DRAMeXchange: Samsung Sweeps 40% of Global DRAM Market, Total Revenue Decreases Slightly by 1.9%
According to DRAMeXchange, a research division of TrendForce, total revenue for the global DRAM industry in 2Q11 was approximately US$8.1 billion. Although DRAM ASP (average selling price) increased slightly due to expectations of supply disruption from the Japan earthquake, ProMOS’s decrease in wafer start volume and Powerchip’s increase of non-DRAM products caused overall revenue to decrease slightly, 1.9% from 1Q11.
From the market perspective, as DRAM manufacturers continued to increase output from 30nm/40nm advanced processes, DRAMeXchange predicted that in 2Q11 overall chip production would show a growth of 9% from the previous quarter. However, as DRAM manufacturers undergoing technology migration but were still in the adjustment period characterized by low yield rates, a large volume of low-grade chips entered the spot market. As a result, spot price was not on par with the contract market in 2Q11, falling by more than 10%. According to estimates by DRAMeXchange, while the market had been in a state of oversupply this year, affected by supply chain disruption from the Japan earthquake DDR3 2Gb contract price increased approximately 10% in 2Q11. However, taking into consideration the decrease in spot market price, the temporary rise in contract price had not been beneficial to overall revenue. This upward trend was merely a reflection of PC-OEMs restocking their DRAM inventory, as opposed to an indication of strong end demand.
As in 1Q11, Korean and American DRAM manufacturers continue to release profitable figures, while among Taiwanese DRAM manufacturers, only Rexchip and Winbond show small profits. All other manufacturers are in the red. Looking towards 3Q11, the short-term upward momentum of price brought about by the Japan earthquake has already dissipated, the overall economy remains weak, and end demand shows no sign of recovery. DRAMeXchange expects these factors to exacerbate the state of market oversupply and cause a further decrease in ASP. In addition to concentrating their efforts on technology migration to improve cost structure, DRAM manufacturers also need to adjust their product combinations if they hope to escape this storm of falling DRAM prices in one piece.
Revenue Ranking of Global DRAM Manufacturers
Looking at the revenue ranking for global DRAM manufacturers, Korean makers Samsung and Hynix retained their stronghold in first and second place. Samsung’s 2Q11 revenue saw a 2% increase from the previous quarter, with market share over 40%. Benefitting mainly from a large increase in 35nm products, DRAMeXchange estimated that Samsung’s bit growth should have been nearly 10% for 2Q11. Moreover, product quality was not seriously affected by the migration process, which further aided Samsung’s revenue and market share figures, and was also an indication that decline in spot price had had limited effect on the Korean company. As for Hynix, affected mainly by short-term capacity loss from technology migration and a previous quality issue with a small batch of 44nm products, their revenue and market share had remained roughly the same as it was in the previous quarter. Looking towards 3Q11, Hynix hopes that end client certification of the 38nm process will proceed smoothly as they aspire to begin shipping by the end of 3Q11. Their strategy of increasing the proportion of mobile and server DRAM remains unchanged, and the company is expected to have healthy results even in the face of the falling DRAM price storm.
As Japanese manufacturer Elpida had a higher proportion of products on the spot market, they were more noticeably affected by the decrease in spot price. Fortunately, benefitting from the addition of Powerchip’s output since June, their revenue increased 4.8% from the previous quarter, and market share grew by 1% to 14.4%. However, as Elpida had a higher proportion of spot market shipments than other companies, they were still in the red in terms of profit. As for Micron, while their financial report showed that they are profitable, the company’s revenue and market share decreased by 7.1% and approximately 1% respectively from last quarter. The decline was mainly attributed to a lower than expected yield rate from the 42nm process. Additionally, they were about half a year behind other international manufacturers in the adoption of advanced processes. Looking towards 3Q11, conclusions can only be drawn about Micron’s future market share after observing whether or not the 42nm process yield rate improves and if there is significant progress on 37nm product samples deliveries.
As for Taiwanese DRAM makers, Powerchip Semiconductor Corp., now part of the Elpida camp, has officially become Elpida’s commodity DRAM foundry since June. Powerchip’s output for the end of 2Q11 has already been totaled into Elpida’s sales, leading to a revenue of US$210 million for the Taiwanese manufacturer, a nearly 90% decrease compared to the previous quarter. In 2H11, Powerchip has withdrawn completely from the branded DRAM market, capitalizing on the opportunity to transform into a non-DRAM product OEM. As for Nanya, sales for 2Q11 increased 9.7% from the previous quarter, mainly benefitting from a 17% increase of 42nm process products.
In regard to Winbond, as they had entirely withdrawn from the commodity DRAM market and were focusing on production of more profitable specialty DRAM, their revenue increased 2.1% from last quarter. In the future they will focus on product profitability. ProMOS decreased wafer start volume due to insufficient cash for purchasing materials. As a result of such financial difficulties, ProMOS’ revenue and market share fell 24.5% and 1.3% respectively, and their effect on the DRAM industry will be insignificant. (Figure-1)
Branded DRAM Industry Market Share by Region
Looking at market share by region, Korean manufacturers took 65.7% of the market, while American and Taiwanese DRAM makers had 11.1% and 8.4% market share, respectively. Korean manufacturers’ market share continued to expand, firmly demonstrating the maxim of “big gets bigger”. (Figure-2)