【出版时间及名称】：2010年4月台湾内存行业研究报告
        【作者】：摩根斯坦利
        【文件格式】：pdf
        【页数】：41
        【目录或简介】：
DRAM - improving product mix, optimizing fab efficiency,
de-leveraging, and generating free cash flow on capital
discipline. As the cheapest DRAM stock globally, we
see the most upside for Winbond ahead. 2. Near-term
concerns – Nanya and Inotera are facing losses in 1Q10
on 50 nm transition, causing bits decline in 1Q10 on
yield loss. Nanya/Inotera should resume significant
DRAM bits growth in 3Q10 when 50 nm yield improves.
Investors will need to assess DRAM price impact in
3Q10 after firm 2Q10 pricing (better than expected) if
Nanya/Inotera grow bits by ~50% Q/Q. 4. Balance
sheet constraint - Powerchip has to reduce capex to
repay debt. Powerchip benefits near term on firm DRAM
pricing with minimal capex for technology migration.
Powerchip will face risks in 2H10 when Nanya/Inotera
completes 50 nm transitions that pressures DRAM price
again.
What’s New? We upgrade Winbond to OW from EW
with higher TP/EPS on improved profitability.
What’s Next? 1. DRAM price should stay firm in 2Q10
at below US$50 for DDR III and ~US$40 for DDR II 2 GB
module on inventory build for 3Q10 demand and 50 nm
migration constraint by Immersion tool delivery. Spot
price has peaked at ~US$3 for 1 gigabit chip. 2.
Consumer-oriented PC makers are seeing cost
pressure that will likely limit PC DRAM per box at 2 GB
while corporate PCs are adding more DRAM on 64 bit
OS migration. 3. This cycle is unique in that CPU MHz
power is no longer the selling point where DRAM and
storage are the only quantifiable technology in PC box.
Bull case scenario - 1. 2009-11 is similar to 1993-95 on
32 to 64 bit transition (DOS to Windows) and capacity
shutdown during the Financial crisis (Sumitomo epoxy
fab explosion in 1993). DRAM is half way through
recovery as constraint today is held by bottleneck
Immersion tool equipment that should get resolved by
2011 (Atif Malik’s research). Where could we be
wrong? If DRAM price stays relatively firm at ~US$2.5
(mainstream 1 gb) in 2H10, Powerchip could benefit
more than Inotera/Nanya as it should generate more
cash flow on minimal capex while costs at Inotera/Nanya
would rise initially on 50 nm capex due to yield loss.