Time flies, and 2006 is already halfway through. The global semiconductor industry in the first half of 2006 is growing steadily, much better than the same period last year. Although the growth rate is not very high, for an industry with a value of more than 200 billion US dollars, even a 5% increase is worth more than 10 billion US dollars. The semiconductor industry is changing and is more competitive than before. Looking back on the past six months, there are some important events that will have a significant impact on the future of the semiconductor industry.
CSIA and WSC signed a memorandum of cooperation
During the World Cup, the China Semiconductor Industry Association (CSIA) also "kicked" a crucial kick. June 15 was an important day. CSIA Chairman Yu Zhongyu signed a memorandum of cooperation with Brian Halla, the 2006 rotating chairman of the World Semiconductor Council (WSC) and chairman of the Semiconductor Industry Association (SIA). The signing of the memorandum marked a new chapter in the cooperation and exchanges between the Chinese semiconductor industry and the world's major semiconductor producing countries and regions.
Applying to join WSC will help the Chinese semiconductor industry to truly integrate with the global semiconductor industry, including data statistics, IP protection and international cooperation. On the other hand, starting to work under the international unified caliber will make the Chinese semiconductor industry more stringent and urgent. Therefore, the Chinese semiconductor industry must take this opportunity to keep track, find the right time, and strive to achieve breakthroughs in some areas.
Transfers are accelerating and mergers are intensifying
Although the industry unanimously believes that the global semiconductor industry will no longer experience the ups and downs of growth like in 2000 and 2004, but will instead experience a more moderate growth, the competitiveness of the semiconductor industry means that the rule of "the big ones get bigger" will hardly change.
The shift of the industrial chain to higher value-added development is a trend, and the semiconductor industry must also follow this rule. At present, the back-end packaging and testing industry has undoubtedly shifted to China. As for the chip manufacturing industry, advanced countries have begun to implement the "Fab-lite" strategy. The more successful one was Motorola's 28 semiconductor factories, which eventually only remained through mergers and sales. Even the new 8-inch production line built in Tianjin, China in 2000 at a cost of US$1.9 billion was eventually sold to SMIC for US$260 million in equity. But these 9 factories are all outstanding.
LSILogic also successfully transferred in 2005, selling its chip manufacturing plant to become Fabless; Agilent sold its chip line to become a tester supplier. Recently, there have been many mergers between other small factories, such as Germany's Xfab's merger with Malaysia's First Silicon; Japan's Epson spent $81.42 million to buy back its shares in a joint venture in Japan from IBM.
Also attracting global attention is the transformation of Infineon's memory division, with the new company "Qimonda" continuing to venture into the memory market; and Philips' decision to spin off its semiconductor division and list it on the stock market at the end of this year.
Whether it is Infineon or Philips, one thing in common is that they are implementing the "light fab" strategy, that is, the semiconductor industry chain is upgrading, and even fab manufacturing is gradually withdrawing and instead shifting to chip design, IP trading and third-party design companies.
Market situation remains unclear
The spring forecast of WSTS (World Semiconductor Trade Statistics), a well-known American organization, indicates that the global semiconductor industry will grow by 10.1% in 2006, reaching US$250 billion, and will increase by 11% and 12.8% in 2007 and 2008 respectively.
This figure is higher than the original WSTS forecast in the fall of 2005, which predicted an 8% growth in 2006. Overall, the semiconductor market has not seen cyclical fluctuations, and only some regions and product groups still have cyclical patterns. Among them, the Asia-Pacific region has the fastest growth, not only due to demand growth, but also relying on the transfer of production capacity. Recently, another organization Gartner said that due to the rise of the global PC market, it has raised its forecast for the semiconductor industry in 2006 from the original 9.5% to 10.6%.
On June 28, iSuppli made a new forecast for the global semiconductor industry in 2006, raising it from the original 7.4% to 7.9%.
