How Intel knocked itself out of the smartphone chip market

Former CEO Paul Otellini admitted as much in a 2013 interview with The Atlantic. Intel has now bailed out of the smartphone chip market while Apple is flying high with its iPhones, based on its own A-series chips.

Intel has cancelled its upcoming Atom chip lines for smartphones, including Broxton and the Sofia 3GX, Sofia LTE and Sofia LTE2 commercial platforms. That decision ends close to a decade of futility with Intel trying to outmaneuver rivals like Qualcomm, Apple, and Samsung, which make mobile chips based technology licensed from ARM.

Intel still has a mobile strategy, but it's a major change in direction. The chipmaker is looking ahead to speedy 5G technology, which won't be limited to mobile devices. The technology will also be used in sensor devices, industrial equipment, cars, drones, and robots.

The long-term view of 5G will help Intel play a role in the fast-growing Internet of Things market, with Gartner estimating 20.8 billion connected devices to be in use by 2020. Intel is making baseband radio processors, communications equipment, back-end appliances, and field-programmable gate arrays (FPGAs) that will drive 5G deployments.

Intel will continue to ship existing Sofia and Atom tablet chips, code-named Cherry Trail, to device makers. Cherry Trail chips will be followed by Pentium and Celeron chips code-named Apollo Lake, which are targeted toward hybrids and laptops.

The Atom smartphone chip cancellations pull the curtain back on an ugly past in which Intel shot itself in the foot with bad timing and ill-advised executive decisions, analysts said.

In early to mid-2000s, under then-CEO Craig Barrett, Intel started laying down an end-to-end mobile and networking strategy, which included making networking equipment and mobile phone chips.

In a 2005 interview with IDG News Service, Barrett insisted Intel's mobile chips were popular with mobile phone makers. Intel's ARM-based StrongARM processors were a big part of its strategy, and the biggest competition at the time came from Texas Instruments' OMAP processors.

But Otellini, who replaced Barrett, deemphasized the mobile and networking strategy to focus on Intel's core PC market. Intel's PC offerings then came under pressure when AMD became a serious competitor with innovative chips, said Dean McCarron, principal analyst at Mercury Research.

As new CEO in 2006, Otellini made one of his biggest announcements when Apple switched Macs to Intel's x86 chips. Intel in late 2006 sold the StrongARM assets to Marvell for US $600 million.

Two major factors later changed Intel's view on smartphones -- the success of the iPhone, and how it cut into PC sales. Tablets then began to hurt PCs sales after the iPad was released in 2010, and those devices did not use x86 chips.

"That painted a picture of, 'oh there's a market that's not a PC market'," McCarron said.

Intel started retooling the Atom processor -- designed originally for netbooks -- to fit into Internet-connected devices like mobile Internet devices (MIDs), which were small-screen computers with Web browsing capability. Intel developed some prototype phablet-like devices based on a chip called Menlow, which was announced in 2008, but MIDs never took off.

At the time, Intel viewed the MID as a smaller version of a PC that was Internet enabled. It couldn't perceive growth in the smartphone market because of mobile broadband connectivity issues and data limits at the time.

PCs are "your expertise, that's your worldview, and that kept them from seeing the phone as a client," McCarron said.

The next Intel smartphone chip called Moorestown was announced in 2010 but was too power hungry for smartphones. The first Intel-based smartphone was Lava's Xolo X900, which was released in April 2012 in India, followed by handsets from Orange and Lenovo. Those handsets had a new Atom processor code-named Medfield.

While Intel struggled to make a competitive mobile processor, its failure in mobile software also burned the company. Intel in 2007 started working on Linux-based Moblin, which was merged with Nokia's Maemo into a new OS called Meego in 2010. The OS was then merged with LiMo into what is now Tizen. Intel ultimately warmed up to Android in 2011, but it was too late.

Intel was trying to force its OS down customers' throats as Apple and Google were taking over the market, said Jim McGregor, principal analyst at Tirias Research.

"They had a chance to jump ship to go with Android, but they didn't do it," McGregor said.

Intel also spent billions retooling chip manufacturing with mobile in mind, with the focus being on low-power chips instead of performance. The goal was to catch up with ARM in power efficiency by using its manufacturing advantage.

The investments didn't advance Intel in the mobile market but have helped bring longer battery life to laptops and tablets.

Another mistake was the high priority placed in the now-declining tablet market. Intel CEO Brian Krzanich, who replaced Otellini, set a goal to ship 40 million tablet chips by the end of 2014 using heavy subsidies on Atom chips. The company shipped 46 million chips that year, but the effort hurt Intel's profitability, and Krzanich decided not to repeat that strategy with smartphones. 

A move away from smartphone chips gives Intel a clean slate to envision an alternative, and more importantly, profitable mobile future. Intel's smartphone chips had thin margins, and now the company will focus on expensive server chips that will deliver Web services to mobile devices. Intel will also continue developing 5G modems.

Intel discontinued one mobile project almost a decade ago only to regret it, but this time, the company may have taken the right decision, especially with smartphone prices falling fast and the promise of 5G.

"I don't think it's shortsighted, it's realistic," McCarron said.