Shifting Sands the Future for Semiconductor Foundries

Shifting Sands The Future for Semiconductor Foundries

By Enrique Duarte Melo and David Michael

New dynamics are combining with historical efficiently. Companies such as Taiwan determinants of success to reshape the Semiconductor Manufacturing Company semiconductor sector. The first article in our (TSMC), Global Foundries, and UMC have series on these “shifting sands” explored how grown into major players alongside the challenges of managing scale, dealing with integrated device manufacturers (IDMs) market volatility, and innovating at the speed such as Intel, Samsung, and Texas Instru- of Moore’s law increasingly share space on the ments (TI). As we move further into the strategic priority list with the demands of the post-PC era, however, foundries—as well growing mobile-device market, a fragmenting as design companies and IDMs—face big universe of devices with short life cycles, and changes and choices in how they do new technologies to support sophisticated business. Major shifts in product mix, big applications such as advanced graphics. technology challenges, and increasingly Nowhere are these challenges coming together demanding economics will cause foundries in a more concerted fashion than for found- to rethink their basic value proposition. ries. The long-term revenue outlook is strong for foundry companies as a group. Individual- Foundries face broader strategic options ly, however, they face major shifts in both and imperatives than ever before in terms technology and marketplace demand, necessi- of the products they manufacture and the tating a comprehensive reexamination of technologies they employ. At the same strategies and business models if they are to time, the mushrooming cost of construct- perpetuate the same levels of growth that ing new, advanced technology capacity have propelled their rapid rise. heightens the financial pressure to make the right bets. Only the biggest companies, emiconductor foundries were such as TSMC, will have the ability to serve Sborn out of the need for the fast-grow- a wide array of customer segments. For ing semiconductor sector to construct and others, new capabilities, in addition to manage manufacturing capacity more scale and utilization, will rise in impor-

For more on this topic, go to bcgperspectives.com tance, especially as large, deeply resourced grown in size and has also been carved into competitors, such as Intel and Samsung, a greater number of slices. seek footholds in the foundry market. Intel, for example, has recently contracted The long-term outlook remains strong, but to manufacture Altera’s field-programma- it will continue to be fueled by varied ble gate array chips using its 14-nanometer forces: the explosion in digital devices with tri-gate transistor technology. These differing power consumption needs, the dynamics have ramifications for the entire rising demand for low-cost systems and value chain, from design companies to devices in emerging markets, the need to original equipment manufacturers. process ever more vast quantities of data, and the continuing penetration of semicon- Foundries will have to develop the ability ductors throughout other major sectors of to assess end-user—often meaning consum- the economy. Gartner projects the semicon- er—demand and needs. Which products ductor market will grow at about 6 percent will generate new sources of revenue? a year—approximately twice the rate of Which technologies are the best fit? Which GDP growth—to more than $400 billion in design customers and IDMs can a foundry 2016. Demand at foundries will increase best work with in order to meet this even faster. Revenues grew more than 7 demand? percent from 2011 through 2012 and are expected to increase more than 7.5 percent This is no small challenge. Big, integrated in 2013 to some $37 billion. companies such as Intel and TI have very different strategies, priorities, and needs Fabless design companies, long the biggest from those of younger designers such as customer segment, will drive some three- Broadcom, Qualcomm, Marvell, and quarters of demand at foundries, but MediaTek. Companies that straddle much increased outsourcing at IDMs will also be of the value chain, such as Apple and responsible for a significant share. The top Samsung, have their own competitive ten fabless design companies and the five dynamics. Working hand in glove with the largest outsourcing IDMs are expected to right partners to establish a strategy and generate approximately half of foundry vision for the future will become critical in revenues. This customer concentration will creating value for many foundries. become an increasingly important industry dynamic. (See the exhibit “Foundries Face a Significant Concentration of Revenues.”) From Standardization to Proliferation Managing scale and productivity and As the PC era took off, rising demand for keeping pace with Moore’s law are still key faster, more powerful computers and more strategies for foundries. So, too, are meeting and more memory put a premium on scale demand for a widening array of semicon- and high rates of capacity utilization in the ductors with varying features and capabili- chip-manufacturing process. PC architec- ties, such as sophisticated graphics and ture converged around a few design power management, and adapting to longer standards, and foundries enjoyed rapid and R&D timeframes and shorter product life profitable growth as a result. In recent cycles. The biggest driver of demand will be years, however, the sources of demand have the exploding number, and changing nature, proliferated to include sophisticated new of digital devices. Computing and consumer consumer devices such as smartphones and electronics represent 60 percent of the tablets, as well as industrial applications end-user market for chips. Recent BCG such as embedded processors in automo- research in the U.S. found that consumers biles. These devices and applications have today own an average of 2.9 digital devices different processor needs from those of (including computers, smartphones, and PCs, a phenomenon that has resulted in the tablets), almost twice as many as three years increased use of architectures such as ARM- ago, and they are likely to own 4.1 such based design. As a result, the device pie has devices in three years’ time.

