<font color=red>Richard, someone who has a different perpective re the Intel/RMBS relationship....it must bite to be a Rambyte:
____________________________________________________________
Intel Drops a Bomb on Rambus
By Jim Seymour
Special to TheStreet.com
7/26/00 6:21 PM ET
Wednesday afternoon, commentary editor Leland Montgomery stirred up a hornet's nest in the RealMoney.com Columnist Conversation hot box. He passed along a link to a ZDNet posting of a Reuters story, which reported Wednesday's 13% drop in the shares of Rambus (RMBS:Nasdaq - news - boards), the high-hopes/high-skepticism outfit that believes it has a new memory architecture that can run a lot faster than today's other choices in high-speed RAM for computers.
Intel (INTC:Nasdaq - news - boards) had just ended its exclusive deal with Rambus to use Rambus-technology memory -- the company is an intellectual-property play that licenses its technology; it doesn't make chips. Intel then rubbed salt in the wound by announcing that, rather than forcing PC manufacturers planning to use its hot new Pentium 4 chips later this year to use Rambus-design-memory chipsets in them, Intel is instead designing a chipset to work with the Pentium 4, which allows the use of more conventional "PC 133" high-speed RAM.
There are lots of nearly hidden subtexts here:
Rambus memory has created serious, publicly acknowledged problems in compatibility for Intel over the past year.
Rambus memory costs about three times as much as conventional high-speed memory.
There have been grave doubts about whether there was enough production capacity in the world to produce enough of the hard-to-manufacture Rambus-design chips in order to allow Intel to release the Pentium 4 CPU this fall on schedule, and for PC makers to get real machines into customers' hands.
Intel has a warrant from 1996 to buy 1 million (now 4 million, thanks to a 4:1 split) shares of Rambus at just $2.50 if it meets a test of 20% of its chipsets going out the door in two consecutive quarters based on the Rambus technology. So Intel is obviously pulled two ways: get the Rambus-design product out there, exercise the warrant, and pick up a few hundred mil more in investment income ... or to hell with the Rambus bucks, full-speed ahead on P4.
Now, the biggie: The only reason to pay the price for Rambus designs is its claimed superior performance (read: speed) when coupled to high-speed CPUs. The extent of that performance edge -- and maybe its existence -- has been challenged widely, which has led to pretty-spectacular fluctuations in Rambus' stock price, from 14 to 135 over the past year. (Rambus was trading around 75 late Wednesday, down about 12% on news of Intel's decision.)
Now, however, Intel itself has brought that performance advantage into question. Earlier this month, Intel posted on its own Web site the results of a comprehensive test of Rambus' memory vs. other, cheaper high-speed memory, using test-bed PCs built around 933 MHz Pentium III CPUs. Both PCs tested had the same amount of memory -- 128 MB, which is becoming the new standard -- and so the test was on level ground, with one using much cheaper "PC 133" memory and one using a Rambus design.
Intel's own tests show the standard PC 133 memory actually outperformed Rambus designs in 11 out of 14 of the benchmark tests -- including all six productivity benchmarks (which measure performance when executing typical business software tasks) and most of the multimedia and 3D benchmarks.
Computer testing always reminds me of the hoary cliche about statistics -- There are lies, there are damned lies, and then there are statistics -- but I'm inclined to believe that while some personal interpretation of test results is always possible, Intel was very unlikely to post bogus results for the world to see on its own Web site.
Now, arising from those messy subtexts, comes Intel's announcement Wednesday that it is ending its exclusivity deal with Rambus, and that it's going to release in the fourth quarter, with the Pentium 4 CPU chip, a supporting chipset that makes it possible to use conventional high-speed RAM with the P4.
Pretty clearly, Intel has decided its business decisions have to be focused on assuring the success of one of its two new landmark chips (the other being the 64-bit Itanium processor, also due in Q4), not on its investment plays.
Good call, Intel.
This is crummy news for the legions of Rambus true believers. And they are legend, believing they sniffed out a great new stock early on, and believing that right down to the wire.
I'm not saying Rambus is dead; far from it. Heck, it has license fees coming in from Double Data Rate DRAM (DDR), a juiced-up version of PC 133 memory that may, in fact, prove to be the choice that kills off "pure" Rambus-design memory.
Toshiba and Hitachi have signed license-fee settlements with Rambus on their DDR designs, in response to Rambus' litigation against many memory-chip makers. (But I should note that many see the Toshiba and Hitachi settlements as examples of corporations eager to get out of the courtroom, not technically savvy decisions affirming any violation of Rambus' intellectual property rights.)
No forecasts of doom. Just a tough, tough moment for Rambus.
--------------------------------------------------------------------------------
Jim Seymour is president of Seymour Group, an information-strategies consulting firm working with corporate clients in the U.S., Europe and Asia, and a longtime columnist for PC Magazine. Under no circumstances does the information in this column represent a recommendation to buy or sell stocks. At time of publication, neither Seymour nor Seymour Group held positions in any securities mentioned in this column, although holdings can change at any time. Seymour does not write about companies that are current or recent consulting clients of Seymour Group. While Seymour cannot provide investment advice or recommendations, he invites your feedback at jseymour@thestreet.com.
-------------------------------------------------------------------------------- |
