Wind River Analyst Day Summary
1. Jerry Fiddler opened the analyst day with an excellent visionary statement, extending past developments into current requirements and future expectations.
2. Tom St Dennis smoothly introduced the new WIND organizational structure. Last summer, Wind River Systems organized around vertical and horizontal business units. The organization has been enriched with the ISI merger and the acquisition of EST. Horizontal units are (1) professional services, (2) platforms (OS, tools, etc.) and (3) HSI, hardware-software integration, which essentially is EST. Vertical units are (1) networks, (2) consumer products and (3) TDI, which stands for Transportation, Defense and Industrial Automation.
3. While most analysts probably knew about WIND's new organizational structure, I doubt any of them fully appreciated the value-proposition embedded in each vertical business unit. Curt Schacker, Marketing VP, tackled this deficiency by fleshing out the organizational skeleton, first by presenting the market forces that demand increasingly high-valued vertical solutions, and second with an illustrative walk-through. He presented the following about DSL:
WIND is in 47 of 52 known company's DSL products, or 90%. This is not the same as market share, but is indicative of a dominant market presence. DSL CACG is 230% according to credible market research.
Growth X Share = Units
ASP for DSL modems averages about $1 when WIND supplies the underlying OS and development environment. WRS Networks is developing "Tornado for DSL", modeled after Tornado for Managed Switches, which enables easy addition of second-generation features like VoIP, ATM, Multiple protocols, Network Address Translation, VPN, Firewalls and other Security. The royalty ASP for Tornado for DSL will be in the $3 to $5 range, depending.
Units X ASP = Revenues
When I plugged these number into a spreadsheet, assuming 1999 DSL units = 200,000, and a gradual transition to high-value sales, I got DSL revenues growing from about $180K to about $70 million by 2003.
The VP then pleaded with the analysts in attendance NOT to view WIND as a DSL company. DSL is only an example of a market segment WIND owns on the early end explosive growth, just like a lily pond.
4. WRS Networks, which gets credit ONLY for vertical solutions that lie on top WIND's basic platform solution, Tornado/VxWorks, expects to grow to 50% of total company business quickly with sustained growth over 100% annually. Telecommunication and data communications (Network's domain) currently represents about 50% of WIND's $350 in revenues, but the amount credited to the new unit, with new verticals, is much less, I would guess around $40 million. Dave Frazer, the General Manager, clearly bases his expectations on a number of market segments considerably larger than the DSL example (the vertical portion of which he owns).
These include:
(a) A network storage device that uses the Intelligent RAID building blocks jointly developed by WIND and Intel and is based on Linux.
(b) InfiniBand use of Intel's IOPs each of which pay a significant royalty to WIND.
(c) Intelligent LAN (iLAN) NIC's being jointly developed with Intel to enable servers and workstations to offload TCP/IP, felt to be absolutely necessary to access fast LANs at wire speed without affecting performance of the CPU. Recall that Intel owns about 50% of the NIC business, and has the market weight to push iLAN to prominence just like they are doing with RAMBUS.
(d) A vertical, shrink-wrapped software market on top of iLANs to sell VPN, VoIP, firewalls, security, encryption, compression, switches, routers, etc. The latter two can be developed by plugging in Intel's IXP 1200 into an iLAN.
(e) Tornados for DSL, Cable Modems and Set Top Boxes to be available soon.
(f) Tornado for Managed Switches. In particular, broadband optical switches is booming, and is positively huge. I suspect they would have preferred to present the Broadband Optical Switch lily pond, but they haven't gathered up all the requisite data.
(g) And we didn't even get to WIND's reference design for IXA, which is Intel's new initiative for commoditizing the network equipment space.
5. WIND Services disappointed me in Q1 with tepid 10% growth. The explanation given was that ISI hadn't integrated Dr. Design into their sales channel. The situation now is that WIND's massive sales channel is being tapped for professional services, and the demand is overwhelming, limited only by capacity. Expect services to grow between 40% and 50% year-on-year starting as early as Q2.
