Intel giving up on the smartphone battle to focus on the looming 5G chipset war
On April 19th Intel announced a restructuring initiative aimed at accelerating the company’s shift from a PC driven business to a business which connects the cloud with the increasingly large base of smart or connected devices. As part of the company’s pivot, areas including memory and programmable solutions such as field programmable gate arrays (FPGAs), 5G and other connectivity technology, and the continuation of Moore’s law are key points of emphasis.
Intel will shed up to 12,000 employees globally as part of the restructuring effort with many of the actions expected to be communicated within the next 60 days. The company expects about $750 million in cost savings in 2016 with annualized savings reaching $1.4 billion by the end of next year. Last week in the wake of the announcement several media outlets began to report that Intel will be ceasing its development of the Atom SoC aimed at smartphones and media tablets. This includes generations of Intel’s “smart or feature phone on Intel architecture” (SoFIA) products, the low cost integrated modem/application processor targeted toward entry level smartphones and tablets as well as the Broxton applications processor which was meant for the high-end or premium smartphone market.
Not gone or forgotten
Although Intel’s shift appears to de-emphasize the smartphone and media tablet markets, the company continues to hold interest there through partnerships with or investments in Chinese semiconductor vendors such as Rockchip and the Spreadtrum subsidiary of Tsinghua Unigroup. Intel partnered with Rockchip in the spring of 2014 to accelerate development of its SoFIA product for the entry level tablet market. In the fall of that year Intel agreed to invest up to $1.5 billion in Spreadtrum for what could be a 20% stake in the company. During the course of 2015 Intel had completed the first phase of the agreement investing $966 million with $500 million required in the second phase pending regulatory approval and other closing conditions. SoFIA SoC are fabricated at merchant foundry TSMC, and are not reliant upon Intel’s manufacturing processes the way x86 processors are. Thus the potential still exists for SoFIA 3G and LTE solutions to target entry level media tablets and mobilized computing devices even though Intel is shifting focus away from these products.
Although efforts surrounding SoFIA are winding down, Intel plans to continue participating in the standalone or thin-modem market where it had some successes with OEMs such as Samsung in the past and expects traction in the future. The company has stated that it expects the XMM7360 (3rd generation LTE thin modem) to ship in volume in the second half of 2016. Apple has been all but confirmed to have initiated a dual source strategy for the modem in some models of the next iPhone. Intel is also continuing to invest in technology for the product segment. As recently as October 2015 the company demonstrated its commitment to the thin-modem market by acquiring Via Telecom’s CDMA assets, strengthening its position in markets such as North America and China where CDMA capabilities can be especially important. Combined with its thin-modem efforts Intel’s Curie platform will remain a critical piece of its strategy as well with focus shifting to the IoT.
The giant’s head turns slowly
Intel’s shift from mainstream smartphone and media tablet products began well before last week’s revelation that the company would pivot away from its Atom products in these markets. Intel’s larger mobile communications efforts were initiated in 2011 with the purchase of Infineon’s WLS business. From 2010 (the year prior to the larger IMC effort) to 2013 Intel’s annual research and development (R&D) expense increased by 61% from $6.58 billion to $10.61 billion, driven at least in part by Intel’s focus on smartphone and tablet products. Intel began to reduce its R&D investment in these markets in 2014 although overall research and development spending has continued to rise due to increasing process technology development cost and increased investments in server, Internet of Things, and new device product categories. Intel’s R&D expense was $12.13 billion in 2015, an increase of 84% since 2010 as it continued to shift focus on the product categories which are core to its business strategy or have the most potential to drive future revenue growth.
As Intel moved away from investing in smartphone and tablet products operating losses in its mobile communications group mounted, accumulating to over $9 billion between 2012 and 2014 before the company re-organized the segment under the client computing group last year. After such disappointing results in the space Intel management is likely looking forward to leaving the mainstream smartphone and tablet market behind with an eye toward the future.
Intensifying the focus
The advance to 5G will be very different compared to the development of 4G cellular technology. On the road to 4G there were two different paths, the first was WiMAX, a technology pre-dating LTE and with a shorter lifespan in which Intel had invested. While backing the wrong technology may have been the beginning of a long and challenging 4G journey for Intel, it shouldn’t face the same decision with 5G. The next generation of mobile wireless technology will be more unified than 4G was with the 3GPP providing common industry backing and acting as a ratification body.
While the commercial reality of 5G is still several years away, Intel’s involvement at the earliest stages represents a fresh opportunity for the company. By continuing its efforts in the modem space and exiting the highly competitive smartphone and tablet markets with the split from Atom, Intel should be able to tune its focus to compete more effectively with market leader Qualcomm. While 5G modems will require new chipset innovations the company’s 3G/4G efforts will continue to add value providing backward compatibility to new products. Shedding the Atom efforts in smartphones and tablets may provide Intel the ability to be more acquisitive while developing 5G technology. Purchasing capabilities in the RF front-end may be appropriate as this could provide Intel the ability to offer more complete solutions to its customers when 5G commercialization arrives.
Intel management recently highlighted a “core belief” that economic principles of Moore’s Law will continue as process technology moves from 14nm to 10nm and below. Regardless of if the law continues to hold true, Intel should be able to capitalize on its semiconductor manufacturing capabilities to optimize scale and cost in the forthcoming 5G era. Therefore, Intel is taking a long term view to win the larger chipset war looming with 5G by making a tactical retreat on a losing battlefront of smartphone and tablet SoCs.