以下為本週外資研究報告中，關於光寶 2301、閎輝 3311、友訊 2332、合勤 2397、健鼎 3044、聯電 2303、華邦電 2344、與東元 1504等個股研究報告。提供各位建立產業與相關公司的基礎認知，切勿作為個人買賣依據或對外散播。
研究報告的更新方式為….增列於本文最下方！
Lite-On Tech (光寶 2301) - Macquarie Research
Positive momentum
Lite-On Tech (LOT) announced great results. 3Q08 EPS was 30% above our EPS estimate at NT$0.94 (+50% QoQ, -8% YoY, bonus adjusted).
Perlos also substantially beat expectations, its operating margin improved from 2.6% in 2Q08 to 6.7%, compared to our estimate of 3.0%.
A second buyback of 40m shares was announced. In total, LOT will buy back 70m shares or 3% of the total outstanding shares, for employee bonuses.
Our view vs consensus: LOT is one of our top, long picks in Taiwan given its solid balance sheet, healthy cashflows, improving fundamentals, and cheap valuations. Despite macroeconomic headwinds, we are happy to see LOT execute well in a difficult environment and encourage investors to notice the positive changes taking place at this company.
欲閱讀全文者請按 Lite-on Tech 2301 光寶 研究報告
Silitech Technology (閎輝 3311) - Macquarie Research
Better 3Q08, but gloomier 2009
We maintain Neutral rating on Silitech. We lowered our target price from NT$94 (14x 2009E EPS) to NT$72.5 (12x 2009E EPS). Given the weaker industry outlook ahead and the likely EPS decline in 2009, we expect
Silitech’s valuation to be de-rated.
When to revisit? For the long run, we still like Silitech’s leading industry position and superior technology. We think investors might revisit Silitech should there be better 1H09 order visibility and/or product diversification (into light-metal casings) that enables Silitech to gain more tier-one client orders.
欲閱讀全文者請按 Silitech 3311 閎輝 研究報告
D-Link (友訊 2332) - Citi Investment Research
Buy: Better Op Margins in 3Q08; More Conservative 09 Forecast
Our view vs. consensus — While consensus is looking for a sharp decline in 2009E and is worried about weaker margins, we think the company’s exposure to emerging market telco projects and growing presence in US/Europe SMEs will help D-Link cope with the slowdown. We expect D-Link’s unit shipments to be flat in 2009, while sales and earnings are likely to decline by 5-7% yoy in 2009E, driven mainly by blended ASP declines. In addition, margins should also be higher for the company’s SME solutions due to fewer customer rebates.
欲閱讀全文者請按 D-Link 2332 友訊 研究報告 – 花旗 ，另提供 D-Link 2332 友訊 研究報告 – 高盛 予各位一併參考。
ZyXEL Communications Corp. (合勤 2391) - Goldman Sachs Research
Disappointing 3Q result; but visibility higher on carrier exposure
ZyXEL reported 3Q parent results, while consolidated results are not available yet. 3Q earnings of NT$151mn was well below our estimate by 50%. Although 3Q sales were in line with our estimate, operating profit was 47% lower than our estimate. We attribute the lower than expected GM / higher opex ratio to fierce competition in the networking space and ZyXEL’s inefficient opex control. The non-operating loss was roughly in line with our estimate. Forex loss impact due to Euro depreciation was mitigated by the reverse of inventory loss and bad debts provision.
ZyXEL is trading at 13X ‘09E EPS, below its historical trough of 15X. Downside risk for the share price could be limited in the near term, given ZyXEL’s strong cash position (2Q08 cash is 27% of 10/28 market cap) and low valuation.
欲閱讀全文者請按 Zyxel 2391 合勤 研究報告
Tripod Technology Corp (健鼎 3044) - Macquarie Research
Conservative outlook
We reaffirm our Outperform rating, but reduce our TP from NT$103 to NT$60 to reflect earnings estimate reductions and a lower valuation multiple due to weak market sentiment and limited earning growth.
Tripod’s share price declined by 37% and underperformed the Taiex by 11% during the past month due to limited liquidity and the company’s exposure to DRAM and TFT. We believe the stock has been oversold at less than a 6x PER (lower end of PCB peer trading range), despite Tripod’s strong execution, low-cost structure and improving balance sheet. Tripod is our top pick in the PCB space, and our TP NT$60 is derived from its average 2008/09 ROE-g/COE-g with a 30% discount, implying a 2008/09E PER of 8.5x.
欲閱讀全文者請按 Tripod Technology 3044 健鼎 研究報告
UMC (聯電 2303) - BNP Research
New target price of TWD8.80
We lower our target price to TWD8.80 (TWD10.60 earlier), which is based on 0.6x P/BV from the previous 0.7x. Our new valuation reflects: 1) 15% valuation de-rating of foundry manufacturers in the past four
weeks; and 2) UMC’s potential risk from ProMOS investment. We maintain HOLD for UMC. We believe UMC’s ROE will fall further during the downcycle.
Downside risk during downturn
We believe UMC is likely to underperform its 1Q09 outlook. Our 1Q09 revenue forecast for UMC is 8% q-q down. We believe UMC’s ROE could fall further in the 4Q08-1Q09 downcycle. We believe UMC is likely to
see further significant order cuts because of its weaker position in foundry industry (we notice its market share at 65nm process is behind Chartered).
