Integration plus cost has lead to fewer suppliers at leading-edge levels; this can mean trouble for supply. But, new diversification trends may create opportunities for partner and sector growth along the supply chain. So, does a post-PC device entail a post-PC supply chain or not?
The tremendous demand for tablets and smartphones is driving changes in our industry, from chip design to supply chain alignments. These changes hold great opportunities but also noteworthy challenges, for supply chain partners, device markets, new hardware ecosystems, component innovation, and even the pursuit of Moore's Law itself.
While smartphones and tablets are spurring overall semi growth and integration along supply chains and within devices, across the board there are increases in semi content as more and more aspects of business, transportation, and daily life rely on technology, in both mature and emerging economies (cf. this IC Insights reporton semi content growth forecasts). However, along with this growth are the new demands from consumers and enterprises, regardless of market sector: mobility, connectivity, multi-tasking, power efficiency, and multi-media capabilities.
These feature demands translate into new component requirements at a time when both R&D and fabs for next generation wafers and architectures come with astronomical price tags. These investment costs do result in more sophisticated components and devices, and that is what people are demanding. Somewhat of a higher price is being accepted, but as 'smart becomes average,' price pressure strikes and leads to shrinking margins which can lead to a shrinking number of suppliers.
So, what do we make of these changes? There is certainly a healthy influx of new design trends and new drivers for components, but what will the supply chain supporting these demands look like? Growth is certain. The real question is growth for whom, where, and for what. The recent and terrible events in Japan underscore both the interdependence of partners along the supply chain as well as the fragility of these networks. As Japan moves to recover and then rebuild itself in the wake of these catastrophes, there will also be a renewal of suppliers as new businesses arise and older partners rebuild and refocus their capabilities and connections.
Overview of market opportunities
Prior to the March 11 events in Japan, the market situation was one of anticipation and seasonally strong demand from consumers and businesses in both mature and emerging economies for a variety of end-products across a variety of market sectors (for example, tablet/PC, smartphones, handsets, industry, automotive, entertainment and infotainment, and health care IT). With roughly a two-week excess or buffer in inventories in the supply chain, there were moderate concerns of inventory build ups should demand not come to fruition. Now, with many questions about what will happen as this buffer is depleted, the more pressing question is how will suppliers support demand given not only component shortages but more critically, the upstream material supplies that are essential to the most fundamental parts of chips (e.g., aluminum, resin, wiring, etc.).
Despite the now confounding problem of managing a wide-spread shortage situation, the end-product/device forecasts are not just growing in unit volume, but importantly seeing increases in semiconductor content to meet new feature demands (read more about the forecasts and devices in the two companion articles in this issue of MarketWatch Quarterly here and here).
With all of this demand for both more unit volume and greater semi content per device, the industry data for the components, especially ICs, should also be strong. One recent description of the World Semiconductor Trade Statistics (WSTS) January's IC sales data was "shocking." (cf. this IC Insights report) Yes, everyone has been worried about the impact of tablets and smartphones cannibalizing sales from PCs and related devices, and you might just jump to the conclusion that this is where the bad news meets the good news. Except that this use of "shocking" is intended to be good, really good, as in historically good:
[E]xtrapolating January 2011's IC figures with the 1999-2010 average growth for February (3%) and March (28%) suggests IC sales for 1Q11 could top [US] $71.6B, a 14% surge (and a 27-year high). Even a worst-case scenario using the historically poorest sequential IC growth rates for February (07%) and March (19%) translate to slight growth in 1Q11 (3%). (cf. the Solid State Technology article summarizing the IC Insights report)
Granted, as we are reminded here by EETimes, IC Insights tends to be more bullish on IC forecasts and WSTS data, so a tempered forecast representative of a wider group of analysts is more likely in the 5-8% growth range for 2011. Still these data speak to solid growth, which entails opportunities. The new, overarching question mark is will upstream material fab, and down through assembly lines at other locations be able to handle the production and supply void left by Japan for the interim?
