The company was founded in 1990 with purpose to develop chip-to-chip interface technologies in general and memory systems in particular. I have no idea if Mark Horowitz, Mike Farmwald and other Rambus’ founders had an idea about the ongoing trends in early nineties, but they decided to found a company that would develop standards for the whole industry. Probably their strategy was absolutely correct in some way – they realised that the industry needs faster interfaces and memory devices already in 1990. One thing they had not learnt is that the industry does not like loners.
Early nineties appeared to be a big milestone in development of computers: they combined everything valuable from the eighties and they forced the industry to develop the way it does now. On thing very important there was the fact that proprietary computer systems like IBM’s PS/2, Atari and some others died then. For example, in 1987 IBM made an attempt to recapture the market with its new line of PS/2 personal computers, microchannel architecture and proprietary OS/2 operating system, but eventually failed as Intel, Microsoft and others forced their technologies to become standard without asking licensing fees or enormous prices for them.
Rambus decided to go almost the same way as their unlucky predecessors – they joined JEDEC in late 1991 or early 1992 for the purpose of participating in JEDEC's efforts to develop a first-generation standard for SDRAM, however, tried to promote the technology they had invented – the RDRAM. Since they wanted other firms to license the technology, the JEDEC ejected their offering and adopted traditional DRAM technology that was not covered by any Rambus patents. After that Rambus revised its strategy and broadened their patents on SDRAM technology. Although they received the broadened patents in 1996, they did not file any cases in regards patents infringements till 1999 because Intel decided to adopt RDRAM as their next-generation memory standard. Apparently, Intel, as well as memory makers did not succeeded in successfully launching the first PC intended RDRAM chipset, i820, and it became clear that there were hardly any prospects for RDRAM on the market because of low yields, licensing fess and high prices. After that Rambus started to sue memory makers about the patents infringements of the Los Altos, California-based developer. Eventually, manufacturers of nearly 50% of the worldwide production of DRAM technology have allegedly signed licenses with Rambus, but three patent infringement cases are currently pending. Since RDRAM is almost dead on the PC and its memory market share gradually slides, the only way to survive for the technology company is to receive royalties from the industry players. This approach and strategy resembles a pupil who failed to complete a task, but still wants to have something for his efforts.
Apparently, the approach works well – the company again declares profits for the fourth quarter 2002! For its first fiscal quarter ended December 31, 2002, net income was $5.5 million (22% of revenues), compared to $6.2 million in the same period in 2001 year and $5.9 million in the previous quarter. Earnings per share for the December quarter were 6 cents, compared to 6 cents in the same period last year and 6 cents in the previous quarter. Revenue for the quarter was $25.7 million, up 3% over the same period last year and up 5% from the previous quarter.
According to the report, December quarter results include $24.3 million (or, 94.5% of the revenue!) in royalties, up 12% over the same period last year and up 5% from the previous quarter. Royalties were up sequentially for SDRAM and DDR memories and memory controllers as well as for RDRAM memory and memory controllers. Given that RDRAM market share is below 5% these days, and was about 7% in the fourth quarter last year, we can conclude that Rambus received lion’s share of the revenue from the SDR/DDR DRAM memory makers.
Besides the ongoing legal disputes with Micron, Hynix and Infineon, Rambus is also sued by The U.S. Federal Trade Commission that accuses the Los Altos, California memory firm of unfair competition and alleged destruction of some documents 1998 that concerned Rambus’ participation in JEDEC’s deliberations in early nineties. The FTC wants to find out if Rambus disclosed all their patents during the JEDEC meetings or not. If they had not – the unfair competition will be proved and companies mentioned above will not have to pay fees to Rambus.
What will happen, if Rambus becomes unable to receive any SDR/DDR-related royalties eventually? Of course, there still are a number of companies to pay the Los Altos, California chip-to-chip interface firm for their proprietary techniques like RaSer and Yellowstone, but will it be enough for the company to continue its work on the high-performance interfaces? Currently the company spends $6.5 million per quarter on Research and Development and their main source of revenue does not seem to be the royalties for their own products, but for SDR/DDR memory. Therefore, if Rambus loses the legal disputes, they will eventually either go bankrupt or transform into a very-very small company that will not be able to spend even $6.5 on R&D and will probably develop various concrete solutions on a contract basis. In this case almost all technologies developed by Rambus will become obsolete eventually, as the giants of the industry will develop their own more advanced techniques. With no prospects and chances it may be reasonable for Rambus to become a part a company that will utilise their technologies and develop them further. Though there are a lot of companies that can buy Rambus (current market capitalisation of RMBS is $700.9 million), there are not many companies that may receive any benefits as a result of this acquisition. To name a few, there is Intel, Samsung, Sony, Micron, HP, Cray (and maybe Microsoft to kill Sony’s PlayStation series) as well as some others who need high-performance chip-to-chip interfaces in their products and may be interested in technologies that were invented by Rambus.
To sum up, Rambus may either live forever for royalties from certain DRAM makers, or fall into oblivion in case FTC manages to prove their unfair competition. In case the latter happens, there are two ways for the company that spent a decade to sue everyone in the industry just for the fact that they had refused to adopt technologies developed by Rambus in early nineties. So, what will happen to Rambus? No idea here, but keep in mind: the industry does not like loners and current management board should have learnt the lesson well by now.