6. TDI. Military/Aerospace are doing OK. Industrial Automation is holding its own. But automobile transportation finally is starting to hum. WIND is the only company that has an OSEK-compliant kernel with development tools for auto controls AND a presence in the in-the-cab systems like navigation systems, Internet in the car, radar warning systems, entertainment consoles, etc. WIND owns over 50% of the auto navigational system segment, for example, which is beginning to ramp -- another lily pond. WIND's OSEK kernel (which has never had a reported bug in its two plus years of production existence) is in a growing number of control systems, like fuel injector control systems. Since the OSEK kernel starts with a 1.5KB footprint, it is being eyed as an OS for low-end consumer products, not just for auto electronic controls. And WIND's MatrixX design tool, used strategically in many auto companies, does lead to downstream design wins (in at least one case without any need for hand-written software.) The latest version of MatrixX, in development for over a year, ships in a month.
Incidentally, WIND mostly pursued development tools for OSEK, relying on 3Soft's OSEK kernel, while ISI invested heavily in the OSEL kernel and, of course, MatrixX. The combined company's resulting AutoWorks meshed these developments together virtually without overlap.
7. Embedded Java finally is becoming a real-world standard for consumer devices, because of the platform independence, inherent security, etc. WIND is Embedded Java, since WIND alone accounts for more than half of all embedded Java design wins.
WRS Consumer Products expects to account for 20% of total revenues in the future, quite a task if WRS Networks takes 50%, Services about 30%, autos taking off in TDI, Platforms using its broad base to garner significant sales, and HSI becoming a key technology to integrate software with rapidly emerging semiconductor products.
8. WIND has a Center of Excellence (CofE) deal with Intel for embedded Pentiums, which has been extended to the StrongARM. We know WIND announced a new CofE with the MIPS that extends to licensees. WIND has an unannounced CofE with Hitachi, and plans to complete agreements soon with Motorola and IBM re the PowerPC, ARM and another one or two. WIND's objective to cover with CofE's the vast majority of the semiconductor usage of WIND's active customers. CofE's are a way for semiconductor companies to join with WIND to develop fully integrated, optimized hardware-software platforms for beginning development by high-level application programming teams. I think CofE's form a self-reinforcing process that, while started because WIND is the leading software provider on all these popular platforms, their effect will be to extend WIND's dominance. In some cases, the semiconductor company promotes additional sales by entering a VAR agreement whereby WIND does not charge the end-developer a royalty but receives it directly from the semiconductor company (like the IOP agreement with Intel). WIND now has 27 VAR deals, although not all are with semiconductor companies.
9. CFO guidance is unchanged with 37% revenue growth for the year and 51 cents EPS. Now that the merger is pretty well digested, expenses will start to be controlled better, mainly using headcount. Data systems should be in place progressively over the year, enabling WIND to be more transparent in the future. I presume this means more lily pond stories.
10. WIND's just-announced Cirrus version of VxWorks removes any remaining barriers to VxWorks substituting for proprietary OSs for High Available systems, like what controls carrier-level communication devices. (WIND software is almost always in line-cards today, but now can be in control as well. This enables WIND's entry into a new super-high-end vertical market for WIND. I suspect is Cirrus lives up to its goals, WRS Networks will select it in its next version of network storage solutions.
11. Jerry demonstrated an Internet PCMS Switch and Appliances, including Pingtel's VoIP phone, details of which you can find on WIND's web page.
wrs.com
12. In a mid-day Q&A with all analysts in attendance, the focus was on the Linux threat. While the company appeared comfortable with Linux as an open-source technology they can utilize like they did with GNU compilers and other open-source technologies, I think some analysts remained unconvinced. They seemed to fear that Linux is a threat, not so much to WIND technology as their to their business model even though WIND is now positioned to profit mostly from verticals above the OS. In fairness to the concerned analysts, let me say that this is an exquisitely complex issue to decide before it plays out. Since exquisitely complex analysis is what this thread is all about, I will try to post a complete explanation of why Linux does not pose a credible threat to WIND.
13. After a short lunch, the analysts broke into small groups and rotated through breakout sessions with business unit GM's.
Allen Benn |
