Does Rambus (Nasdaq: RMBS) have a sustainable advantage worthy of a Rule Breaker? Sustainable advantage can come in four important ways. Yesterday's column covered business momentum and inept competition. Today's serves up patents and visionary leadership. (And, if you want to know sustainable advantage cold, consider participating in The Motley Fool's Quest for Rule Breakers 2001 online seminar. But hurry, registration closes tomorrow!)
Patents vs. production
To some, Rambus is a company that doesn't make anything. It's just a so-called intellectual property (IP) company. It thinks the big ideas, but licenses its patented technology to others -- computer memory manufacturers such as Samsung and Toshiba -- who roll up their sleeves and get dirty manufacturing, and who pay Rambus a royalty on each product sold to systems companies.
Why is this bad? Some say that an IP company is at the mercy of third-party manufacturers, certain to wait on the sidelines, too timid to invest in the technology until they hear Wagnerian choruses of customer demand. And, even if they ramp up production, third parties are free to be inefficient or inept, while the IP company is helpless.
Au contraire. For Rambus, the third-party manufacturer adoption, quality, and cost issues are over. Rambus' deal with Intel (Nasdaq: INTC) that led to Intel's Pentium 4 (and a new Intel chipset including a Rambus memory controller), and the necessary associated RDRAM and Rambus ASIC cells (RACs -- logic integrated circuits containing Rambus technology) production, has solved the demand-side problem: getting someone to build RDRAM and RACs right and in sufficient quantity to make them affordable (though there's much debate about how affordable).
Rambus' product is not RDRAM or RACs, it's technological innovation. The company's expertise is not selling and marketing, it's research and development. It should not spend its time building and maintaining expensive semiconductor fabrication plants, but engineering, perfecting, and patenting the next-generation Quad Rambus system, doubling the bandwidth of current RDRAM. It's free to spend the next 10 years of its patent protection doing that, letting others make 10 years' worth of investment in RDRAM production infrastructure, forming the base of an enormous value chain, at the same time that original equipment manufacturers from Agilent Technologies (NYSE: A) to EMC (NYSE: EMC) build 10 years' worth of diverse products with Rambus technology embedded, ready for the upgrade. (Can we say "Windows 3.1, Windows 95, Windows 98, Windows 2000?...")
Any discontinuous innovation -- a completely different solution to the bandwidth problem -- will face this formidable value chain and substantial switching costs. A sustainable advantage moat!
Patent defense: big, bad Rambus
Rambus' technology is protected by patents with expiration dates ranging from 2010 to 2019. It licenses those patents to DRAM makers in exchange for royalties on the DRAM and RACs to systems companies. But Rambus wants licensing deals not only for its own RDRAM, but for SDRAM and next-generation DDR SDRAM, both of which Rambus believes infringe on its patents. Six of the top 10 DRAM makers -- including Samsung, the largest, with 21% of the DRAM market -- have agreed to pay Rambus royalties for all memory sold (and higher royalties for DDR SDRAM, RDRAM's putative competitor). Four others have demurred -- three of which are in litigation with Rambus over the issue.
Patents give Rambus the chance to speed the movement to RDRAM and RACs -- by charging higher royalties for DDR SDRAM -- and a war chest to enhance R&D and/or buy what it needs in the future to lead the pack in enhancing bandwidth (data transfer rates) -- or even expand into other profitable businesses. Of course, Rambus' strategy could drive customers to technology outside of Rambus patents, such as embedded DRAM, which combines logic and DRAM on the same chip, if cost and performance are equal.
But patents won't fight competition, help Rambus survive tweening, and give it a possible life as a Rule Maker (our dream for all Rule Breakers). Think Xerox (NYSE: XRX) after its initial breakthrough photocopying patents expired. Rambus needs vision at the top.
1990. Remember the game gear kids used then? My workplace's computers were 286s. The Internet? People were talking about Gophers a few years later. Rambus' vision was to imagine then the need for increased data transfer rates between the CPU and memory long before anyone else, and find a solution. Rambus' 1990 vision was as sci-fi then as the PlayStation 9 is to us in a world of the PlayStation 2.
Then, Rambus made another strategic, visionary decision: not to compete with other semiconductor manufacturers and invest in costly, redundant facilities, but to be free to engineer technological innovation. To do this, it needed a business power to persuade the powers that be to build Rambus technology into their future. Enter Geoff Tate.
A Canadian, Geoff Tate joined Advanced Micro Devices (NYSE: AMD) after receiving his MBA from Harvard. In 10 years, at age 36, he was senior V.P. for microprocessors and logic. Then he quit: "I wanted to swing for the fences and do something really big. I wanted to be involved in something entrepreneurial and make a big impact on the world."
He battled cancer a year later, and that only emboldened him to push Rambus all the way -- to the key deal with Intel that the led to the Pentium 4-dependent RDRAM, the SDRAM and DDR SDRAM licensing strategy, and Rambus-beyond-the-PC. He's successfully led the battles, while his company maintains its 80% engineer staffing goal.
Possible Rule Breaker returns are substantial, especially if Rambus can continue to secure royalties for SDRAM and DDR SDRAM, as well as RDRAM and RACs. Based on rough royalty guides provided in Rambus' latest SEC Form 10-K, Rambus could secure an average royalty of 2.5% for all but a fraction of the projected $60-100 billion DRAM market in three to five years.
Let's be conservative and put the market at $80 billion in five years, for $2.0 billion in royalties. With a rounded 100 million shares outstanding, that's $20 dollars in sales per share. A reasonable price-to-sales ratio for a high-margin IP company would be 20, the same as for another so-called IP company, Qualcomm (Nasdaq: QCOM). That equals a price of $400 a share, 10 times the current $40 share price. Taking the high side of the revenue estimate in three years leads to $500 a share, or 12.5y/3x. Rambus satisfies the 10x/5y criterion easily.
Rambus, in which I'm an investor, has business momentum, patents to bolster its progress and fund its future, flailing competition, and visionary leadership. Or the company could well be a one-trick pony, just a prince of fifth-generation DRAM and nothing more, alienating customers through lawsuits, and driving them outside of Rambus patent protection to different technologies. Considering yesterday's and today's columns, and Brian Lund's analysis of the other RB criteria, what do you think: Is Rambus a Rule Breaker?
A) Rambus is a Rule Breaker, pure and simple.
B) Rambus is no Rule Breaker.
C) No opinion: too techie for me -- too far from "buy what you know."
Tom Jacobs -- TMF Tom9 on the discussion boards.