Dr. Pai : Well, yes, the surge of DDR SDRAM prices was mainly contributed to the supply constraint caused by the delayed 110nm migration and DRAM makers' diversification on product mix. Most major makers such as Samsung, Infineon, Hynix and Micron are increasing capacity allocated to NAND flash or CMOS image sensor to meet the rising demand on handsets, DSC and USB drives. Also, some DRAM makers increase capacity allocated to low density SDRAM and foundry services, therefore, reducing the capacity growth for commodity DRAM. As for Nanya , the yield rate on 110nm is great so far since we use the most advanced equipments in both our 200mm fabs and the Inotera fab. Our ramp up schedule of the110nm was delayed for one to two months, due to the late delivery of some equipments, and the extended qualification time on this new equipment.

I expect a pretty tight supply for the second half of 2004, especially for the contract market. The estimated demand for the second half is expected to be higher than the estimated supply according to our market research and the forecast demand from our OEM customers. Though I am positive for Nanya to have a significant bit growth from Q3 to Q4, the total DRAM supply, especially for the contract customers, will still lag behind the demand.

DX: Regarding the 110nm micron migration, many DRAM makers have experience bottlenecks at the migration process. You have mentioned that the delay for Nanya is caused by the prolonged delivery of equipment and the extending of qualification on the new equipment . Can you tell us more ?

Dr. Pai : Both our 200mm fabs and the joint venture, Inotera are using the most advanced equipments, including ArF(193nm) scanners which cost more than US $20 million each. The equipment suppliers couldn't meet our demanding delivery schedule. Using the ArF also requires some new photo-resist, and therefore, extends the learning curve. Both factors slightly delays our schedule on the 110nm volume production. Besides these, we are very optimistic on the ramp up progress. Nanya will have both 8-inch fabs migrated to 110nm in the third quarter. Only 14K in our fab-1 will remain at 0.14um micron technology node for specialty DRAM and foundry service at the end of 2004.

DX: For the brand value enhancement, Nanya has made great progress in building Nanya and Elixir brands for different markets such as contract and spot markets. What is your next step ?

Dr. Pai : Nanya's goal is to provide DRAM products with quality and stable supply to our customers. Nanya brand is for our OEM customers, who take 60% to 70% of our total output, and the Elixir brand is for channel customers in the spot market which takes the remaining. We try to keep the allocation ratio fixed within a range to secure both markets with a stable supply. When the OEM customers' demand turns stronger than our channel customers, what we do is to expedite our capacity ramp up in Inotera rather than choosing not to meet our channel customers' demands.

DX: Nanya has made a bet on DDR SDRAM in 2001. This strategy turned to be a success for Nanya by increasing market share significantly and was one of the only two DRAM makers which made profit in 2002. What is your strategy for DDR2 and projection of market share of DDR2 to the end of 2004 ?

Dr. Pai : As Intel aggressively pushes DDR2 to be the next mainstream memory specification, we believe DDR2 stands a good chance to replace DDR SDRAM soon. With DDR2, the memory subsystem can support the front side bus beyond 1GHz, which will open the road map for system builders and users who demand higher performance. Nanya will be introducing DDR2 with 110nm technology and will be switching to 90nm technology later. Our goal is to have 50% bit growth in 2004 and to double that in 2005, with a major portion of the added capacity going to the higher performance and lower power DDR2 products.

Now, our cost for DDR2 has been controlled to be within 30% over DDR1 SDRAM. The advantages of DDR2 should make it acceptable to the market within some premium over DDR. However, the current pricing and good demand for DDR SDRAM will make the transition to DDR2 very delicate for both OEMs and DRAM makers. Nanya has both the technology and the capacity ready to meet our customers demand once DDR2 demand picks up.

DX: In addition to DDR2, what's the product mix strategy and roadmap for Nanya Technology Corporation ?

Dr. Pai : Nanya is a company focused solely on DRAM solutions, and therefore, we don't plan to enter other market in the next few years. As for DRAM, we believe 256Megabit DDR devices will still be produced and used in high volume. However, 512Megabit products will become mainstream products in the market when the expected crossover on price and cost occurring later this year. Nanya's 512Megabit DDR1 SDRAM, designed by our own technology team, showed very good results on the first cut silicon. It will only takes4 chips of the 512Mb DDR to build a 256MB modules and we are expecting an increase in the demand from notebook PC which requires lower power consumption.

At the beginning of this year, we were expecting 512Mb DDR SDRAM to account for a significant portion of our total DRAM output, with 512M DDR1 and 512Mb DDR2. The portion of DDR2 could be readjusted according to the market demand. Nanya also plans to launch low power SDRAM to the market in the fourth quarter of this year. The low power SDRAM will be 256Megabit devices, lead free, FBGA package or packed in MCP form for used in DSC and camcorder markets. We will be increasing our foundry service to produce 1T Pseudo SRAM up to10% of total capacity.

By the end of 2004, Nanya will have a complete product lines that covers the unbuffered DIMM for desktop PC, SO-DIMM for notebook PC and registered DIMM for servers and workstations. This progress will raise our ASP and profit margin on DRAM products.

DX: Inotera is a joint venture between Infineon Technologies and Nanya. Infineon Technologies has excellent accomplishment in trench technology. However, as DRAM makers are aggressively migrating to advanced technology to reduce the cost, technology constraints seems causing the migration slower and harder. What do you think the constraints will be for trench technology compared to stack technology ?

Dr. Pai : I can not comment on stacked technology since we do not have all the information on this technology.
However, we are very confident with the development in trench technology. We have already qualified the 90nm trench technology, and are working good process in the 70nm trench process. Basically we have in our hands the trech technologies we need for the next 5-8 years. If there's a problem beyond 70nm, I would say, that it probably will be by the equipment supply. We know 193nm scanners is suitable till around 70nm, but are not sure what equipments will be used for 50nm technology at this moment.

DX: What will be the capex plan for Nanya Technology Corporation and Inotera for year 2004 and 2005 ?

Dr. Pai : Our capital expenditure plan is US$260 million for Nanya and US$ 882 million for Inotera for 2004. The 2005 numbers have not been proposed yet.

DX: What's the cost for DDR SDRAM produced by 8-inch fab and 12-inch fab as for Nanya ?

Dr. Pai : To continue providing our customers with quality DRAM products, Nanya has invested over 13% of revenue on research and development. The fully loaded cost for our 0.14um technology in 200mm fab is around $3.6 per chip. We expect the cost to be further reduced with 0.11um mature yield rate in 200mm fabs and the cost for 0.11um developed in 300mm will be even lower.