Please refer to each numbered footnote for a more detailed explanation of my reasoning for these figures.

Notes:

1. Memory market projections are notoriously wrong. The memory market was $35 billion in 2000, nearly $12 billion in 2001, and is being projected to be over $18 billion in 2002. The fourth quarter alone for this year is projected to be over $6 billion. Last year it was a disaster for the memory manufacturers, as they lost billions because of selling memory below cost, Micron Technology lost $959 million before an income tax benefit, Hynix lost $3.9 billion, Infineon lost over $1 billion before an income tax benefit, only Samsung was profitable. With a memory market of $18 billion this year, Micron, Hynix, and Infineon are still losing a lot of money, this obviously cannot continue. The declining costs to manufacture memory will not make the manufacturers profitable with an $18 billion dollar market, because they will be shipping more memory units for that $18 billion dollars, due to the memory bit growth rate which has been increasing over 60% per year in recent years. My estimate for 2004 and 2005 is based upon the assumption that the surviving memory manufacturers will have to sell memory at a profit in order to continue surviving. There is a good chance that Hynix won’t be around as a memory manufacturer by 2004.

2. The SDRAM market share is going down as it is being replaced by DDR
3. The RDRAM market share this year was expected to be about 8% of the memory market in dollars. It will be less in units because RDRAM memory is currently more expensive than SDRAM and DDR. I am projecting RDRAM market share to increase because of a couple of factors. RDRAM is decreasing in price while DDR has to increase for the memory manufacturers to become profitable. As RDRAM approaches price parity with DDR, the demand for RDRAM should increase. According to projections by research companies, more memory will be used in communications and consumer devices than in PCs by 2005. RDRAM has a large advantage over DDR in these applications because of granularity, and this advantage will increase as higher bit density chips replace the lower density chips. A single RDRAM chip can be used whereas it requires at least four DDR chips in an application.
4. The DDR market share is growing very fast and is displacing slower SDRAM.
5. It is estimated that over 200 million PCs will ship in 2005, and at a price of $30 per chipset for memory controllers, that would be a $6 billion market for controllers for PCs alone, not including consumer devices, game machines, and communications. I believe my estimate is not too large.
6. The non-Intel share is broken out assuming Intel has 60% of the market for PCs, consumer devices, communications, etc. That leaves 40% that Rambus could get royalty on, because Intel is already paying a flat $40 million per year to Rambus.
7. I got the number of serdes connections from statements made by Rambus. Rambus stated that there probably would be over one billion connections in 2005. I am arbitrarily splitting that number, some in 2004 and more in 2005.
8. Rambus’ share of the serdes market is based upon several considerations including the fact that Intel is one of five companies Rambus announced, who have already signed serdes licensing agreements for Rambus serdes IP. The Intel serdes agreement will provide income to Rambus over and above the flat $40 million dollar per year Intel agreement on Rambus’ memory IP. I feel Intel will be responsible for Rambus’ gaining a significant serdes market share. Also the fact that Intel is using Rambus serdes IP, should help Rambus sign additional serdes licensing agreements with many other companies. Rambus has also announced the fastest single link serdes speed of 6.4 Gbs, which is called RaSer V, and Rambus has 10 Gbs by 2003 on their serdes roadmap. Rambus’ solution of just selling IP to incorporate into their clients’ chips, has a much lower cost, uses less power and board real estate than competing discrete serdes chip solutions.
9. The royalty of $2 per connection is a figure I was given at an Intel Developer’s Forum. I was told that the slow speed links (3.2GHZ) would be around $2 per connection, while higher speed links would be more, the higher the speed, the higher the cost.
10. The income for Yellowstone is based upon Rambus’ last conference call in which they said that Yellowstone and serdes would represent about 10% of their income in the next twelve months (probably from contracts). When asked, how much for Yellowstone and how much for serdes, they answered about fifty fifty. Ten percent of a years income would be about $9 million, so Yellowstone would be about $4.5 million by the middle of 2003, so I think $8 million is a reasonable figure for 2004 and a doubling of that by the end of 2005.
11. The flat $40 million dollars per year from Intel is from a signed $200 million contract Rambus has with Intel, giving licensing rights to Rambus’ memory patents to Intel for five years.
12. Rambus’ last quarter total expenses were $16.1 million, including $2.4 million in legal fees (about $64.4 million per year). I believe expenses of $100 million in 2004 and $110 million in 2005 are reasonable estimates.

This projection contains forward-looking statements. These statements are based on current expectations, estimates and projections about the Company's industry, my beliefs, and certain assumptions made by me. You can identify these and other forward-looking statements by the use of works such as "may", "will", "should," "expects," "plans," "anticipates," "believes," "estimates," "predicts," "intends," "potential," "continue" or the negative of such terms, or other comparable terminology. Forward-looking statements also include the assumptions underlying or relating to the foregoing statements. Actual results could differ materially from those anticipated in these forward-looking statements as a result of various factors, including those identified in Rambus’ recent filings with the Securities and Exchange Commission, including its recently filed Form 10-Q, and also including the uncertainty of new technologies; and the uncertainty regarding the technical and market demands for such technologies. All forward-looking statements included in this projection are based on information available to me on the date hereof. Hager Technologies assumes no obligation to update any forward-looking statements.