continue ...
> During the last 18 months or so, Intel has been working with Rambus to
> make Direct Rambus DRAMs the PC main memory standard in the future.
> In order to make this goal a reality, Intel and Rambus have had to
> convince the leading DRAM suppliers to develop Direct Rambus DRAMs and
> Intel has had to design a system-logic chipset that will interface
> with this new memory standard. With the help of Rambus' ASIC
> interface and Intel's powerful design team, we believe Intel's chipset
> (code named Camino) will be ready for production in the first half of
> 1999. While this chipset will continue to support the current SDRAM
> standard, Intel has reaffirmed its position that Camino will not
> support any alternative DRAM architectures except the current
> generation of SDRAMs and Direct Rambus DRAMs.
>
> We believe Rambus and its key partners have continued to meet their
> important development milestones ever since Intel decided to make
> Direct Rambus DRAMs the next standard for PC main memory applications.
> The Rambus solution addresses the bandwidth gap that has developed as
> MPU speeds increased significantly faster than DRAM performance, and
> we believe Rambus has proven its technology with its Base Rambus
> DRAMs. Rambus' DRAM partners have been shipping 500 MHz parts since
> 1995 and the performance level increased to 600 MHz last year.
>
> Direct Rambus DRAMs will operate at 800 MHz and they will offer peak
> bandwidth of 1.6 gigabytes per second. Rambus has already delivered
> its Rambus ASIC Cell (RAC) interface to be incorporated into Intel's
> Camino system-logic chipset. In addition, Rambus announced on
> February 17 that it delivered the final Direct Rambus DRAM interface
> spec to its 13 DRAM partners. We believe the majority of Rambus' DRAM
> partners have made significant progress in developing their Direct
> Rambus DRAM designs, and the leaders appear to be on schedule to
> sample their first parts in the second or third quarters of this year.
> Based on data provided at Intel's Developer Forum on February 17, it
> appears that Hyundai will sample Direct Rambus DRAMs in the second
> quarter, and Fujitsu, Hitachi, LG Semicon, NEC, Samsung, and Toshiba
> are scheduled to provide samples in the third quarter.
>
> .And RMBS Should Become A More Compelling Investment As We Approach
> The Launch Of Direct Rambus-Based PCs Next Year
>
> We believe Intel will introduce the final iteration of its Pentium II
> product line (which is code-named Katmai) in the second quarter of
> 1999. Katmai incorporates 70 new instructions to become the second
> generation of Intel's MMX instruction set. We believe Katmai will
> debut at 500 MHz, and the new instructions should enhance the 3D
> support of Intel's MPUs. By the time Katmai systems are introduced
> next year, it appears that at least 8 DRAM manufacturers will be in
> volume production of 64M Direct Rambus DRAMs. In order to support
> Intel's Katmai product launch, we expect PC OEMs to begin to purchase
> Direct Rambus DRAMs in the first quarter of 1999.
>
> While we wait for Direct Rambus DRAMs to show up in PC main memory
> applications in the first half of next year, we believe near-term
> earnings expectations remain conservative. On March 9, Matsushita
> announced that it would license Rambus' high-speed memory interface
> technology for use in consumer electronics products. Given the
> position of its Panasonic subsidiary, we believe Matsushita could help
> Rambus successfully penetrate the consumer electronics market over the
> long term. In the near-term, the license with Matsushita gives us
> confidence that Rambus' license revenues will grow sequentially in the
> current quarter, and we believe the March quarter will meet or
> slightly exceed consensus expectations. Rambus is scheduled to report
> its second-quarter results after the market closes on April 14.
>
> Although design activity has remained strong and Rambus has continued
> to develop new market opportunities, the use of Rambus DRAMs in the
> Nintendo64 video game remains the largest consumer of the Rambus
> architecture. However, Rambus DRAMs are currently being used in PC
> graphics and multimedia subsystem applications from Compaq, Dell,
> Gateway 2000, Hewlett-Packard, IBM, Micron Electronics, and NEC. In
> addition, Rambus technology is featured in digital televisions from
> Panasonic and LG Semicon, and Rambus has continued to garner design
> wins in the networking market. Thus far, the sales of ICs using
> Rambus' technology has exceeded $1 billion, and we continue to believe
> that the penetration of the PC main memory market will enable Rambus
> to address the majority of the DRAM market within five years.
>
> Upgrading RMBS To Outperform And Our 12-Month Stock-Price Target Is
> $60
>
> We believe Intel's decision to not support any future DRAM
> architectures for PC main memory applications except Direct Rambus
> DRAMs suggests that Rambus' long-term future is solid and our five-
> year growth rate of 75% remains unchanged. The DRAM market was around
> $20 billion last year, and although DRAM revenues will likely decline
> in 1998 for the third consecutive year, we believe capital spending
> cut backs by the major suppliers will enable the DRAM market to be in
> the midst of a strong cyclical growth phase as the Rambus architecture
> reaches mainstream status. Based on a stronger DRAM market, the
> success of Direct Rambus DRAMs in PC main memory applications, and the
> adoption of Rambus DRAMs in networking, PC graphics, and other digital
> video applications, we believe the company's earnings will reach $5
> per share or more in 2002.
>
> As a result of its unique business model, Rambus has the potential to
> become a very profitable company. The most significant royalty stream
> would be from Direct Rambus DRAMs used for PC main memory in 1999 and
> beyond. Although RMBS remains expensive on the company's current
> earnings power, we believe the stock is undervalued when a present
> value analysis of Rambus' future earnings power is calculated.
> Consequently, we have upgraded our rating on RMBS to Outperform from
> Neutral, and our 12-month stock-price target of $60 is derived by
> applying a 25% discount rate to a P/E multiple of 30 applied to
> calendar 2002 earnings of $5 per share. When we initiated coverage of
> Rambus following the completion of the company's initial public
> offering on May 14, 1997, we used a discount rate of 30%. However,
> given the passage of time and Intel's emphatic position that it will
> not support other DRAM architectures beyond synchronous DRAMs, we
> believe the use of a lower discount rate is appropriate.
|