But the industry has a different voice. Another famous organization, SIA (Semiconductor Industry Association), has adjusted the growth of the global semiconductor industry in 2006 from 11% to 9%, and believes that the growth of the semiconductor industry in 2006 may be only 4%, which clearly indicates that the global semiconductor industry will slow down.
Analysts from Handlebanken, another bank investment analysis agency, interpreted the WSTS report in April and believed that Intel's processor chips had a 21% decline in quantity and a 40% decline in price, and expected Intel's second-quarter performance to decline, with sales adjusted from $34 billion in 2005 to $33 billion in 2006. Its old rival AMD was also inevitably affected, with corresponding sales of $3.4 billion in 2005 and $3.3 billion in 2006. Therefore, on June 27, the global semiconductor industry was revised again, from 6% to 5%.
There are three main types of chip products in the world: CPU, memory including flash memory and FPGA (field programmable gate array). At present, due to the fierce competition between Intel and AMD, the CPU market has a downward trend.
The global memory industry has seen a significant increase in demand due to the rising demand for storage capacity in MP3, mobile phones and computers, but the pressure to reduce prices has also increased. In the first quarter alone, the price of NAND flash memory dropped by 30%. Therefore, the global memory industry's total sales in 2005 was US$47.19 billion, and in 2006 it will be US$49.25 billion, with a net sales growth of 4.4%.
In any case, compared with the downward trend in the beginning of 2005 due to the increase in inventory, the global semiconductor industry is doing well in the first half of 2006. The most fundamental reason is the growth in demand for the two major terminal products, computers and mobile phones. According to the latest data from IDC, the growth of PCs in 2006 has been raised from the 10.5% predicted in March to 10.8%, while the growth in demand for mobile phones can reach 18%. In addition, a very obvious factor is that the global investment in semiconductor fixed assets in 2006 reached 51.3 billion US dollars, an increase of 12% compared with 2005. The main increase came from Japanese chip manufacturers and DRAM industry.
However, there are also negative views, such as JPMorgan Chase's chief semiconductor analyst, Bowen Xia, who is worried that the IC inventory problem will explode in the second half of the year and is very worried about PC inventory. Another example is that in the second quarter, the global fabless factory inventory has reached a five-year high, and it is predicted that the global semiconductor demand in the second half of the year will not be as good as before, and may face downward pressure.
It is normal to have different views on the global semiconductor industry. However, the semiconductor industry is no longer what it used to be. The strong growth trend is no longer there. The semiconductor industry is becoming more mature, and the chances of being affected by various external factors are increasing. In addition, Intel and AMD, the leaders of the global chip industry, have been fighting each other recently, and it seems that no one has won. On the contrary, the rumors of the sale of the communication chip department and layoffs have been confirmed, which will have a huge adverse impact on the global semiconductor industry, especially the semiconductor index.
Memory is still a bright spot
Nam Hyung Kim, a well-known analyst at iSuppli, recently pointed out that the global NAND market has ended its downward trend.
Since 2006, the DRAM market has been able to develop steadily and no longer experience the ups and downs of the past, and NAND Flash has played an indispensable role. Since the first quarter of this year, although the global NAND Flash price has dropped sharply by 30%, which is enough to compete with the decline in 2005, many DRAM manufacturers are still not afraid of difficulties and continue to transfer DRAM production capacity to NAND Flash.
The reason why DRAM manufacturers are so determined is that NAND Flash has a high profit margin, which is much higher than standard DRAM. Therefore, so far, no DRAM manufacturer that has switched to NAND Flash production capacity is willing to switch back to standard DRAM. Under this circumstance, the increase in DRAM production capacity comes mostly from the improvement of process technology and the addition of 12-inch wafer production capacity, resulting in the DRAM industry still being the main player in the total investment in the semiconductor industry this year.
Judging from the demand for NAND Flash, after a quarter of price drops, by the end of March, consumers had begun to further stimulate demand due to the sharp drop in the overall average selling price of NAND Flash, and NAND Flash prices had reached the bottom. It is rumored that Samsung Electronics and Hynix will raise the price of NAND Flash in the near future.
