Written by Lisa Ann Cairns, Ph.D.
The diversification and customization of components, software and device designs is expanding, a trend rooted in the ending of Moore's Law. A clear example of this trend is found in the rise of the "Internet of Things." The recent CES 2014 revealed component architectures and packaging solutions that have embedded "Smart" into a rapidly-rising number of "Things."
Moving into 2014, there are a number of developments in the semiconductor and electronics industry that hold promise for solid, long-term growth. Perhaps the most important is the broad-sweeping diversification that is underway. This diversification is combining the dual design aspects of component innovation and user experiences. Additionally, this diversification reaches across market sectors to support new growth that will move us beyond traditional computing and mobile devices to fuel a broad-based and pervasive set of growth opportunities.
As we saw in real devices and prototypes alike at the International CES 2014, the diversification trends range from specific shifts in consumer electronic (CE) devices to the realization of the Internet of Things (IoT), which opens new markets, verticals, device types, and applications for semiconductor expansion. In the case of CE, there is a two-part diversification: the spread of heterogeneous computing in the architectural hardware designs for devices, and the spread of "smart" computing to meet users' demands of their devices for a seamless, connected, multitasking, and integrated experience. In parallel with CE diversification is the expansion of machine-to-machine (M2M) connectivity that is IoT. This IoT diversification branches across markets, devices, and use cases, furthering ubiquitous computing. Just as the architectural diversification of devices supports smarter CE devices, the M2M expansion supports the realized IoT by extending so many "Things" and increasing the number of interactive ("smart") devices in our daily lives.
Moving from Moore to more
Today's diversification trends are rooted in a critically disruptive event in the semiconductor and electronics industry, a dramatic change to the very tenets and trajectories of design and manufacturing for decades: the ending of Moore's Law. We have seen this point coming for a few years, but 2014 will likely mark the real shift away from keeping up with Moore's Law (i.e., the previously-held, long-term tenet defined by Gordon E. Moore stating that the number of transistors on an IC would double roughly every twenty months). Moore's Law has been a mainstay of research and development (R&D) and design within the semiconductor industry since the mid-1960s, when Gordon Moore codified his descriptive forecast for ICs.
Although Moore's Law was never intended to hold for more than a decade or so, the industry adopted the tenet and integrated the every-other-year pattern of new rollouts into the strategic R&D fabric of innovation. Now, with multiple challenges set to end Moore's Law (particularly physical and material constraints, skyrocketing fab equipment costs, and equipment and testing limitations, among other roadblocks), it is time to look beyond and perhaps be freed from the (unintentional) constraints of Moore's Law. The new opportunities beyond Moore were well-put by Henry Samueli, co-founder of Broadcom, in an interview with EETimes' Rick Merritt:
From our [Broadcom] perspective [this situation] gives us more breathing room to be clever about design. Most people run to the current process node as fast as they can. That's going to change. Instead of running to the next node, you will come up with new architectures and circuit designs, and that will create more opportunities on the design side, bringing more value to a company like Broadcom.
Samueli's message is important, and has been echoed across the industry over the past year-plus. The costs of the shrinking architectures and increasing complexity of ICs are reaching a point where it is not really feasible to continue from an ROI perspective. Already, we have seen the number of chip manufacturers and foundries dwindle because the only way to keep up has been to consolidate or to exit. It is no longer possible for new foundries and fabs to compete, due to the insurmountable barriers to entry, namely the pure cost of facilities and equipment as well as extremely sophisticated engineering and testing.
The design complexity and shrinks have brought us to a critical point in the industry, a point from which – instead of trying to continue innovation solely on the one path that is Moore's Law – we can look to explore new device and use cases with all that has been architected, manufactured, and put into successful designs. In other words, we can now be freed to vertically explore how the expansive component set we have can be configured to realize new device and computing opportunities across many market sectors and use cases. In many ways, this freedom to diversify was at the crux of Intel CEO Brian Krzanich's keynote that opened CES 2014.