Despite strong IC growth forecasts, there are many market challenges that were existent even before March 11th. These challenges range from the competition underway in the tablet/PC sector and concerns over the effects of device convergence (tablet versus notebook, cloud versus HDD) to the macro-economic (will oil prices surge over US $100/barrel and remain there, dampening consumer and business spending). One caveat, the macro-economic scales have tipped in light of the rebuilding and restructuring efforts necessary in Japan.
Rising oil prices do actually hold growth for some sectors of the semiconductor supply chain, particularly for growth in automotive electronics (cf. this IC Insights forecast for automotive semis).
As prices at the gas pump rise, and businesses and consumers alike look at transportation costs, the adoption rate increases for semi content intensive electric vehicles (EVs) (cf. here and here from our MarketWatch resources, and this recent report from Manufacturing Business Technology on Toyota's 3 million Prius sales numbers). Additionally, there is increasing demand for the already well-received hybrids, as witnessed by both recent volume figures, additional competitors and an expansion in the variety of automobile types in the hybrid and EV categories (cf. this recent report from WSJ.com on new family models from Toyota and this analysis (subscription required) from Financial Times regarding market strategy).
The rise in automotive semi content is vital to the industry in terms of demand for a diverse set of components, which then support a broad-based supply chain (see this article for the latest ideas from Ford in EETimes). By having the cascade of semi content from high- to low-end vehicles for consumer to business uses, there is opportunity for moving older technologies along the chain while refreshing inventories with increasing volumes of newer components. Furthermore, the rise in hybrid, EV and even simply more fuel-efficient vehicles worldwide supports innovation of a mixed set of components as well as the leading integration demands of portable devices (mobile and tablet/PC) with the increased capabilities of the automobile itself (real-time navigation, safety, lighting, audio-video, monitoring and warning systems, Bluetooth capabilities, etc.).
Component innovation driven by 'ecosystem' challenge
The new 'device ecosystem' is essential to understanding today's drivers of a strong market forecast whether at the component, device or customer level (cf. this IC Insights report and this SIA report). In terms of component drivers, the device ecosystem has cemented a change in the way that these smart wireless devices (SWD) are being incorporated into daily life, whether for recreational, health, transportation or business uses.
The device ecosystem, as discussed in this MarketWatch Quarterly companion piece, is the combination of capabilities (namely, hardware, software, and service) that together shape the user's experience with the device, whether that device is a phone, a PC, or an automobile. Importantly, the 'always on, always available' demand for devices today informs what capabilities are seen as required of components. Components must be able to support the device's full, customizable, user experience (the components must maximize the software and quickly support multi-tasking, requiring various processors to be integrated in function, if not form) for a maximum amount of time (lowered power consumption), preferably regardless of location (ability to capture the mobile service signal and alert the user to where their location is, upon demand), and provide multi-media functionality and interconnectivity that optimize the users' experience and needs as those needs change in real-time (audio, video, text, and cloud computing/SaaS for data management and transfer).
While these demands of devices seem par for the course from a user perspective, the challenges for component design are considerable. To the benefit of the industry, there are some general component designs that are transferable across sectors (e.g., healthcare, automotive, tablet/PC, mobile/wireless): multi-core architecture, fast (≥1GHz) clock frequency, touchscreen interface, high resolution graphics (GPU) integrated with MPUs, strong RF signal, baseband processor, multiple antennae, low power consumption and higher battery lifetime, to name some of the top capabilities. Couple these sophisticated capabilities with the shrinking size and weight and one is immediately looking at solutions requiring 4xnm to 2xnm node.
Additionally, the integration of processors while minimizing power and heat while increasing speed, means integrating more functionality into ICs. Given the space constraints of the devices at the forefront of demand, SWDs, we quickly move to smaller sizes and integrated stacking increases in importance (cf. the recent report on application processors by Petrov Group, available from DigiTimes 8 March 2011). Multi-core processors are already becoming standard for smartphones and the rate of adoption and growth is showing 'smart becoming average' quickly, which means multi-core will soon be standard throughout the smart wireless device (SWD) sector (cf. "Handsets: Making the Call," Citigroup Global Markets February 22, 2011).