| Shifting Sands 2 Foundries Face a Significant Concentration of Revenues

Semiconductor revenue Foundry revenue composition composition, 2010 ($billions)

Texas Instruments Top-ten fabless 39 Revenue from fabless Toshiba design companies STM

Other fabless 26 Renesas Other IDMs

Outsourcing IDMs 91

Revenue from IDMs Others1 143

2 Tota l 299 Revenue from OEMs

0 100 200 300

Sources: Gartner; BCG analysis. Note: IDMs = integrated device manufacturers; OEMs = original equipment manufacturers. 1Includes Intel ($42 billion), memory companies ($33 billion), IDMs that run foundry services ($31 billion), and IDMs with minimal outsourcing ($38 billion). 2Includes a small contribution to the foundry segment and is not a focus of this analysis.

The Shifting Economics of Power nanometers to an estimated 18 months for Technology issues and financial constraints 28 nanometers and beyond. complicate the strategic choices further. The market continues to expect process The inevitable result is diminishing pay- technology advances (such as decreases in back on ever-bigger investments—a size from 45 nanometers to 32 nanometers difficult prescription for value creation. to 22 nanometers and beyond) that simul- Foundries also face increasing complexity taneously consume less power and provide in design with the advent of 3-D integrated increased performance. This puts a premi- circuits and the transition to larger wafer um on precision tooling (particularly diameters (from 200 millimeters to 300 lithography) and yield optimization, which and 450 millimeters). Several key technolo- is leading to exponential growth in capital gies are approaching the limits that physics costs for advanced technology nodes. places on further advancement.

Most companies cannot keep up with A new, big bottom line looms: as the market leader Intel, which already has amount of capital required to fund devel- 22-nanometer chips in production and is opment of advanced nodes grows, only a working toward the production of 14-nano- few of the biggest players have the balance meter chips in the near future. Three of sheet capacity to lead in process develop- the next-largest players trail significantly. ment. Some current leaders soon will be- At the same time, the rising cost of invest- come followers, and companies used to win- ment in each new generation of technolo- ning on technological edge will need to look gy makes it harder to create the necessary for new sources of competitive advantage. returns. In addition, while the market expects faster ramp-up cycles to meet the introduction of new products, the time New Capabilities for an from “tape out” to initial production for Evolving Marketplace foundries expands with advanced nodes— The net impact for just about all foundry from 12 months for 65 nanometers and 45 companies will be the need to rethink

| Shifting Sands 3 options and magnify bets. Not all found- ing of customer needs, which include the ries have evolved equally, of course. With following: 2012 revenues of $16 billion, TSMC is the market leader by a considerable margin; •• Process technology needs (technology the next two largest players, Global nodes and transistor characteristics Foundries and UMC, and are about such as gate materials) one-quarter its size. Each company has its own strengths in technology and manufac- •• Intellectual property needs (such as turing, but most do not have the resources specific input-output or I/O, physical to serve every design company or IDM or design or PHY, and analog technologies) to develop the intellectual property necessary to meet the increasingly varied •• Markets served—such as logic, embed- technological demands of a fragmenting ded software (including all its subcat- market. They will have to make choices, egories), and memory based on their current market position, technological capabilities, financial •• Business interface needs, including strength, and customer relationships. Not technical assistance with yield analysis the least of these decisions will be choos- and design impact, high-priority “hot ing where to commit constrained resourc- lots” for rapid testing, volume of chips, es and how companies will position cost pressure, geographic location themselves for the future. For example, do needs, manufacturing diversity, pricing they want to be technology leaders or to models, and investment needs focus on niche segments? China’s SMIC, for example, has decided to focus on •• A customer’s propensity for outsourc- second- and third-generation technology ing throughout the design and manu- applications. facturing process

