Intel (INTC -2.40%) recently announced it will stop investing in SoFIA, its Atom-based mobile SoCs for low-end Android devices, and Broxton, its Atom chipset for high-end smartphones and tablets. This move marks a big retreat from the mobile market, but it wasn't surprising -- Broxton was already pushed back from its original launch date last year, and certain SoFIA chips had also been delayed.
Image source: Intel.
Intel isn't completely abandoning tablets, where Atom chipsets have gained some traction. Instead, it plans to pitch Pentium and Celeron chips based on its new Apollo Lake platform to low-end PC and tablet makers. But in smartphones, it's clearly thrown in the towel. The only toehold Intel will have left in the market is a partnership with Chinese chipmaker Rockchip for installing SoFIA SoCs in low-end phones, but now that Intel has stopped investing in SoFIA, that partnership will inevitably wither.
$10 billion wasted for a 1% market share
Intel spent more than $10 billion over the past three years to gain mobile market share against ARM (ARMH) licensees like Qualcomm (QCOM -2.36%) and MediaTek. Its strategy was to subsidize OEMs with generous "contra revenues," which included deep discounts on Atom chips, co-marketing agreements, and financial assistance in redesigning logic boards.
However, that costly strategy only attracted a few second-tier smartphone makers, and Intel processors powered just 1% of the world's smartphones at the end of last year. Meanwhile, industry benchmarks showed that high-end Atom chips like the Atom Z3580 simply couldn't compete against Qualcomm's Snapdragon 810 and Samsung's (NASDAQOTH: SSNLF) Exynos 7420.
Samsung's formidable Exynos 7 processor. Image source: Samsung.
Intel's attempts to turn around the mobile business were haphazard. CEO Brian Krzanich poached high-profile executives from Qualcomm, which reportedly caused internal conflicts at the insular company. CFO Stacy Smith implemented a plan to boost mobile profitability by reducing contra revenues and the cost of producing chips. But since that move meant fewer perks for OEMs, major partners like Asus started using Qualcomm chips instead. Mobile chief Aicha Evans resigned last month, but subsequently withdrew her resignation for unknown reasons.
Losing the 4G battle, preparing for 5G war
After ditching SoFIA and Broxton, which were 4G chipsets, Intel plans to focus its mobile investments on 5G technologies instead. 5G networks, which are expected to start being deployed around 2020, could offer 100 times faster data transfers than 4G networks. By investing in 5G tech, Intel could offer well-developed modems and chips to effectively compete against Qualcomm and other ARM licensees.
Intel's focus on 5G also explains Evans' return. Evans is an expert in 5G technologies and previously outlined Intel's strategy to tap into that upcoming market's growth. Intel didn't make much progress in mobile application processors, but its LTE modems will likely be installed in a portion of Apple's (AAPL -1.22%) iPhones this year, which marks some progress against Qualcomm. If Intel can follow up that rumored victory with more design wins in modems, it would establish solid foundations for 5G growth.
As for the Internet of Things (IoT), Intel believes the number of connected devices worldwide will soar from 15 billion in 2015 to 200 billion by 2020. To tap into that growth, Intel could pitch new versions of Atom chips for smart appliances, wearables, connected cars, or industrial robots -- which would enable its connected chip business to expand well beyond smartphones and tablets. That growth would complement Intel's new "streamlined" push into programmable chips, memory, and data centers.
Intel's Curie (L) and Edison (R) IoT modules. Image source: Intel.
But isn't Qualcomm doing the same thing?
Intel is spinning its retreat from mobile into an investment in 5G, but the chipmaker certainly isn't entering the 5G market unopposed. Qualcomm has already outlined its 5G plans, and is aggressively expanding into non-mobile markets like connected cameras, drones, and cars. Qualcomm also spent $2.4 billion to buy CSR, a maker of IoT and automotive chips, to accelerate its growth in that market last year.
Therefore, Intel's abandonment of smartphones and its renewed focus on 5G isn't a "pre-emptive strike" on Qualcomm and other chipmakers. Instead, it's an attempt to get its act together before it loses another key market to ARM's licensees.