These innovative drivers are important positives for the industry, directing new technologies and architectures towards devices and sectors where end-products containing these now more sophisticated (and higher priced components) are finding healthy demand and adoption. However, as more is being required, and despite the increases in ASP for BOMs, there is also the requirement of increased efficiencies at the manufacturing side in order to improve margins. The concern over margins is twofold: the cost of fabs is skyrocketing and new investments are favoring new wafer and architectures rather than increased capacity of existing lines; and as the adoption of these demanded devices increases, 'smart becomes average' and the tipping point of economies of scale enters (i.e., higher volume for lower prices puts pressure on costly BOMs).
'Fab-tight' future at 300mm and 450mm?
Taken at face value, these component demand situations would typically point to a market situation supporting a diverse set of value chain partners from design to manufacturing and assembly. What we have presently in the value chain are exciting times for designers, but real strategic challenges for the businesses. In order to achieve the feature capabilities expected of the components and then to have that interoperability translate into real functionality in a cost-sensitive device, maximum efficiency is necessary. This efficiency requirement from production is driving the full adoption of 300mm wafers and the certain move to 450mm quickly.
The move to 450mm is seen in the leading fabs CAPEX plans and equipment orders being placed. The reason for the efficiency requirement and size shift is simple, margins. The reality is skyrocketing costs mean that each wafer produced needs to maximize revenue in exchange for the cost of production at the fab, and along the supply chain. These new wafer sizes (and the die shrinks) are very costly, and these costs increase the barriers to entry for new competitors but also for existing manufacturers to continue to operate (see also our MarketWatch Commentary analysis here). As SEMI recently reported:
In 2010, SEMI's World Fab Forecast counted seven 300 mm volume fabs beginning construction. However, in 2011, only Intel's fab is predicted to start […]. In 2012, three 300 mm fabs will begin construction: one foundry and two integrated device manufacturers (IDMs); two of which are potential candidates for 450 mm-ready cleanrooms. […]
For the first time, SEMI's World Fab Forecast data identify seven facilities (R&Ds, pilots and volume fabs) in the near future that are candidates for 450 mm readiness. The first facilities are expected to come on line in 2013. (cf. here for the SEMI report)
What do these fab forecasts mean for the majority of the supply chain? Quite a bit, actually. The trending toward fewer and fewer fabs raises questions about utilization and what SEMI aptly calls 'fab-tight,' meaning the lack of manufacturers may lead to shortages, preferential order production, and even more worrisome, "companies may not be able to meet new demand by upgrading their technologies any more. Lead-times for a new fab are just too long, typically taking 1 to 1 ½ years from construction start to ramp up to volume production." (cf. the report here)
Tight capacity at fabs does warrant concern for potential shortages, particularly when the present and forecasted demand can only be met by three to four foundries. The forecast from iSuppli for 2011 is that only TSMC, GlobalFoundries, Samsung Electronics, and possibly Intel will remain as foundries (see also our MarketWatch Commentary brief here). These foundries will have to support the major OEMs and smaller EMS businesses alike – that warrants concern for utilization allocation and timing concerns in today's competitive device landscape. As it is, look at the delays experienced even by Apple for its initial AT&T iPhone 4 release last year.
In the wake of the crisis in Japan, one business question is the direction of investment during rebuilding efforts. Will the new and/or repaired fabs and foundries return to the sizes they were serving, or make the move to 450mm? The new influx of investment capital, albeit for not the reasons we like to see, could speed these wafer size conversions.
The 'post-PC' supply chain?
Our generalized industry situation, both from the perspective of device convergence and from new component integration and innovation, was recently well summarized by Lee Kun-hee, Samsung's chairman. In this discussion (subscription required) with Financial Times about the competitive landscape facing Samsung in light of the iPad2 release, Mr. Lee offered that despite the success of his company's competitive device to Apple's iPhone, the Galaxy S smartphone, Samsung remains cautious:
warning his company could face ruin unless it diversifies into new fields such as medical equipment, pharmaceuticals and energy-efficient lighting. The big strategic question is whether the business culture that created Galaxy can be translated to these areas, into which he is investing $21bn over the next decade.
