel is able to boost the initial value-add of IXA by tapping the network effects just beginning for the more mature, developing I2O technology.
I suspect that Intel expects IXA mostly to be lodged in traditional boxes hosted by embedded Intel-Architecture processors, embedded Pentium III's, etc. What OS might they be running? Linux and Windows NT/2000, or perhaps Solaris and Netware, come immediately to mind. But don't forget that Intel has a dedicated team of engineers in residence at WIND to optimize WIND's OS's and tools for embedded I-A processors. This means that VxWorks is a serious contender for this piece of the business as well. Developers experienced working with Tornado for IXA and/or Tornado for IxWorks will appreciate the consistency and reliability of having access to a familar software platform for the host equipment.
WIND's relationship developing with Cisco, if anything, could end up being stronger than the relationship with Intel. For anyone concerned about the threat of free operating systems, imagine this: WIND's OS and tools is beating out Cisco's IOS, not outside in open competition, but inside Cisco where the playing field is anything but level. If the renown Cisco IOS can't compete successfully at zero cost and buoyed by policies dictating its preference within Cisco, it should be obvious that WIND's leadership position can overcome any challenge in open competition.
Cisco IOS is loosing the technology battle to WIND because virtually every acquisition they make includes network devices made using WIND's products. Cisco cannot pay $6.9 billion for Cerent, and then risk destroying its success formula by requiring the company to migrate products to home-grown tools. The problem for Cisco is that this conundrum is only getting worse, not better. Ultimately, it will force Cisco to migrate core products to VxWorks, to maintain some semblance of consistency and compatibility throughout the organization.
Meanwhile, don't forget that Cisco set up a Global Alliance Network program a year or so ago, to extend Cisco IP to the edge of the Internet, i.e., at the entry point to homes and small offices: DSL, Cable Modems, etc. You might recall WIND's announcements supporting this program, including comments during conference calls. The relationship takes the form of a VAR relationship. That means Cisco helps device makers worldwide with a sell-through to WIND's products. How many analysts in NY realized that the DSL product Intel announced was developed in concert with Cisco almost certainly was based on this reference design.
The dots connect not only within Intel and Cisco separately, but they also can be connected at a broader level that simultaneously includes both companies and others. By now you have to be wondering whether these relationships are fully understood in the executive suites of the company's involved? What are the implications if they are, or if they are not? And equally intriguing, why have pundits or most money managers failed to connect the dots?
Can you figure out how this game ultimately will play out?
Allen |
