Take 2 for PC memory

Samsung, Micron Technology, Elpida Memory and other chipmakers are cranking up the production of DDR2 SDRAM, a faster version of computer memory known as double data rate synchronous dynamic RAM, that will start to become available in the next few weeks.

DDR2 will be able to handle data at a substantially faster rate than products with DDR, the most common type of PC memory, while using less energy.

DDR2 will likely sell for a healthy premium for nearly two years, giving manufacturers time to realize profits in a tech market plagued by a vicious boom-and-bust cycle. Memory prices have been known to plunge by 75 percent or surge three-fold in just months.

So far, 2004 appears to be headed for the surge category, which could be bad news for consumers, who have become accustomed in recent years to steadily falling prices.

According to research company Gartner, prices will rise by 23.1 percent during the year, while the amount of memory cells churned out will grow by 52 percent to 164 billion megabytes. Overall, revenue for the industry is expected to grow from $17.5 billion in 2003 to $25.3 billion in 2004.

DDR2 will help drive the gains, with the new memory chips accounting for 22 percent of the memory output in the fourth quarter and 50 percent of the output at the end of 2005, according to Gartner.

In the past month, prices for the older DDR technology have soared: 256-megabit DDR chips went from $4 in the first week of March to $6 this week. Part of the reason is that PC makers have increased their orders to insulate themselves from fears of shortages later this year, said Nam Hyung Kim, principal analyst at iSuppli. (The prices reflect the cost of individual chips, which are classified by megabits and sold to manufacturers. Memory modules, which are classified by size in megabytes and what consumers buy, track chip prices. DDR modules of 256MB currently cost about $70 to $80.)

"All of the major DRAM suppliers are making money at the moment," Kim said. "The DRAM market has gone crazy."

The shift to DDR2, however, will coincide with a manufacturing overhaul that will require producers to smooth the kinks out of new factories, manufacturing processes, and changes to packaging and chip design. A slip anywhere along the line could cause manufacturing yields--and potential profits--to evaporate.

In addition, the market has been notoriously hard to forecast, as demand fluctuations often clash with ebbs and flows in supply.

For now, though, industry executives are remaining cautiously upbeat.

"We're not wrapping dollar bills around memory any more. The price has been stable," said Dan Donabedian, president of Elpida's U.S. subsidiary. "But we can't (afford to) lose money on this one."

What is DDR2?
DDR stands for double data rate, and it has become the standard PC memory over the past couple years. DDR now tops out at 400MHz; DDR2 will debut at 533MHz and climb to 667MHz before the end of the year. Higher speeds lead to greater data throughput, a key driver of PC performance.

DDR2 is also easy on energy. Statistics from Samsung project that DDR2 running at 533MHz uses 65 percent less power than 400MHz DDR does. While this will help conserve notebook batteries, it will also cut back on heat dissipation in desktops and servers, an increasing problem, as processor speeds climb.

The exact performance boost is difficult to quantify, "but suffice it to say that as DDR2 speeds increase from 533MHz to 800MHz, the performance difference (is) notable," said Richard Gordon, research vice president for emerging technologies and semiconductors at Gartner.

The ability to integrate DDR2 into computers will come after Intel and others release chipsets that can connect to it. Grantsdale and other early DDR2 chipsets will likely appear in May, with PCs that use the chips appearing later in the year.

"The server market is going to come up first, followed by the desktop, although one large customer is ramping up desktops before servers," Donabedian said.

The fab five
The difficult part for manufacturers will lie in digesting all the back-end changes in their fabrication facilities, or fabs. Jim Elliot, senior product marketing manager of DRAM at Samsung, has identified five different elements, an unusually high number.

Most large manufacturers will open up new 300-millimeter manufacturing facilities to produce DDR2. These same manufacturers will also shift from making chips on the 130-nanometer manufacturing process to 110 nanometers and below. Typically, opening up new factories and shifting to a new process is a difficult endeavor for semiconductor manufacturers, because they have to wrestle with unpredictable or unknown problems. Mistakes can subsequently lead to pauses in production.

Individual chips will also be able to hold 512 megabits of memory, twice as much as current chips do, a shift in density that occurs every few years and can be tricky to manage.

And DDR2 chips will have to be inserted in new types of packages. Not only does this introduce another element of risk, but manufacturers will also be required to pay royalties.

Finally, DDR2 is fundamentally different from DDR in its design, introducing a greater chance for error. If anything goes wrong, a manufacturer could end up with a higher-than-acceptable number of chips from each wafer that doesn't work.

"There is a difference in this transition," Donabedian said. "There is a lot to bite off."

Memory holes
More potholes await PC makers. Because of a packaging change, DDR2 won't fit on old motherboards. "The layout is different," said Elliot, a fact that's prompting hardware engineers to tinker with their designs.

Besides navigating the manufacturing and design issues, memory makers will also have to contend with the rapid and often seemingly illogical shifts in prices for which the memory market is known.

Right now, for example, 128-megabit SDRAM chips, made with an older version of memory that doesn't provide the same performance as DDR does, sell for more than the equivalent DDR chips. Why? So much manufacturing has shifted to DDR that SDRAM has become somewhat more scarce.

If most manufacturers execute the transitions flawlessly, which would increase supply, DDR2 prices could plunge prematurely. During various points in the recent past, memory makers had to sell fresh products for less than they cost to make because of oversupply.

Conversely, widespread problems could lead to high prices and a lack of demand from PC makers. The current high prices of DDR could also postpone adoption of DDR2.

"Unless demand forecasting is spot-on (and it never is), it is highly likely that we will see severe product mix issues," Gordon said.

The time to enjoy any price premiums will also be short. Currently, 512-megabit DDR2 chips sell for 84 percent more than 256-megabit DDR. The premium will drop to 24 percent by the end of the year and then disappear entirely by the end of 2005.

To ameliorate these factors, Samsung has instituted a program it calls common-die manufacturing, which allows the company to postpone the decision on what kind of chips to make. Generally, it takes about three months to go from a blank wafer to an individual memory chip.

Samsung and others have figured out how to delay the decision on what to use a wafer for--DDR or DDR2--until the final four weeks of the process, Elliot said. In an industry in which customer contracts are renegotiated every two weeks, postponing the decision will let Samsung alter its output to better match demand.

Manufacturers will also heavily target the most profitable markets. "To succeed in DDR2, you need lots of sales into the server market," Elliot said.