As this process plays out, design companies In areas that lack scale or process technol- and IDMs will likely find themselves with ogy leadership, foundries also need to fewer manufacturing partners from which determine if there is an opportunity to to choose. They will be pushed toward serve as a second source of supply. This working more closely with a few compa- raises a different series of challenges, nies that can consistently meet their needs including organizing to perform at similar from a technology and capacity point of levels of service as the primary supplier, view. All players will need to think about managing functions such as inventory and new questions of strategy, competitive variation in manufacturing volume, advantage, and operations. maximizing facility utilization, and pricing. Foundries need to decide—first, foremost, and quickly—which segments of the Design companies need to consider their market they can most profitably serve. options as well. They’ll need a foundry Their next challenge is to build close strategy that can deliver the technology working relationships with those targeted base, product capability, investment customers. They need to gain an intimate capacity, volume capacity, flexibility, and understanding of their strategies and pricing they require. One big question is needs—and get embedded into each whether a sole supplier will be adequate customer’s product roadmap. They also (and advantageous) or if the design compa- need to determine where to place bets on ny can generate sufficient volume to technologies and how to build (or acquire) benefit from a second source. In addition the capabilities required to serve the to competition, of course, secondary market segments that they target. suppliers can provide potentially useful geographic or product fits as well as Developing a successful segmentation insurance against difficulties for, or inter- strategy requires an in-depth understand- ruptions from, the primary supplier.

| Shifting Sands 4 In structuring their supplier relationships, the yield curve quickly, speeding time to design companies will also need to careful- market, and lowering development costs. ly consider goals and objectives—for example, whether they are looking to eliminate fixed costs and move to a more he fabless foundry model has variable cost structure, to free up their Tserved semiconductor customers and balance sheets, to focus on innovation investors well for almost three decades—a rather than utilization, and so forth. near-eternity in the technology sector. Contracts need to align with these objec- Continuing to create value will be a more tives, and pricing will inevitably be a key complex challenge as structural changes in consideration. Does the design company the market combined with the cost and pay per wafer, die, premium die, yield, or technological challenges of ongoing on some other basis? Design companies advances bring new pressures on foundries should motivate foundry partners to get and design companies alike. We believe embedded in their product and technology the foundry model can continue to pro- roadmaps, which is something, as already duce significant value, but doing so will noted, that foundries will be looking to do. require finding new sources of competitive Both sides will have to address tough advantage for a changed environment. The financial questions such as who pays for need to place costly bets makes it all the inventory during a market downturn. more essential for foundries to work even There should be adequate incentives on more closely with their customers in both sides to develop close, integrated meeting the needs of a rapidly segmenting, relationships that can facilitate getting up increasingly complex market.

About the Authors Enrique Duarte Melo is a principal in the Dallas office of The Boston Consulting Group. You may con- tact him by e-mail at [email protected].

David Michael is a senior partner and managing director in the firm’s San Francisco office. You may -con tact him by e-mail at [email protected].

The Boston Consulting Group (BCG) is a global management consulting firm and the world’s leading ad- visor on business strategy. We partner with clients from the private, public, and not-for-profit sectors in all regions to identify their highest-value opportunities, address their most critical challenges, and transform their enterprises. Our customized approach combines deep insight into the dynamics of companies and markets with close collaboration at all levels of the client organization. This ensures that our clients achieve sustainable competitive advantage, build more capable organizations, and secure lasting results. Founded in 1963, BCG is a private company with 78 offices in 43 countries. For more information, please visit bcg.com.

| Shifting Sands 5