Atom chip cancellation puts Intel’s mobile processor plans on ice
Late on Friday night, Intel snuck out the news that it’s bailing on the smartphone market. Despite being the world’s best known processor maker, Intel was only a bit player in the mobile space dominated by Qualcomm, Apple, and Samsung, and it finally chose to cut its losses and cancel its next planned chip, Broxton. This followed downbeat quarterly earnings, 12,000 job cuts, and a major restructuring at a company that’s had a very busy April. Intel is still one of the giants of the global tech industry, but it’s no longer as healthy and sprightly as it used to be.
The bane of Intel’s existence for the past decade or so has been the transition to mobile computing. It wasn’t supposed to be that way. Having secured a commanding lead as the premier provider of desktop PC processors, Intel had a clear-eyed strategy for extending its dominance into the mobile realm.
A series of ignominious failures has left Intel reeling
With the help of Microsoft in 2006, Intel inaugurated the category of Ultra-Mobile PCs (UMPCs), which were the stylus- and touch-friendly precursors to today’s ultra-versatile tablets. They combined low-voltage Celeron and Pentium M chips with Windows Vista, and like everything else touched by Vista, they flopped. Unattainable pricing and inadequate battery life consigned the UMPC to the status of a historical footnote. The same fate befell Intel’s Mobile Internet Device (MID) initiative, which saw the chipmaker pushing and incentivizing its hardware partners to build mini internet tablets like the Nokia N810. Pervasive problems with affordability, battery life, clunky design, and ill-suited software prevented MIDs from ever becoming a mass-market success.
On the software front, Intel recognized the need for a tailored operating system to make the most of mobile PCs and sought to develop its own Linux variant titled Moblin. Moblin never convinced anyone outside of Intel, and was eventually merged with Nokia’s Maemo to produce MeeGo, which in turn merged with Samsung’s Bada and is now known as Tizen. Well, it’s only barely known, even by owners of its most successful product, the Gear S2 smartwatch. The series of post-Moblin software mergers has been merely the consolidation of repeated mobile failures.
Intel’s ventures into mobile hardware and software development show that even a great idea is only as good as its execution. The MIDs and UMPCs of yesteryear were aimed at the same usage scenarios as the phablets and pro tablets of today — but they were compromised and premature, and therefore rejected by the market.
This has cost Intel dearly, with the company lavishing billions on developing suitable processors and modems to put into its various mobile undertakings. The multibillion-dollar mobile costs have spiraled in recent years — a loss of $3.1 billion in 2013 was followed by a loss of $4.3 billion in 2014 — which eventually forced Intel to combine its mobile and PC earnings reports in order to disguise its unproductive spending.
The tragedy of Intel’s mobile failure is that the company foresaw all the threats to its business and acted to preempt them. It just didn’t do so very well. That being said, Intel’s the victim of its bad decision making almost as much as its poor execution.
Favoring WiMAX over LTE was a historically bad decision
One of the fateful choices that Intel made around 2009 was to commit itself to WiMAX as the 4G standard of the future. Qualcomm went the other way, prioritizing LTE instead, and now the latter has a significant lead in designing and integrating LTE modems, while the former is scrambling and struggling to catch up. The total defeat of WiMAX almost wiped out Sprint, its biggest US purveyor and advocate, and it put Intel on the backfoot in adopting the true 4G standard of the future, which turned out to be LTE. At this point, even if Intel were to double its already vast spending, bridging the gap of years of research, development, and experience would be practically impossible.
In spite of its unhappy mobile history, Intel persisted in trying to compete because it knew how central mobile devices were becoming to our lives. Last year, its Atom processors even looked like they had a shot at denting Qualcomm’s market dominance, thanks in large part to the Qualcomm Snapdragon 810 chip’s power and heat issues. There was a small opening, but this year’s Snapdragon 820 is an absolute beast that conclusively shuts the door on any further Intel incursions.
The top three smartphone vendors — Apple, Samsung, and Huawei — each produce their own processors. At Mobile World Congress this year, Xiaomi, another large-scale smartphone maker, co-branded its launch event with Qualcomm. And global names like LG, HTC, and Sony basically only shop at the Snapdragon aisle for their flagship phones. Intel’s most loyal hardware partner is Asus, which makes a habit of announcing interesting new devices at Computex in June and not shipping them until the end of the year. The most feted Intel Atom-powered smartphone to date is probably the ZenFone 2, a distinction that speaks for itself.
Without any unique advantages to its Atom CPU line and no captive market like it has on the desktop, Intel is right to bow out of the smartphone processor race. It’s a merciless competition that has already ousted big names like Nvidia and Texas Instruments, and Intel will be better off figuring out different parts of the mobile computing world where it can participate profitably. CEO Brian Krzanich put „the cloud and data center“ first atop a list of Intel’s new priorities in a recent blog post, reiterating the idea that the company will transition to facilitating connectivity as its main area of competence. Discrete Intel LTE modems will still be around, and the company seems to think it can recapture its mobile competitiveness by being a leader in the adoption of the incoming 5G wireless standards. To that end, Intel doesn’t intend to kill off Atom entirely, and still plans to offer a chip for tablets later this year, codenamed Apollo Lake.
Even Moore’s Law is hitting a wall
To its credit, Intel has always operated under the assumption that mobile computing will eventually supplant the desktop and consign the old PC boxes to niche-use status. We email on our phones, ideate on our phablets, and write and create on our tablets — as Steven Sinofsky, former boss of Windows, recently articulated with respect to the iPad Pro. The primary form of personal computer is changing, which is why ultrabooks and hybrid laptops are so prominent in Intel’s marketing and development efforts. The low-power Core M, Intel told me last year, was the most important variant of the Skylake processor family, and the company’s ongoing mission is to move with its users to more portable form factors.
Intel remains a diverse and strongly profitable company. There will always be PC gamers and video producers looking for the latest and fastest CPU. But while the core business that’s kept Intel going for so many years isn’t disappearing, its importance and primacy are being steadily eroded by the insatiable growth of mobile computing. Even Moore’s Law, the Intel co-founder’s prediction about the constant growth of processing power in chips, is hitting a wall now. Intel’s desktop CPUs are being pushed further back on the roadmap while some of its mobile ones are being deleted entirely.
It’s an uncertain future for what used to be one of the most assured companies in tech.