Gartner predicts that the global mobile multimedia player market will reach 187 million units in 2006, compared with 134 million units in 2005, of which 80% will use NAND flash in 2005. It is expected that in the fourth quarter of this year, Apple will once again launch a high-end NAND iPod with a capacity of 10Gb-12Gb. Gartner believes that the impact of the iPod MP3 market on the global NAND flash market cannot be underestimated.
NAND flash storage capacity has doubled every year since 2002. The first quarter of 2006 was in excess due to seasonal reasons, while the second quarter became tight. A total of 1%-2% excess is expected in 2006. The production of memory in 2006 was 641,031TB (1TB=1024Gb), an increase of 198% over 2005, while the demand in 2006 was 636,572TB, an increase of 194% over 2005. In the NAND flash market, MP3 accounts for 37%, USB accounts for 17%, mobile phones account for 14%, and others account for 5%.
The DRAM market is experiencing a two-year cycle. From the current situation, most DRAM factories' 8-inch factories will be phased out because they cannot continue to upgrade to the next generation of process technology. The industry predicts that from the perspective of cost-effectiveness, the 12-inch DRAM silicon wafers in 2005 will exceed the 8-inch. However, since the production capacity of the newly added 12-inch factories is still relatively small, when the industry gradually loses the DRAM production capacity produced by the 8-inch factories, even if there are a lot of new 12-inch production capacities in 2006, it is likely that it will inevitably fall into a state of "supply exceeding demand". Therefore, it is expected that the price of DRAM will not continue to weaken throughout the year.
Another important factor is that the DRAM industry has decided to switch from the original DDR specification to DDRII. Since the first quarter of 2006, the price of DDRII in the market has been significantly higher than DDR, and the gross profit of each DDRII product is 30% to 40% higher than DDR. As a result, DRAM manufacturers have been working hard to increase DDRII production capacity, which has also led to a decrease in DDR output. It is estimated that the global DDR output in the second and first quarters has decreased by more than 50%.
However, since the specifications of PCs in most emerging markets are mainly focused on cheap low-end products, DDR architecture PCs are still the mainstream. Under this circumstance, the market demand for DDR products has not actually dropped significantly, but the supply of DRAM factories has been greatly reduced, causing DDR prices in the spot market to rise significantly, with 256Mb DRAM prices reaching US$2.4 to US$2.5, the highest price since 2006.
The biggest potential market for NAND mobile flash memory cards in the future is in mobile phone applications, which have recently begun to use NAND as an external storage solution. iSuppli predicts that in 2006, mobile phone applications will surpass digital cameras to become the main market for mobile flash memory cards. It is predicted that the average capacity of mobile phone memory cards in 2005 was 112M, and will increase to 1.6G by 2010, equivalent to a 70% annual capacity increase. In 2005, the number of removable memory cards used in mobile phones increased by 160%.
45nm production is on the agenda
According to Intel, the first chip plant using 65-nanometer manufacturing technology in Portland, Oregon, began manufacturing chips in 2005. The second plant in Chandler, Arizona, began operations earlier this year. On June 23, Intel's third 65-nanometer chip plant began production, and the Irish plant is the most cost-effective.
Intel President Otellini said that Intel has adopted the most advanced 65-nanometer manufacturing process technology to produce a large number of chips in three factories. Among the microprocessors provided by Intel for computers and servers around the world, more than half are produced by 65-nanometer manufacturing process. The company also plans to launch more advanced 45-nanometer manufacturing process chips by the end of 2007.
In addition, TSMC's vice president Jack Sun said that the company will start mass production using the 45-nanometer production process as early as the third quarter of 2007. In addition, market sources pointed out that Taiwan's UMC is also currently committed to developing a 45-nanometer production process. Sun pointed out that TSMC has successfully developed a 45-nanometer system chip development platform and is working hard to make the company one of the first three manufacturers to successfully adopt the 45-nanometer production process. IBM is currently the only manufacturer to break this milestone. The company previously claimed that it had successfully developed a 29-nanometer production process.