Component diversification & integration at the core
Looking beyond Moore, there is quite a bit of growth and excitement ahead. Design architectures and the latest challenges of IoT and expansive component reach are all poised to really take off starting this year. "Smart" everything is the keyword that captures this diversification shift, imbuing "smart" into a rapidly-rising number of "Things." This was, of course, well exhibited at CES 2014 through an impressive new array of devices that can be worn and embedded, and will allow for new connectivity between devices (from appliances to smart devices and cars). What this means is that the semiconductor and electronics industry's next growth wave will come not from the traditional Moore's Law increments, but instead from the push into new use cases where we have previously had little penetration. Creating a "Smart Life" and "Smart City" means creating the right admixture of components and device designs to support the growing interconnected computing that users are demanding and expecting.
The "Smart Life" and IoT are still in early stages – that means there are both growth but also growing pains ahead. Meaningfully connecting devices, appliances, vehicles, wearable devices, and other aspects of our lives into a coherent and seamless user experience is a real challenge, from the component, networking, and software perspectives. For components, the challenges are exciting and not minimal: handling increasingly large data sets over increasingly crowded networks from an increasingly large number of devices, and having a way to enable seamless multitasking by users through a shifting set of devices, such as smartphones, tablets, PCs, internet TVs, and others as the users move through the "smart" environment. Obviously, there is the challenge of interoperability, which is both a hardware and software (and network) issue that must be overcome. Similarly, as Frost & Sullivan notes, "Reliability and performance issues will need to be addressed to allow the potential of IoT to be realized."
One of the bigger changes we are seeing among the major chip manufacturers, such as Intel, is the strategic diversification into more communication systems and mobile processors in order to meet the component demands for smart wireless devices that are both in highest demand and also driving the wider IoT shifts. Low-powered Systems-on-Chip (SoCs) are what mobile hot devices demand, and the competition wins, therefore, center around these component designs, an area where some, like Intel, were not previously competing as earnestly as ARM's designs. In response, in 2013 Intel's Atom chip had significant wins, and there is positive anticipation for the Bay Trail rollout for tablets with 64-bit architectures in 2014. At CES 2014, Intel also made a real splash with its latest Quark announcements supporting options such as dual-OS for computing and the related Edison's small SoC form factor to support wearable device innovation.
The processor wars have characterized 2013, and the anticipation is that 2014 will see a next level of design and application/device wins that will increase the various SoCs, System-in-Package (SiP), and Package-on-Package (PoP) options in the market. Importantly, the proliferation of SoC, SiP, and PoP architectures further the innovation and design diversification that support the ongoing market and device diversification that we see. In other words, what is happening in the component market and what is happening in the system and device architectures are in synchronicity, each supporting and driving important innovation and growth for the industry. We expect, therefore, that 2014 and moving forward will be ripe with new device types, applications, and use cases moving semiconductors into the once less-tapped markets and verticals. The broad-based demand and growth should, in turn, also support inventory control and average selling price (ASP) levels that can support continued steady industry revenue growth. Currently, semiconductor industry growth is estimated to continue on a steady "[…] 3-6% [pace] for the next 3-5 years, which is a good sign."
Last year's International CES 2013 brought core components to center stage more than we'd seen previously. Given the diversification of devices across market sectors, not only have we seen CES 2014 showcasing the latest smart wireless CE devices (especially across the categories of wearables, automotive, and Smart Life), but, importantly, CES 2014 also presented the latest component architectures and packaging solutions that support these new device design builds. Of course, one of the questions in the industry is whether or not we will eventually see chip and packaging integration to the level of a single package for single devices, as discussed in the Spring 2013 issue of Chip Design Magazine, but Intel's Edison and ARM’s IoT solution ideas definitely an important step in that direction.
