How Intel Lost the Mobile Market, Part 2: The Rise and Neglect of Atom

Update (6/1/2020): The report down below might have been published in 2016, but it continue to stands up as a postmortem of what went mistaken with Intel’s cell attempts — with 1 really crucial omission. Back again in 2016, we did not know Qualcomm had been ruthlessly imposing licensing and buying terms that designed it proficiently not possible for manufacturers to present Intel-based mostly cell equipment. I don’t forget pondering why Intel couldn’t locate a single US organization to create a telephone all around its hardware system for love or revenue when the initial Xolo X900 in comparison well sufficient towards a then-existing Iphone.

Intel continue to designed a amount of problems with Atom, as this report discusses, but the simple fact that Qualcomm had a stranglehold on the sector behind the scenes of course had an effect on what variety of achievement Intel was ever going to reach.

I genuinely appreciated the Xolo X900 device I tested all all those yrs in the past, and the Bay Path tablets I had circa 2013 were being wonderful equipment. Atom’s cell attempts will often continue being an attractive might-have-been.

Authentic tale down below.

In Part 1 of this two-element collection, we discussed the distinction among Intel and TSMC’s foundry designs and how these dissimilarities designed it really difficult for Intel to contend in cell. In Part 2 we’ll examine the precise choices Intel designed, the increase and neglect of Atom, and why the company’s excellent foundry technology was not sufficient to conquer the sector.

The typical clarification for why Intel dropped the cell sector is that its x86 cell processors possibly drew as well much electric power or weren’t highly effective sufficient in comparison with their ARM counterparts. Intel’s conclusion to promote its ARM division and XScale processor line in 2006 has been commonly derided as a essential error. It’s a simple, typical-perception clarification with just 1 flaw: It problems signs and symptoms for induce.

All of This Has Transpired Prior to

Intel’s struggles in the cell sector did not get started with Medfield, Moorestown, or even the conclusion to promote its ARM business enterprise and XScale chip division ten yrs in the past. As EETimes reported in 2006:

Intel (Santa Clara, Calif.) spent much more than $10 billion to enter the communications business enterprise in excess of the yrs, but the microprocessor giant dropped its shirt — if not tens of millions of dollars in the arena. The reported communications-chip sale is said to be element of Intel’s program to overhaul the organization. Intel is also established to contain the layoff or redeployment of 16,000 personnel, in accordance to speculation from 1 Web internet site.

Adjust “communications” to cell, modify the amount of fired personnel, and that paragraph could’ve been published today. Intel’s complications in cell aren’t new Santa Clara has been battling to enter new markets for almost 20 yrs. Other articles or blog posts from 2006 emphasize that XScale sales had been fairly small, as had revenue from Intel’s networking and communications division.

From Intel’s perspective, advertising XScale designed perception. Making a cell processor business enterprise all around ARM cores would have limited Intel’s ability to leverage its possess IP and know-how in x86 producing, even though concurrently reducing into its income (Intel would have owed considerable royalties to ARM if these kinds of a structure ever grew to become well-liked). Atom was previously well into growth in 2006 and Intel made the decision to wager on its possess hardware know-how and application growth capabilities.

Atom and the Increase of x86 Just about everywhere

Contrary to well-liked belief, Intel was not caught fully off-guard by the increase of smartphones or the popularity of smaller, Online-connected equipment. Atom growth started in 2004 the Silverthorne main that Intel debuted in 2008 had a TDP of just 2-3W at a time when most cell Core 2 Duo processors were being trapped in 35W territory.


Intel’s initial Atom.

Most people today don’t forget Atom as the chip that released a thousand netbooks, but that was not Intel’s initial program. The organization thought nettops and netbooks would be a market sector for Atom, not the chip’s main system. Atom and its successors were being supposed to launch an armada of Cellular Online Units, recognized as MIDs.