TSMC has been working hard to develop its own advanced production technology recently. The immersion lithography technology developed by the company and ASML may become the mainstream technology in the future. TSMC expects to use the 32-nanometer production process for mass production in 2009.
The OEM industry has entered the era of "super factories"
According to IC Insight, the global foundry market reached US$16.95 billion in 2005, of which TSMC in my country accounted for US$8.22 billion, accounting for 48.5%; it is predicted that the global foundry market will grow by 32% to US$22.3 billion in 2006. From 2005 to 2010, the global foundry industry is predicted to have an average annual growth rate of 21%, which is twice as high as the semiconductor industry during the same period.
In fact, the above estimates may be too optimistic. The global foundry industry may rise steadily, but not too fast.
TSMC's senior vice president of global business and services, Jin Lianfang, said that the global semiconductor industry has entered the "interactive growth period" from the "chaotic period" since the turning point of the Internet bubble in 2000. It is expected that in the next 5 to 10 years, it is unlikely that there will be amazing "killer" products that will bring explosive growth. However, the emerging markets created by the application of "small killer" products extended from existing products will drive the steady growth of the semiconductor industry.
For example, the global mobile phone market has two main directions of development: one is multifunctional smart phones, and the other is low-priced products targeting the needs of emerging markets such as China, India, and Brazil. In addition to mobile phones, the other is low-priced computers, such as the low-priced computer concepts recently proposed by Intel, AMD, and the Massachusetts Institute of Technology (MIT) for emerging markets. These are all driving forces to stimulate the continued growth of the semiconductor industry.
As for the next development of the foundry industry, Jinlianfang believes that there is currently one fabless design company with a scale of more than US$3 billion, four with a scale of more than US$2 billion, and the remaining five with a scale of more than US$1 billion. The foundry industry can support the outsourcing capacity needs of the entire fabless design industry and some IDM factories. At present, about 70% of the orders for 12-inch factory capacity in the world come from foundry companies, which shows that the role of foundries is becoming more and more important, which is also the reason why the foundry industry can continue to grow.
It is worth noting that as the global semiconductor industry enters an era of integration, manufacturers continue to increase investment, expand production scale and economic benefits. Jinlianfang pointed out that the future will
enter the era of "super factories" (12 inches, with a monthly production capacity of more than 80,000 pieces), and those small and medium-sized factories with a production capacity of less than 30,000 pieces will face severe challenges. For those large wafer foundries, they will integrate IC design, back-end packaging and testing, and electronic design automation (EDA) companies, playing the role of irreplaceable ecological integrators.
At the same time, due to the rapid progress of semiconductor process technology, the two foundry giants have entered the ranks of the world's advanced. TSMC's 45-nanometer process technology has made significant progress and will accelerate its efforts to compete for the top three in the world. It is expected to be mass-produced as early as the third quarter of 2007. As for the 32-nanometer process technology, it has also entered the development stage and is expected to enter mass production in 2009. This situation has raised the threshold for entering the high-end foundry field and also caused the plan for Japan's first foundry, which had been brewing for half a year, to completely dissipate.
After entering the era of integration, super factories not only have the ability to integrate the front-end and back-end to reduce customer risks, but their production scale and economic benefits are also more important. The trend of the global semiconductor industry entering "super factories" is irreversible.
The global semiconductor industry is discussing the survival prospects of the foundry model, believing that foundries exist only for fabless companies, and that they will inevitably face price wars in the future due to increasing homogeneity, leading to differentiation, mergers and reorganizations of the global foundry industry.
As chip integration and process complexity increase, the cost of developing new products has skyrocketed, and the interdependence between design and manufacturing has become inseparable. Therefore, the industry has once again raised the argument for the IDM model.
提升卡
变色卡
千斤顶