Beyond Moore's Law, one other shift in the component market that bridges the end-device, CE market is the move to 64-bit processors for smart devices, first introduced by Apple in the newest iPhone 5s in 2013. At CES 2014, we saw more high-end smartphones and tablets with 64-bit processors featured, particularly leveraging the increasingly sophisticated GPUs that are reaching CPU levels to support the gaming sector. It has been argued that 64-bit, quad-core architectures for smartphones are, at best, a chicken-and-egg moment, because there really is not the software available for these smartphones to make full use of these complex cores. However, what we do see here is an example of the innovative trend that can promote different solutions, but that must be balanced against cost and market strategies, as argued by Mark LaPedus in Chip Design Magazine earlier in 2013:
In reality, however, there is no one-size-fits-all technology. As the market continues to splinter into various sub-segments, such as entry-level phones, smartphones, superphones and tablets, there is room for several different architectures and technologies.
As LaPedus continues, what we will likely see increasing in 2014 is that OEMs may step away from the traditional increments in processing power and similar "overkill for today's systems." Instead, they will have to rethink the balance between components and features for cost management – between software capabilities and design and use case options that are more likely to drive demand. In this balancing, we can again see the diversification shift where the next-generation technologies and devices are more likely to consider the balancing of components and software to deliver the best experiences than to simply increase processing power or other component increases more in accordance with traditional Moore's Law step-wise increments.
Consumerization of Smart spreads ubiquity and growth
The consumerization of Smart CE is likely to be a prime theme beginning in 2014; analysts across the spectrum are heralding that this next expansion for CE devices will necessitate the inclusion of a different type of designer for next-generation devices. As Apple has clearly demonstrated, the elegance of the form of the device is as important as its functionality. One of the constraints on the adoption of today's wearables, such as smartwatches, has been the issue with the design of the devices themselves. As we already saw at CES 2014, form and function are a real topic and concern for real CE adoption.
While there may be compelling component architectures and a general sense of form and functional style, overall, there has not yet been much design-appeal in new wearables. For example, to date, many smartwatches have not done as well as we'd have thought. One important reason is that smartwatches have been designed for larger wrists, typically for adult males. This oversight has left out not only the other half of the adult population – women – but also the critical youth demographic where turnover and new device purchases happen on a faster cycle than the older adult populations. Core architectures and feature capabilities are essential to drive demand in next-gen devices, but, as more of these devices will likely fit into the wearable or accessory category and will be intricate parts of users' smart life, being able to express the user's fashion sense and personality will be increasingly important. The crossover and collaboration of device R&D, including fashion and engineering design, would drive demand for these devices beyond what we've seen. It is highly anticipated that CES 2014 will (hopefully) set the stage for the start of more fashionable and customizable smart devices.
But the realm of smart device expansion is not just about improving fashion for wearable CEs. Tomorrow's smart life is expanding through today's increased adoption and demand for smarter devices for most aspects of our life. No longer content to limit our smart devices to the PC, we moved on to tablets and smartphones and "superphones." 2014 is certain to be the real launch into a smarter life, into ubiquitous computing. The Internet of Things (IoT) expansion is where we will be seeing the diverse component solutions proliferate and penetrate a much wider set of markets than before. While TVs, some major appliances, and our indoor thermostats and smoke detectors have adopted smart architectures, the opportunities to expand the number of "Things" that we can "make smart" will provide opportunities for a different type of design innovation, the one beyond Moore's Law, where new applications of existing chip designs can be leveraged to increase return on investment (ROI), continue to control the inventory balances we've reached, and improve user experiences in new and innovative ways.
All in all, as we move into this post-Moore era, one of ubiquitous computing and IoT, or – as some have called it – the newest technological revolution/era, diversification and customization from components to software and device form-and-functional designs are set to expand. The new opportunities across components and markets for the semiconductor and electronics industry support a healthy growth forecast not just for our industry, but also in a greater "industrial revolutionary" realm that can support the realization of IoT.