Chuz -
About 80% of the technology is readily available. Maybe half of the remaining 20% is available to companies like Dell which partner closely with the key technology providers (in particular for Dell, a very close partnership with Intel).
CPQ invests most of it's R&D money into longer term technology (examples below). Unlike IBM, most of CPQ's R&D eventually makes it into product.
R&D is essentially an investment in future products and processes, and thus, should be logically independent of current results.
I agree, and apparently so does CPQ. They place some constraints on R&D through a process called 'portfolio management'. The idea is that there is a target for overall R&D spending (which comes from the board). Within that total, there are guidelines around what percentage to apply to near-term development, which to longer term strategic, and what goes into 'wild cat' R&D. There guidelines are applied by a corporate governance body which reviews the portfolio and approves according to the portfolio guidelines. This allows them to control overall spending without excessively limiting the potential for innovation.
CPQ's goal with this work is to gain a 9-12 month advantage in the market on key technologies. They understand that in order to succeed these technologies must become broadly available, so CPQ has adopted a model of turning much of the technology over to standards bodies or industry leaders like Intel, which makes the work available to CPQ's competitors. But CPQ retains some key IP, and also gets licensing revenue for the technology, currently generating more than $1B annually.
Some examples of recent development by CPQ -
1) CPQ did a joint venture with Corollary to develop an 8-way memory chipset for Intel architectures. IBM and Axil - DG also developed competing technology in this area but CPQ - Corellary was much higher performance (about 4X the competing designs in terms of memory latency). The product was so attractive that Intel bought Corellary to get the technology. Now CPQ will get royalties from Intel for all 8-way chipsets, in addition to some control over future designs.
2) CPQ, MSFT and Intel developed a design for cluster interconnects called the VI Architecture. CPQ (and Tandem) did most of the base ASIC design and produced the reference silicon. There is little doubt that this will be the basis of next-generation clusters, and CPQ will have both design control and first products ahead of the general market.
3) CPQ developed technology for 'hot-plug' PCI which was proposed to MSFT and Intel as the standard. CPQ provided the core silicon designs. Here again CPQ will get revenue through cross licensing, and design control.
4) CPQ developed crossbar technology for 32 way and 64 way SMP designs, initially for Alpha but eventually for Intel (in the Merced timeframe). This provides CPQ with a way to address the largest IT users. Currently only Sun has similar technology.
Is this investment justified to address the highest end of the high end base? Does CPQ's return on this investment justify the expense? So far the answer has apparently been yes. But the business practices which generate value from this investment are very different than those of a 'fast follower' like Dell.
I like the comparison of this business to the Alaskan dog-sled races. CPQ is breaking the new snow, Dell is following, always ready to grab an advantage with fresh, fast dogs. |
