Comment: Regarding the semiconductor fabless company several lessons

Semiconductor industry is moving toward commercialization, few opportunities to differentiate. Fabless companies in the battle for market position and ADC0804LCN datasheet and market share on the occasion, no plant should follow in the field of supply chain trend. Competition is particularly fierce in the year coming to an end, the new year is coming. On this occasion, the author would like to think about the past 10 years, several trends emerged, and ADC0804LCN price and to explore what lessons one can draw. No plant in the topic read, if you think the things mentioned in this article similar to the real companies, it is just your own imagined.

Foundry field is changing the competitive landscape: the foundry industry has become increasingly boring and ADC0804LCN suppliers and not competitive, in fact the beginning of this century, several Taiwan companies took control of the industry. But when China joined the competition, things become very interesting. An upstart and a few smaller start-ups into the foundry industry, making the situation for fabless customers, very exciting.

Enterprises in Taiwan to the mainland will not allow the transfer of the islands most advanced technology, and the U.S. Congress will not approve sale of certain U.S. companies manufacturing equipment to China. Nevertheless, the economics of these new businesses had managed to survive the early ability to provide a camouflage. The emergence of new firms to the mainlands main strategy is to expand the price competition.

After numerous litigation and settlement after the foundry industry, changing the pattern, has led to sharp price decline in OEM (beneficial to consumers), but the winner remains the winner, and the new arrivals did not shake the old market leader in profitability and market share.

Lesson: never be rich and powerful competitors with the Competition price.

I should create funds invested in where? (Fab or assembly - test plant?): 90s of last century is that foundry lucrative, and assembly, testing and packaging (ATP) is the continued plight of manufacturers. Fabless foundry customers are loyal, but often replace the ATP firms. However, in the turn of the century, some things have changed, completely reversed the role of various manufacturers.

First, the communication industry bubble burst. Subsequently, demand for consumer applications, has become the main force to promote the semiconductor. Results silicon costs come under attack, and has dropped significantly. Adding new contract manufacturers, and manufacturing process to a more sophisticated level of rapid transformation, and promote this trend. ATP cost of service and the overall IC increased the proportion of costs, the results for many applications, the silicon cost is no longer necessarily the main factor in determining cost.

Secondly, the reference wafer foundry business practices, ATP vendors strict control over capital expenditure. They also began to greatly reduce the supply, making the supply situation became more favorable to myself. Here, a substrate, the cleaners had burned the factory, or elsewhere, there are two substrate factory bankrupt, it would certainly increase supply and demand imbalance.

Present, very few foundries to make money (probably two or three of the largest enterprises), but many ATP vendors are profitable (at least four or five largest firms.) ATP vendors seem attractive in financial terms, and even private equity companies are beginning to look at them, and has taken some action.

Lesson: an old dog can learn new skills

IP business: the viability of the semiconductor IP business is a hot topic of debate. In addition to several major companies (mainly the processor IP provider), it is difficult to find profitable and scalable IP services. Viability of the key, it seems from the perspective of business model is to bring the IP royalty revenue base, and from a technical point of view of IP should have a degree of customization.

Addition to the processor IP, most other types of IP are difficult to meet the business and technical requirements. For example, free or paid standard cell IP business has been a commodity business, increasingly facing the foundry provided from the library (library) the threat of any of these libraries provided by the manufacturer than IP are all closer to the actual wafers. PHY IP in general the beginning of the business to make money, but because of its standards-based, it will gradually become a commodity (some IPs that change rapidly and some slowly some).

General lack of accepted industry standards for IP quality, this is a problem, but the idea of customers and IP provider, it is not always consistent. The former want to standardization, in order to reduce costs; who often hope to compete with the quality and differentiation.

In addition, customers unwilling to pay large sums (if any) IP user fees (especially in consumer applications), the low-cost IP providers to join, continue to lower the cost barrier, also contributed to the lack of this business to attract power factor. Many have taken place in the field integration, this is not surprising, and certainly there will be more consolidation occurs.

Lesson: business model innovation and technology innovation at least as important.

Innovation gap: silicon to create or electronic design automation (EDA) vendors and supply chain, manufacturing (foundry) part of the innovation. Manufacturers spared no effort in pursuit of innovation to extend the life of Moores Law. Technology has a lot of innovation, such as reducing the size lithography methods, DFM / DFY, and including new material of innovation, including new transistor works.

In wafer production, companies not only focus on innovation, but also frequently invested in R & D above 10 billions of dollars of capital. EDA vendors will usually catch up with technology innovation, and strive to continually catch the silicon manufacturing over the curve ball thrown to him. But EDA vendors

managed development and marketing of the best equipment, so customers can create a simulation to accurately predict the performance of silicon design. On the other hand, ATP vendors (as a group), relatively speaking, not a lot of innovation.

QFP and BGA come out for a long time. Occasionally, these companies can be impressive, such as flip chip or QFN, but most of the time, always being innovative back-end challenges of the field, or the lack of any real creativity.

To wait four to six weeks to get a four-layer package substrate, and only a little more than a month you can get the required processing of more than 30 layers of silicon mask, it really surprise you? This imbalance in innovation will continue? Back-end IC manufacturing provider will call for "innovative or not," the slogan?

Lesson: sometimes the tortoise does run faster than rabbits

Conclusion

Semiconductor industry more heavily dependent on consumer-driven, the supply chain is constantly changing pattern. For industry participants, it is important to step back and pay attention to this gradual change and learn new things. Business model challenges to overcome, innovations need to make up the gap, the need to improve competition, all of which may result in continued industry consolidation. A competitive and stable supply chain for the semiconductor industry is the key to aging gracefully, the industry is constantly maturing.