TrendForce: Arrival of Post-Elpida Era Brings Chance for DRAM Industry Revival

According to DRAMeXchange, a research division of TrendForce, a month after Elpida filed for bankruptcy protection, merger and acquisition rumors began circulating in the DRAM industry. Elpida is attempting to make it through on their own, hoping to rely on financial restructuring to survive. TrendForce indicates, in the past few years the DRAM industry has been facing decreasing profits, and the PC industry has been affected by tablet PC and smartphone cannibalization. The situation is only getting worse, and Elpida shows overreliance on PC DRAM. Meanwhile, Elpida’s operations focus heavily on the spot market, where the impact of economic fluctuations is the most severe. Even if Elpida is able to go up against first-tier manufacturers in terms of process technology, as soon as they are faced with an economic downturn, the maker’s revenue declines and market share begins to fall.



Elpida’s future remains a focus of the DRAM industry – the maker must find a way to solve all of its financial difficulties. The Japanese government’s stance will have key influence in the matter, and Elpida is looking for any viable solution that would enable its survivial. TrendForce indicates, selling Rexchip for cash and turning the Hiroshima fab into an IC design facility would help propel the DRAM industry forward, minimizing the amount of damage to all involved. TrendForce presents three possible outcomes to the Elpida situation:

Scenario 1: Elpida receives funds, debt managed
Current rumors indicate the Japanese government intends to provide funds to save Elpida, under the condition that Japan’s domestic DRAM and NAND flash product lines are consolidated. Furthermore, in order to make inroads into the mobile market, Elpida will lower the proportion of commodity DRAM produced at its Hiroshima fab, increasing mobile DRAM wafer starts instead. However, due to Elpida’s limited NAND flash technology, the maker would still be unable to compete with the Korean manufacturers in the MCP (multi-chip packaging) sector. Therefore, there is a possibility that Toshiba, with second place in NAND flash revenue, will form a technological partnership with Elpida under the guidance of the government. If successful, the makers would be able to target the high-growth tablet PC and smartphone markets with MCP products, enabling them to go head-to-head against the Korean manufacturers.  

Commodity DRAM would be produced mainly by Rexchip, and technology migration to the 25nm process would be accelerated. From a supply and demand perspective however, even if Rexchip handles the commodity DRAM production, the commodity DRAM market will still remain in a state of oversupply –average DRAM price will not increase by much. As for Elpida’s liabilities, the maker may also negotiate for a debt-equity swap or even a partial debt write-off with the banks. It is unlikely Elpida will back out of the DRAM industry completely, as that would provide the least benefit to DRAM chip price.

Scenario 2: Elpida sells fabs, short-term suffering due to technology migration adjustments
If the Japanese government gives Elpida a loan extension, the maker must find a way to survive on its own. After consolidation of the nation’s LCD industry, the Japanese government has set their sights on the domestic semiconductor industry as well – merger discussions are currently underway with Renesas Electronics, Fujitsu, and Panasonic. The government also intends to divide IC design and chip production into separate businesses, with a future venture by INCJ managing IC design. Chip manufacturing may go to Elpida’s Hiroshima fab, which would lower the maker’s commodity DRAM risk.

As for Elpida’s financial difficulties, TrendForce believes it is likely the maker will sell Rexchip or its Hiroshima fab. After the sale, there will be a painful adjustment period due to technology migration when wafer starts will decrease, which may last for six months or so. From the supply and demand perspective, there is a chance the industry will see balanced supply and demand in 2H12 – makers with strong technology migration skills may see profitable commodity DRAM prices.

Scenario 3: Elpida announces bankruptcy, DRAM prices skyrocket
The Japanese Ministry of Economy, Trade and Industry has not yet decided whether it will help Elpida. In the end, if they decide against it and Elpida backs out from the industry completely, a total of 165K wafer starts per month (including Rexchip) will be withdrawn from the market. The global DRAM industry will temporarily lose 12% capacity, and the oversupply situation will become turn to undersupply in the short term. DRAM contract price would have a chance to hit US$2.35, an increase of over 100%. The DRAM spot market, which by nature is prone to price fluctuations, could hope to see a price of US$2.5 due to generally low stock levels.



Additionally, TrendForce believes, if Elpida backs out of the mobile DRAM market, Korean maker Hynix would benefit the most. With both DRAM and NAND flash capacity, Hynix’s product lines are complete, and the maker is second only to Samsung Semiconductors. Hynix, as the supplier with the second-highest revenue figures, and Micron would absorb Elpida’s market share.

Rexchip is undoubtedly the greatest variable affecting future DRAM capacity. Regardless of Elpida’s fate, Korean, U.S., and even Taiwanese DRAM manufacturers are potential candidates for a Rexchip merger or acquisition. If Rexchip changes hands, it will take a minimum of six months to advance Elpida’s current technology to match that of Rexchip’s new parent company; capacity utilization rate would decrease and yield rate would require an adjustment period as well. TrendForce estimates PC DRAM total yearly capacity would see a decrease of 225K if the sale occurred, which would significantly reduce DRAM market oversupply and completely balance the 2H12 sufficiency ratio, currently forecasted at over 10%. Add to that the possibility that manufacturers may increase inventory levels or double book orders to guard against a severe price rebound, as well as stronger demand due to seasonal sales in 2H12, there is a significant chance PC DRAM price will recover considerably in the second half of the year.