Email this article Format for printing Receive via handheld

Rambus Is King of the Hill

Rambus' business model used to depend on the adoption of its RDRAM technology. Now that it has begun to assert its patents on elements of SDRAM and DDR SDRAM, its revenue stream looks far more secure. But it's still a tenuous matter to build your business on patents alone.

By Brian Lund (TMF Tardior)
July 19, 2000

Seems like only three months ago that I wrote that investing in Rambus Inc. (Nasdaq: RMBS), a semiconductor intellectual property company, was a matter of faith. At the time, its market cap stood at $5 billion, about 100 times sales. Well, now its market cap is just shy of $9 billion. Yesterday's third-quarter earnings report lifted trailing 12-month sales to $57 million, so Rambus is now valued at about 160 times sales.

Now I'm thinking it may be worth it.

Royalty revenues are key
Royalty revenues are the key to Rambus' future. The company licenses its Rambus Dynamic Random Access Memory (RDRAM) technology to semiconductor companies and collects royalties based on sales. Royalty rates range from 1% to 2.5%, depending on the license and passage of time. As more original equipment manufactures adopt RDRAM, royalty revenue grows.

Until recently, Rambus' valuation depended entirely on its big brother, Intel (Nasdaq: INTC), designing chipsets around RDRAM. Sure, Rambus gets some revenue from chips sent to Sony Corp. (NYSE: SNE) for use in the PlayStation 2; in fact, almost all of the royalty revenue Rambus currently earns is from sales to Sony. (Nintendo also used Rambus in the Nintendo 64, but those sales are falling.)

RDRAM sales to increase
The big money, however, was in RDRAM sales to computer original equipment manufacturers (OEMs) like Dell (Nasdaq: DELL). Those sales are starting to come, though we won't see the effects until next quarter. Dell recently introduced the first sub-$1,500 PC using RDRAM technology, which begins to allay fears that OEMs wouldn't adopt RDRAM for high-volume products because it is more expensive than standard Synchronous DRAM (SDRAM).

I argued last quarter that large-scale adoption of RDRAM was a very open question. It still is. Prospects appear pretty good, since Intel's latest chip, Willamette, which will be marketed as Pentium IV, should feature RDRAM. It will be optimized to RDRAM, unlike the i820 debacle last fall. Pentium IV is supposed to appear in the second half of this year, bringing revenue into both companies about a year from now.

Royalty payments from Toshiba and Hitachi
Whether that happens or not, however, Rambus now has a new revenue stream: royalty payments from SDRAM and its next generation, double-data rate (DDR) SDRAM. Brian Graney reported on Rambus' recent settlement of its lawsuit with Hitachi (NYSE: HIT) that set a precedent in this area. Rambus now has contracts with Toshiba and Hitachi for royalty payments on SDRAM and DDR. While terms have not been disclosed, they are likely 1-1.25% for SDRAM and 2-2.5% for DDR. Those two companies make up about 5% of the DRAM market. Rambus will begin to realize revenue from these contracts next quarter.

Rambus says that all SDRAM and DDR likely infringe its patents. It is in negotiations with other manufactures to establish royalty payments. If Rambus can enforce its patents, it will take in at least 1% of all DRAM sales, no matter what the type. That's big money. If estimates are realized, Rambus is looking at a minimum of $500 million in annual royalty revenues by 2004, with a top of about $1.25 billion. Since there is no cost of goods involved, a huge majority of that should flow to the bottom line. Wow.

Eating crow
This development has changed my outlook. Before the Hitachi settlement, I didn't think that Rambus would be able to defend its patents. I found it hard to fathom that Rambus could establish that it owned essential parts of the technology behind DRAM. At the very least, it would seem to create an antitrust situation. Well, Rambus apparently does own DRAM. Until someone else challenges that ownership or Rambus' right to enforce it, the company will try to draw royalty payments from each manufacturer. That'll teach me to interpret intellectual property law.

The double-edged sword of patents
Building a business on patents is a two-edged sword, however. Patents can be gotten around. They can become obsolete. They do expire. Rambus is doing more research and development, trying to stay ahead of the curve, and patenting every move along the way, but its business doesn't have a lot of protection in a storm. As a going concern, I'd prefer to have a business built on customer loyalty and manufacturing know-how. Philip Fisher wrote in Common Stocks, Uncommon Profits, "When large companies depend chiefly on patent protection for the maintenance of their profit margin, it is usually more a sign of investment weakness than strength."

Rambus is like a pharmaceutical company in many ways. Last quarter, it had one blockbuster product in clinical trials. Now it's still got that product, but it also has a major one in use. That's a much more tenable position. Investing in Rambus may no longer be simply a matter of faith, but it is still a risky business.

Your Turn:
We've got a lot of Fools that know infinitely more about Rambus than TMF Tardior on the Rambus discussion board. If your interest in the company is piqued, do yourself a favor and join the discussion!

Suggested Links: