Semiconductor Equipment . . . Credence Systems said it offered $150 million of subordinated notes to qualified institutional investors that are convertible into common stock. The notes will be due in 2008, pay 1.5 percent interest per year and be convertible into stock at a rate of $11.31. The company intends to use the proceeds from the sale for general corporate purposes, including working capital and potential acquisitions.
Intel’s roadmap decision seen as positive for Cymer. Adams Harkness believes that Intel decision to drop planned adoption of 157nn lithography as a positive for CYMI, as it allows co to drop R&D efforts on its F2 products and decreases the likelihood of a competitor taking significant share with a new 157nn product.
Intel has informed suppliers of lithography machines that it will not buy the next generation of tools but instead hopes to extend the life of its current generation of 193 nanometre tools until it can jump straight into the successor of 157 nanometre tools, called Extreme Ultraviolet (EUV) tools. ASML Holding said she understood Intel had adjusted its " roadmap" and wouldn't purchase 157-nanometer machines from ASML, as had been previously expected. The value of the order wasn't disclosed. Intel decided not to adopt the 157-nanometer machines and won't be placing an order with another supplier, she said. [Briefing note: According to analysts ASML's top 5 customers in 2003 are Samsung (20% of revenues), IBM (11%), TSMC (8%, likely 2nd half loaded), Intel (7%), and Micron (7%)].
Semiconductors . . . Semiconductors are in a tepid cyclical recovery, with 2003 revenues estimated to grow at 5%. Companies still have no pricing power as overcapacity persists. The good news is that the worst ASP decline is over, and 2004 looks significantly better than 2003 from a supply perspective. Long-term growth rates of the semiconductor industry have slowed to 7%-8%, due to lower secular growth rates for the traditional demand drivers such as PCs and declining prices in almost every end-application. In addition, overall IT spending has not picked up, new growth drivers do not constitute a significant portion of demand, and valuations are generally not inexpensive. Early in 3rd quarter might be the right time to overweight the industry and trade the PC-centric semiconductor stocks Investors should wait for 2nd quarter PC weakness to be priced in (2Q motherboard shipments are tracking below normal seasonal declines). Look for leading indicators of PC builds and sellthrough to gauge the magnitude and timing of 2nd half seasonality. Over the next six to 12 months, pick stocks that do not depend on explosive growth in end markets. Look for companies with market share gains, leadership and competitiveness that can be sustained through the current prolonged downturn, and successful business model transformations that could lead to a higher valuation or a re-rating of the valuation multiple.
Merrill Lynch analysis of semi group suggests that several stocks, including market favorites such as Texas Instruments, appear expensive. Thinks Infineon, STMicro and Integrated Silicon are more reasonably valued.
Marvell price target raised to $35 at ThinkEquity because of better visibility. ThinkEquity was pounding the table on stock ahead of its release, calling MRVL its ''best mid-cap stock going into its earnings report'. The firm is raising its price target to $35 from $30. According to ThinkEquity, visibility improved significantly, prompting MRVL to bump 2004 revenue guidance to $760-$790 million from the $710-740 million range (consensus $724 million).
Bear Stearns initiated AMD with an Underperform rating. The 2nd quarter 2003 weakness is not priced into the stock. The downside risk for AMD’s potential 2nd quarter revenue shortfall outweighs any potential medium-term upside gain to be generated by its new products – Athlon-64, the launch of which is still four months away. Checks with top-tier Taiwanese motherboard makers imply that PC end demand is deteriorating through 2nd quarter 2003. End-demand weakness is more prevalent in the clone markets. Since microprocessors account for 65% of AMD’s revenue, and >50% of AMD’s microprocessors are for the clone market, we expect AMD’s 2nd quarter revenue to decline 8% Quarter over Quarter (vs. company guidance of a flat quarter or Quarter over Quarter sales improvement). AMD may not be able to grow its flash memory business in 2nd quarter because: 1) Samsung is expected to increase NOR flash supply significantly through 2003; and 2) Intel may cut prices for it key flash memory customers to regain market share in 2nd quarter. Therefore, AMD’s expectation for growth in its flash memory business in 2nd quarter may also be at risk. Expect its flash revenue to decline 7% sequentially. While AMD begins a new product lifecycle this year by launching its 64-bit microprocessors, do not expect the new products to have much material impact on the company until 2004. AMD’s near-term outlook is fraught with uncertainties and execution risk is a also major concern. Moreover, estimate AMD will only turn profitable in 3rd quarter 2004. In wafer processing technology, AMD lags behind Intel; it plans to commence its 0.09µm production in 2004. AMD’s partnership with IBM on 0.065µm and 0.045µm technologies should help AMD to catch up in the process migration to better compete with Intel. It is our belief that AMD can become more competitive if it goes fabless because it can allocate more resources to R&D and new product introductions without the burden of large capex being required to upgrade its own production facilities. From a financial perspective, AMD had about $1 billion in cash and cash equivalents and short-term investments at the end of 2002. It has about $290 million and $280 million of contractual cash obligations plus interest payments in 2003 and 2004 respectively. If you include the $650 million capex for 2003, AMD would require more than $900 million in cash this year. Since estimates are that AMD will generate $382 million and $817 million in cash from its operations in 2003 and 2004, respectively, the company’s cash position will not present a big problem in the coming two years.
Bear Stearns initiated coverage of ATI with a Peer Perform rating. After advancing ahead of NVIDIA to gain performance leadership late last year, ATI has shown consistent execution and delivery with its most recent round of product launches. ATI's fundamentals are sound and we would use any dips in the stock price as an opportunity to overweight the shares. In addition to sustaining its performance leadership, ATI is well positioned to benefit from growth in the fastergrowing notebook segment of the PC market, given its dominant share; is likely to make a push in the desktop integrated chipset market this year; and its consumer business should contribute to earnings growth from 4th quarter. There is also promising potential for new game console contracts. Due to ATI's new distribution strategy and increased add-in-board partnerships, the company has recently increased its penetration in the clone PC market -- clone sales increased 50% and 25% Quarter over Quarter in 1st quarter 2003 and 2nd quarter 2003
respectively. The most significant challenge ATI faces in 2nd half 2003 is to make its next-generation high-end chip available by the Fall to compete effectively with NVIDIA's NV35 chip. Although analysts like ATI's fundamentals and expect new product introductions to be catalysts for stock price appreciation, caution stems from the challenges it faces in the near-term, including: 1) weak end-markets; 2) possible share loss in the notebook integrated segment, 3) possibility of NVIDIA gaining share near term with its competitive entry-level products, and 4) weak GameCube sales.
UBS Warburg analyst Thomas Thornhill downgraded Micron Technology to "reduce" from "neutral." Thornhill said lower long-term average selling price trends will offset higher shipments and lower costs. He also sees no near-term "serious" catalysts for pricing improvements.
Marvell reiterated Reduce at UBS, saying that little information was provided to dispel firm's concerns regarding the release of competing Intel, Broadcom and Realtek Gigabit Ethernet products in the 2nd-half of the year; and although industry indications are healthy. The firm believes the strongest growth for Marvell in HDD may be behind the company. Price target goes to $17 from $15.
Intersil target raised to $25 from $23 at UBS due to ongoing checks and public announcements that suggest an improving likelihood of upside to firm's model stemming from power management product opportunities.
The Micron downgrade this morning by UBS Europe to a sell reflects lagging data from only one broker on the DRAM markets (as opposed to multiple sources), which suggests pricing is coming down with increased unit volumes. On the basis of this data point, the analyst suggests this development augurs negatively for Micron in light of its DRAM business and downgraded the stock. However, a trading desk is out with data from multiple sources in the DRAM markets which shows spot DRAM pricing increasing, not decreasing with majors Samsung, MU and Infineon witnessing increases in spot USA pricing. In addition, Taiwan and Hong Kong also witnessed increases in spot IC pricing overnight for both majors and non-majors with speculation of brokers holding inventory in anticipation of prices rising. The sentiment suggests as prices continue to rise and there continues to be a lack of supply of non-majors low priced DRAM in the channel, it will also help increase the price of majors as well. The firm is therefore positive on DRAM majors such as Samsung, MU and IFX and would buy on any weakness in light of this downgrade to MU.
Investment positives for Intel include new products, specification upgrades, process migration and market shift to notebooks help ease margin pressure. Although there is pressure on profit margins due to competition and a sluggish PC market in 2003, expect Intel’s gross margin to stay near the 50% level this year following the launch of a new Pentium 4 in April 2003 that should boost overall PC performance because it supports 800MHz Front Side Bus (FSB) and dual-channel DDR400. Intel also plans to launch the next-generation Pentium 4 (Prescott) in the second half of the year. Prescott will be manufactured on 0.09µm process and is able to scale up to 5GHz. Expect this process migration and specification upgrade to ease margin pressure. A further contributor should be the market shift to notebook PCs from desktops; the ASP of desktop processors is roughly two-thirds the ASP of mobile processors, and mobile processors account for 18% of Intel’s microprocessor shipments.
Intel has launched a new handset chip with integrated microprocessor core, a DSP baseband, flash memory and SRAM targeting the handset market. The Manitoba chip has already won support from eight handset companies in Asia. Although do not expect the new product to have a significant near-term impact on Intel’s revenue, estimate the new chip could bring in an additional $650 million in revenue if it (and its successors) can capture about 10% of the handset DSP processor market in 2005.
Standard Intel Architecture Servers (SIAS) are mostly entry-level servers, and server processors account for about 8% of Intel’s total revenue. According to Gartner Dataquest, Intel-based server revenue will exceed RISC-based server revenue for the first time in 2003, reaching about $20 billion. This is mainly because of a trend toward distributed computing whereby server workloads are shared across clusters of smaller and cheaper servers, lowering the total cost of ownership compared with high-end servers. Server processors have 4.0-4.5 times higher ASPs and, therefore, higher margins than desktop processors. Although expect Intel to face some competition from AMD’s Opteron processor later in 2003, believe Intel should be able to maintain the lion’s share (~90%) of the x86 server processor market.
Centrino is an example of Intel’s shift in branding strategy. Instead of using the brand for one product, the Centrino brand consists of three Intel products (mobile processor, chipset and wireless LAN module). It covers not only PC-related products but also helps Intel to cross-sell non-PC wireless LAN modules or other communications-related products through its microprocessors. Although customers can buy individual products from Intel, they can use the Centrino brand and Intel’s marketing dollars only if they buy all three products from Intel. This is just the beginning of a new branding strategy from Intel. Expect to see more of this in future as the company leverages its strength in PC-related products and diversifies into other product areas.
Intel has the best cash position in the semiconductor industry, with $12 billion in cash and cash equivalents at the end of first-quarter 2003. It has the most advanced manufacturing technologies and one of the strongest R&D teams within the industry. Intel also has the best-known brand within the semiconductor industry and the PC world because of its excellence in marketing. It is best prepared for any PC upturn.
Investment risks are out there as well.
Intel is well prepared for an upturn, but no compelling driver for growth. Intel has launched numerous initiatives in various product lines such as Centrino and Manitoba, positioning it well for an upturn. However, no analysts can identify a compelling driver of revenue and earnings growth like Pentium or the Internet in the 1990s. As for the overall PC market, expects PC shipments to grow only 6% this year versus a CAGR of 18% in 1995-2000. Expect Intel to grow its revenue by 2.6% in 2003. Although we expect 2004 revenue and earnings to grow 13.1% and 37.6% respectively, if the corporate PC replacement cycle is further delayed into 2005, or if there is strong takeup of AMD’s Hammer processor in the server and desktop markets, Intel’s 2004 growth forecast will be at risk. While Intel has tried to diversify into non-PC segments, so far PC processors are still a large part of its business and convergence with communications products is expected to be a long and slow process.
AMD lags behind Intel in its new product launch and manufacturing process migration to smaller geometry. However, following AMD’s recent partnership tie-up with IBM on 0.065µm and 0.045µm process development, the two companies will become a formidable force in manufacturing advanced microprocessors and will compete more effectively with Intel. AMD is likely to close the gap with Intel in the migration to advanced process technology.
AMD launched its Opteron processor for the server market in April 2003 and plans to launch the Athlon-64 desktop processor in September. IBM is the first major server supplier to announce plans to launch products based on AMD’s Opteron server processor later in 2003. AMD’s PC processor market share increased to 20.2% in 2001 from 16.7% in 2000 following the launch of the Athlon processor in second-half 1999. Expect Intel’s microprocessor market share to decline slightly to 80.9% in 2004 from 82.3% in 2003 and 83.7% in 2002.
Samsung plans to ship 60 million 128Mb-equivalent units of NOR flash memory this year, up sharply from only 3 million units in 2002, following the conversion of its DRAM capacity to NOR flash capacity. This is equivalent to bit shipments of 7.5Gb in 2003 and accounts for an estimated 13% of total bit supply. The surge in NOR supply is likely to put pressure on NOR flash pricing this year and Samsung will likely gain share from the world’s leading NOR flash memory suppliers, including Intel.
Although Intel has tried hard to diversify into other non-PC markets, such as Manitoba for handsets, these products are still at an early stage of development and may not be successful in future. Their market acceptance will be influenced by competitors’ strategies and Intel’s product performance.
NVIDIA has clearly lost performance leadership and market share to ATI in the high-end desktop GPU market recently. However, although NVIDIA is playing catch-up in the high-end enthusiast segment, the company’s new entry-level products for the desktop and notebook markets are competitive relative to ATI’s products in the value segment, which should help offset some of the market share loss in the high end. NVIDIA has the opportunity to regain performance leadership if ATI fails to bring its next-generation high-end chip to the market in the fall, when NVIDIA launches the NV35. Initial reviews of the NV35 have been mixed, with some reviews indicating that the balance of power has shifted to NVIDIA, and others indicating that the ATI’s Radeon 9800 Pro still has a slight edge over the NV35 in some applications. Analysts are more optimistic about NVIDIA’s chipset business: After a slow start with the first version of nForce, NVIDIA has seen stellar growth momentum with its nForce2 chipset, the company is well-positioned to gain share from VIA and SIS in AMD-based chipsets. Although it does not have an Intel bus license, the AMD-based chipset market is providing a significant growth opportunity for now, where the company had an 11% market share in first-quarter 2003. NVIDIA’s strategy of producing a non-integrated version of the nForce2 chipset, which is increasing its total available market to the entire AMD desktop chipset market.
Analysts are not more positive on the stock for the following reasons: 1) the stock is not inexpensive, given its current valuation at 32x forward 12 months earnings; 2) ATI is a much stronger competitor to NVIDIA than it used to be, and are not too optimistic about the likelihood of NVIDIA gaining back performance leadership soon, given ATI’s consistent execution and delivery in the past nine months. Concerned that NVIDIA may not be able to catch up with ATI in its next-generation high-end chip, as the NV40 is not likely to be launched until early 2004, whereas ATI has indicated its next-generation R400 could be available by fall 2003. This could lead to further market share loss to ATI; 3) NVIDIA has limited room for share gain in the discrete desktop segment given its two-thirds share; 4) One of the significant longer-term concerns is NVIDIA’s lack of an Intel bus license for chipsets, limiting its market opportunity to the smaller AMD-based market; and 5) the uncertainty surrounding NVIDIA winning a contract for the next-generation console from Microsoft.
Analysts forecast that NVIDIA’s revenues will decline 3.0% in fiscal 2004 (January year), due entirely to a 56% decline in Xbox revenues. In the most recent reported quarter ended April 2003, NVIDIA’s Xbox revenues declined by a sharp 80% year over year, as Microsoft is in the process of clearing excess inventory after weaker-than-expected holiday sales in 2002. For fiscal 2005, forecast that revenues will improve by 8.8%, and that earnings will grow at a faster rate, at 11% and 25% in fiscal 2004 and 2005, respectively, project that margins will gradually improve. Margins should improve as a result of a decline in Xbox sales as a percentage of total revenues, yield and cost improvements on 0.13-micron production, and improvement in the product mix. NVIDIA is trading at a trailing P/E ratio of 37x, above its historical median of 32x, and at 32x our forward-12-months EPS estimate. But historical multiples are an inappropriate comparison, as the company went public only in January 1999 and most of the data are from the bubble years 1999-2000. Mid-2004 fair value estimate is $18.50 based on a P/E multiple of 20x our mid-cycle earnings estimate, which is the sum of earnings estimates in second-half calendar 2004 and first-half 2005. 20x multiple is based on ATI’s pre-bubble historical median from the 1995-98 period. In the near term, for second-quarter 2003, anlaysts are forecasting an 11% increase in revenues, at the low end of the company’s recent guidance of a 12%-18% quarter-over-quarter revenue increase, based on our checks with motherboard suppliers, which indicate that second-quarter shipments are trending below normal seasonality. NVIDIA’s inventory level also was higher than normal at the end of first-quarter 2003, which makes us slightly more cautious. The decline in Xbox sales is expected to largely reverse in second-quarter 2003, which implies guidance of a 5%-10% quarter-over-quarter increase in the core PC business. Note that because of the expected decline in margin, only forecasting that EPS will increase by a penny, to $0.13.
NVIDIA has gradually gained share of the notebook space since it entered the market two and a half years ago, in November 2000. Over a period of three quarters, NVIDIA has more than doubled its share of the notebook discrete segment, from 15% in second-quarter 2002, to 31% in first-quarter 2003. NVIDIA continues to roll out competitive notebook products, including the recently launched GeForceFX Go5600 (NV31M) and GeForceFX Go5200 (NV34M). While at the same time ATI also came out with a DX9 notebook product, the Mobility Radeon 9600, in the lower-price segment, ATI’s Mobility Radeon 9200 will have a struggle competing against NVIDIA’s NV34M, as it is based on an old core (Radeon 8500) and does not have DirectX 9 support. NVIDIA is the first-to-market with a DX9 mobile GPU for the entry-level segment, and we believe there could be a further market share shift to NVIDIA, until ATI’s more competitive performance mainstream product moves down to the value segment, when it could recoup some share of notebook graphics.
NVIDIA is the undisputed leader in the desktop GPU market, with a consistent two-thirds market share of the segment for the past several quarters. The large size of this segment (38% of the total graphics market) also makes NVIDIA the largest overall graphics chip supplier; its overall market share was 31% in first-quarter 2003. Due to its historical performance lead, the company continues to have a strong brand, an important factor in the white-box market where system builders use the individual components as a way to market their own system-level products. The recent rollout of the GeForceFX product line, while not having the highest-performing enthusiast product, is very competitive. In particular, the NV34 (GeForceFX 5200), which features “DX9 for $79,” will allow NVIDIA to maintain its dominant share of the low-end segment of the market.
NVIDIA had a slow start in the integrated chipset market: The nForce was launched in second-quarter 2001, and in third-quarter 2002 the company captured only 7.7% of the AMD chipset market. In the January and April 2003 quarters, however, due to the success of the nForce2, NVIDIA’s chipset shipments increased by 95% and 80% Quarter over Quarter respectively, and the company captured a 6.6% share of the desktop integrated market in 1st quarter 2003, up from 3.2% in the prior quarter. Its share of the AMD-platform chipset market grew to 11.4% from 8.6%. The company shipped more nForce2 chipsets in its April quarter than the number of nForce shipments in its life. NVIDIA is well-positioned for accelerated market share gain in the market for AMD-based chipsets, with the nForce2 having been very well received by customers, and with the nForce brand showing strength. Benchmark tests also reveal the nForce2 as having the performance leadership for Athlon XP chipsets. The share gain will be aided by the fact that the nForce2 is also offered as a discrete chipset, which increases the total available market to the entire AMD-based desktop chipset market, which is 10%-15% of the total chipset market. In first-quarter 2003, NVIDIA had only a 1.7% share of the total chipset market. The company is also well-positioned for chipsets for the Hammer processor.
2020insight.com |