Why does this so well summarize our present industry situation? If Apple's iPad2 is a 'post-PC device,' as Steve Jobs coined during his keynote address unveiling the device, and if Samsung warns of ruin without diversification despite successful market penetration of their own device suite, then we need to take seriously the changes happening along our supply chain.
As discussed above, the costs of production from wafer on down are skyrocketing, and those costs are directly related to the major advances in the capabilities of the components. Today's components, particularly in leading-edge end-products (from phones to cars and used in diverse locations from sofas and desks to operating rooms), are very sophisticated in design, are highly integrated into system-on-chip designs and stacked ICs, are smaller and use less power, and are costly. But all this integration coupled with cost has whittled our supply chain to a smaller set of partners able to participate at these leading-edge levels (see this report on 'fabless and designless' supply chain directions from EETimes). Fewer manufacturers and contract manufacturers can mean trouble for supply (see this EETimes report on Apple and TSMC negotiations). With leading OEMs like Apple now squeezing margins for competitors with the iPad2 release, how many can continue to compete? (see this NYTimes report on price and tablet competition)
One upside: if competition is so squeezed from the tablet space, as is being discussed in the industry press recently, some are choosing to ramp up release dates, such as is rumored for HP's TouchPad (see this PCWorld article), while others are rumored to be pursuing more the PC-like aspects of their tablet devices. Might more of a close-out by Apple, actually push the competition to redesign notebooks rather than compete head-to-head with the iPad?
As mentioned above and in the introductory note to this Quarterly, as rebuilding takes hold in Japan to repair that which is repairable, will new business investments show a greater diversification of suppliers in number and location? Will there be greater attention paid to the complexities of massive and deep supply chain breaks, as has happened due to the terrible losses of life, business, and facilities in Japan? Disaster planning may be more at the forefront of management's strategic plans than it had been until this unfortunate turning point. This diversification will open streams to new and additional partners, providing growth opportunities within many sectors of the semiconductor supply chain.
Growth is certain but survival strategy is paramount
Does a post-PC device entail a post-PC supply chain or not? And, do new sectors, like automotive, demand specific changes to the established supply chain? What are the lessons we will learn from various disaster plans after the tragic events in Japan that have significantly affected so many people, businesses and supply chains?
From the perspective of end-product and market share, the question of device convergence versus cannibalization comes back into play. Is a tablet a device that will eradicate the PC market, or is it the next-generation PC device and therefore the next, innovative progression in a long line of revolutionary devices? Importantly, tablets' bill of materials (BOM) are strikingly similar in ASP and content to notebooks, as discussed thoroughly in the companion articles here and here in this issue of MarketWatch Quarterly, (cf. Credit Suisse Equity Research: Semiconductors 2011 Outlook […], 11 January 2011). In that case, the argument for a tablet to be a next step in PC innovation is logical. However, at the root of the wide-scale, global demand for tablets is the radical departure in design of tablets; these devices are based on smartphone designs, not on PC designs (see our discussion on this 'disruptive technology' here in a recent MarketWatch Quarterly).
But, as mentioned above, tablets are driving component innovation toward highly integrated chipsets, favoring multi-core, complex processors and ICs, MCUs and sensor interfaces with ICs on shrinking nodes to realize the demanded capabilities of these new, compact and lightweight devices. Many of these same innovations in chips are also being driven by other growth sectors, like automotive.
Whether post-PC or not, these high technology demands from the tablet and automotive markets are reminiscent of the technological innovations witnessed over the past many years within the PC sector. The sophisticated global supply chain developed for the PC-centric electronics world will need to adapt – but it will continue to function to support these exciting new sectors.
Who will be participating in these post-PC supply chains? How will the diversification of opportunities grow the semiconductor industry in light of growing demand for an increasing number of end-products with an increasing semi-content ratio? Finally the most uncertain question at the time of press for this Quarterly is what will the new supply chain ecosystem look like as rebuilding efforts begin in Japan. Will there be a tendency to diversify locations and number of suppliers to promote disaster management, or will there be a return to the track that our industry was on prior to March 11? Whatever comes of the tumult we are experiencing now, in the end, the rebuilding of people's lives in Japan is paramount to all else.