欲閱讀全文者請按 UMC 2303 聯電 研究報告 – BNP ，另提供 UMC 2303 聯電 研究報告 – 美林、UMC 2303 聯電 研究報告 – 德銀 予各位一併參考。
Winbond Electronics (華邦電 2344) - Morgan Stanley Research
Time to Exit Commodity DRAM
Quick Comment: Impact on our views. Winbond may exit commodity DRAM as its technology source
Qimonda will be exiting. Going forward, Winbond aims to de-emphasize commodity DRAM and focus on niche memory. In our view, Winbond should preserve its cash and consider fab divestments to become a leading fabless memory maker to differentiate in Taiwan DRAM. As all Taiwan DRAM makers are trading below replacement costs, there would be more than sufficient quality 12” DRAM fabs in Taiwan for Winbond to leverage on its niche memory designs.
欲閱讀全文者請按 Winbond Electronics 2344 華邦電 研究報告
Teco Co (東元 1504) - Macquarie Research
Taiwan High Speed Rail risk
We provide an update on Teco’s business operations and clarify its Taiwan High Speed Rail (2633 TT, NT$4.9, Non-Rated) issue following China Steel’s (2002 TT, NT$20.6, UP, TP: 27) write-off on its THSR investment.
Currently, Teco is trading at 7.7x 2009E PER and 0.4x P/BV. The cheap valuation makes Teco look interesting to watch; however, we suggest investors wait until the market stabilizes and the THSR write-off risk winds down. We retain Outperform with target price of NT$13.
欲閱讀全文者請按 Teco 1504 東元 研究報告 – MACQ ，另提供 Teco 1504 東元 研究報告 – 高盛 予各位一併參考。
11/27 新增
Tripod Technology Corp (健鼎 3044) - NOMURA Research
Hibernation
Despite its strong cost competitiveness, execution, and market-share gains in HDI and NB PCBs, Tripod is not immune to macro woes, especially when its major segments — DRAM, TFT-LCD and HDD (accounting for a combined 57% of FY08F sales) — are facing severe oversupply/slowdown. Initiating at NEUTRAL.
欲閱讀全文者請按 Tripod 3044 健鼎 研究報告 – NOMURA
12/28 新增
ZyXEL Communications Corp. (合勤 2391) - Goldman Sachs Research
Relatively defensive but macro headwind remains strong, Neutral
We now assume ZyXEL’s 4Q08 sales will decline 10% qoq as our channel checks suggest ZyXEL is experiencing a shipment delay from an Asian telecom company. We have also trimmed our 4Q GM assumption to 30.9% (from 32.3%) due to its worse product mix in the quarter (heavier reliance on carrier business).
欲閱讀全文者請按 Zyxel 2391 合勤 研究報告 – 高盛
D-Link (友訊 2332) - Goldman Sachs Research
Struck by slowing emerging markets and FX, maintain Neutral
Our channel checks show a steady worsening of end demand, even in the Christmas season, and D-Link mgmt now expects Dec sales to continue to trend down. As a result, we now expect a high teens qoq decline in DLink’s 4Q sales and a deteriorating operating margin, and forecast a 5.5% sales decline in 2009.
欲閱讀全文者請按 D-Link 2332 友訊 研究報告 – 高盛
01/20新增
D-Link (友訊 2332) - Citi Research
Buy: Government Projects to Resume After Near-term Shortfall
Lowering target to NT$30 from NT$34 — We are lowering D-Link’s target price to NT$30, equating to 1.0-1.1x book, to reflect weaker sales and earnings outlooks due to project delays in emerging markets and concerns over potential losses in 1H09. But we expect infrastructure projects, less focus on the US and
Europe retail segments, and lowered OPEX should help D-Link’s operating margin recover gradually by 2H09E.
欲閱讀全文者請按 D-Link 2332 友訊 研究報告 – 花旗
01/20新增
Winbond Electronics (華邦電 2344) - Morgan Stanley Research
Time to Pick a Side
Quick Comment – Conclusion: Since the 2000 tech bubble burst, Winbond has been in loss making with the
exception of 2004/06 industry upturn. 2009 semiconductor industry decline will be the worst ever in history, dwarfing the 31% Y/Y decline in 2001 given global recession. Winbond’s 12” fab has the smallest
scale in Taiwan but it’s ironically the busiest – producing all kinds of memories at all geometries – commodity, specialty, mobile, NOR, etc; 90/80/75/65 nm, etc; trench, buried wordline, etc. As fab capacity is now in excess, Winbond should restructure to become a fabless memory maker that creates shareholder value on product design not opportunistic fab return of commodity cycle. In the deepest ever DRAM downturn, we believe management needs to devise the right exit strategy before the situation becomes worse. Management is now expressing openness to 1) DRAM consolidation and 2) fab spin offs to become fabless design house – we view this as a step in the right direction.
欲閱讀全文者請按 WINBOND 2344 華邦電 研究報告
UMC (聯電 2303) - Goldman Sachs Research
Xilinx to switch a foundry partner at 45nm; maintain Sell rating
It is not a surprise to see Xilinx switching foundry supplier (see our report “Foundry competition in 2009 (2): foundry customer’s perspective” dated 12/17/2008). Xilinx is likely the only leading semiconductor company that depends on UMC for manufacturing technology. We believe that Xilinx is cautious on UMC’s long-term commitment and ability to drive yield curve in light of the escalating development cost. According to its 10K, Xilinx sold its remaining UMC holdings in 1Q08. We would watch closely on who
would become the main technology supplier to Xilinx at 32nm. Meanwhile, we note that, five years after inception, Samsung’s foundry business has only three leading edge customers – Qualcomm, Marvell, and Xilinx, due to concerns over its IP protection, low yield, and potential termination of foundry business. We believe that Samsung is unlikely to become a meaningful foundry competitor anytime soon for the above reasons.
欲閱讀全文者請按 UMC 2303.聯電 研究報告 – 高盛
