According to Gartner, the semiconductor market will continue to experience above-average growth of about 6 percent a year, reaching approximately $500 billion in 2016. This rate is about twice that of world GDP growth and on par with other technology sectors, such as software and services and telecommunications. Growth in semiconductors will be driven by the proliferation of smart devices, the automation of everything from automobiles to zoos, and the need to process and store an explosion of data.
In this sense, the outlook is auspicious. But it is worth bearing in mind three factors. The first is that the semiconductor sector historically has experienced extended periods of growth interrupted by sharp downturns that last 12 to 18 months. As a result, the sector has suffered from revenue and earnings volatility, and companies have had varied success returning value to shareholders.
The second and more significant factor is that not all companies have had sufficient scale to create value for shareholders. In the past, it has generally been the biggest players that have succeeded, although some companies—such as Qualcomm and ARM—have developed new business models and approaches to scale, also creating value. Most of the rest have underperformed. (See Exhibit 1.)
The third factor is that Moore’s Law is increasingly straining balance sheets and income statements and affecting the ability of companies to play in multiple markets. The physical constraints of silicon real estate and the sheer cost of increasingly sophisticated R&D are threatening the financial viability of continued investment.
To create value going forward, companies must, as they have in the past, have the ability to manage volatility and address issues of scale. However, they must also determine whether and how to pursue Moore’s Law and how to differentiate their products to meet new market demands. At the same time, companies will face other challenges from a handful of big and far-reaching trends that are converging now.
Evolving Sources of Demand and Shorter Life Cycles. The semiconductor value chain has always been complex and fragmented. (See Exhibit 2.) Now, it is in significant flux owing to broader market trends. Chief among them is an end-user customer base that is evolving from predominantly business users to consumers and that is buying an expanding variety of primarily mobile devices. According to iSuppli, for example, the number of smartphones shipped worldwide will exceed 1 billion in 2015, more than double the total number shipped in 2011.
Another significant trend is that more and more semiconductors are being used in a wider variety of devices and products. For example, Strategy Analytics expects the global automotive semiconductor market, worth $23.7 billion in 2011, to grow at a rate of more than 9 percent per year through 2016. The market for purpose-built consumer devices is also rapidly expanding.
These evolving sources of demand are leading to continued growth in the sector, but they are also changing it. Consumer and customer requirements for continuous improvement and innovation have exerted pressure on semiconductor designers and manufacturers, particularly when it comes to producing chips that require less power. At the same time, these requirements have led to shorter product life cycles, making the longevity of demand for any particular chip increasingly uncertain.
Such changes are calling long-standing business models into question.
Industry giants such as Intel and Samsung have built fortress-like positions through vertical integration, controlling all aspects of their design and manufacturing process. This model made sense when the largest customers for semiconductors were computer hardware manufacturers, which built PCs and laptops according to specifications that had evolved into industry standards, driving down costs dramatically. As PC and laptop growth slows, and the proliferation of mobile devices increases, the demand for a wider selection of chips with varying functionality is challenging large semiconductor companies to manage issues of step changes in manufacturing technology, product complexity, and manufacturing capacity utilization.
Although we do not expect the integrated model to disappear, it is coming under sustained attack from the fast-growing “fabless foundry” approach, which breaks down the design and manufacturing process and involves different companies concentrating on distinct functions or intellectual property blocks. Fabless designers (such as Avago Technologies, Broadcom, Marvell, MediaTek, Qualcomm, and Xilinx) and high-efficiency, low-cost foundries (such as Globalfoundries, Taiwan Semiconductor Manufacturing Company [TSMC], and United Microelectronics Corporation [UMC]) have staked out large and growing positions, attempting to flank their vertically-integrated competitors (such as Intel, Samsung, and Toshiba). The trend toward deconstruction is most pronounced in the digital segment, where product life cycles are short, capital requirements are high, and scale matters. Vertically integrated manufacturers remain especially strong in the analog segment, where product life cycles are significantly longer and proprietary manufacturing processes have been developed.
Regional Market Concentration. The semiconductor sector is global, with design and manufacturing taking place in multiple Asian, European, and North American locations. Nevertheless, in recent years, as more original-design manufacturers have opened facilities in Asia, so, too, have more chip makers, particularly those seeking skilled labor, low costs, and proximity to customers. As a result, Asia has become the largest regional market for semiconductors by a substantial margin, a position we expect it will continue to hold. Semiconductor companies without a significant presence in the region will find it increasingly difficult to compete. And equipment manufacturers will find themselves facing new low-cost competitors from the region.
A Focus on the User Experience. Since the PC was introduced, the high-tech industry has largely focused on chips—their speed and power. However, the technological and financial challenges of keeping pace with Moore’s Law have caused the industry to shift away from this obsession toward advancing the end-user experience. This has helped usher in the post-PC era.
Smartphones and especially tablets are the new focus of consumer demand. Gartner expects sales of tablets to reach 370 million units by 2016, a much faster ramp-up than any other consumer-electronics or mobile device. Intel’s announcement that year on year, its 2012 third-quarter revenue declined by $700 million or 5 percent, which includes an 8 percent revenue drop in the company’s PC Client Group, is evidence of the move away from PCs to purpose-built devices.
With these new devices in hand, users are fast embracing digital ecosystems built around operating systems such as Apple’s iOS and Google’s Android. Although task-based applications, such as word processing, have been—and will continue to be—popular uses, these new ecosystems provide an enhanced and broader user experience. The applications that run in these environments are increasingly interactive and give users the ability to share information in real time. And the devices that host them incorporate a wide variety of technologies that enable richer human–computer interfaces, such as voice and handwriting recognition. Companies such as Amazon.com, Apple, Baidu, Google, Microsoft, and Samsung are responding by creating and presiding over content- and capability-rich environments. As the ecosystems and applications continue to evolve, the devices that access them will need an ever-expanding variety of chips that can interface with software, supporting integrated solutions.
Increasing Pressure on Equipment Manufacturers. Semiconductor equipment companies face particularly complex challenges as financial and technological constraints lead their customers to pursue different strategies. Some designers and manufacturers are pursuing new technologies, such as extreme ultraviolet (EUV) lithography, fin field-effect transistor or FinFET, and through-silicon via interposer. Others are reluctant to invest in new technologies and equipment absent widespread demand, which has failed to develop partly because their customers are doing more with existing technologies (adding multiple layers, for example) and partly because they doubt that new technologies can be delivered and scaled quickly enough in a world of shortening product life cycles. At the same time, equipment manufacturers continue to struggle with the highly volatile nature of the sector, with its large swings in revenues and, more important, margins, over time.