Units like the Gigabyte MID M528, proven earlier mentioned, glance hopelessly quaint today, but Intel was evidently thinking about the long term of cell computing. The organization envisioned an ecosystem of netbooks and MIDs pushed by its possess custom x86 architecture, a target the push dubbed “x86 in all places.”

Conflicting Priorities

There were being evidently executives at Intel who comprehended how essential cell would be to the company’s extensive-term long term and pushed for intense positioning and product ramps. Regrettably, all those attempts were being stymied by others who were being concerned about the effect Atom and the small-expense equipment it was supposed to enable would have on Intel’s main business enterprise. MIDs and afterwards netbooks were being supposed to be bare-bones, small-expense equipment, beneficial as secondary machines and for essential jobs, but no much more.


Intel was working on Atom SoCs in 2008, but its initially absolutely unified chip would not ship till 2012.

Intel’s put up-launch frame of mind in direction of Atom is finest summarized as benign neglect. Though the chip went by numerous revisions to integrate factors and reduce costs, Intel refused to commit the resources that would have designed Atom a finest-in-course player in the cell sector. From 2008 to 2013, Intel released a expense-lowered model of its Nehalem architecture, the Westmere 32nm die shrink, a new architecture with built-in graphics (Sandy Bridge), a large-stop fanatic system (Sandy Bridge-E), a new 22nm CPU with FinFET technology (Ivy Bridge), a further architectural refresh (Haswell), and a 2nd-technology fanatic system (Ivy Bridge-E). Which is two full tick-tock cadences for Intel’s massive-main business enterprise, even though Atom did not even make the bounce to 32nm till 2012. Its single architectural refresh to day arrived in 2013, just right after the launch of Ivy Bridge-E.

In spite of remaining to begin with starved for resources, 32nm Atom chips were being competitive in the midrange cell sector. With Medfield, Intel seemed to have turned a corner, but the company’s models normally failed to locate much traction in the sector. Only Intel’s contra-revenue technique gained the organization considerable pill sector share, and all those gains were being only sustained by large fiscal losses.


Intel’s cell and communications revenue and losses from Q1 2013 by Q2 2014.

Atom was not the difficulty — Atom was the solution Intel did not have the guts to chase.

The Tricky Conclusions Intel Did not Make

Intel failed to acquire traction in cell since it was not prepared to hazard upsetting the financial product that had transformed it into a titan of computing. The company’s fabs, producing procedures, and resources were being geared in direction of large, expensive processors, not churning out big numbers of small-expense cell cores. Prioritizing Atom in excess of Core would’ve essential the organization to retool at least some of its fabs to emphasize throughput and decrease costs in get to contend with the ARM processors constructed at Samsung and TSMC. It would’ve intended decrease gross margins and fewer revenue for every device marketed.

Intel's Atom Roadmap

Intel tried to pace issues up at some point, but both equally its 14nm system and the up coming-technology Goldmont CPU main were being delayed.

Intel did take steps to enhance its competitive standing vis-a-vis ARM and ARM’s foundry partners, but it rarely took them rapidly and typically failed to adhere to by. Intel acquired Infineon Wi-fi in 2011 for $1.4 billion, but to this working day all of its publicly introduced wireless solutions, which include the XMM 7480 modem, are continue to constructed on 28nm at TSMC. Smartphones and tablets have often applied SoCs, but Intel did not launch its initially Atom-based mostly SoC till 2012 — 5 yrs right after the Iphone released and 4 yrs right after Atom’s possess debut.

SoFIA integrated Atom SoC, details

The SoFIA partnership with TSMC lifted eyebrows, but not revenue.

One particular factor we want to tension here is that Intel’s conclusion to protect its main (Core) business enterprise and product margins might have been mistaken, but it was not ridiculous. Refitting fabs, building know-how in SoC structure, and porting modems from TSMC would have essential large funds infusions and take considerable quantities of time. If Intel had released Atom with an intense program to put the chip in smartphones by 2010, issues might have performed out really in a different way. By the time the organization woke up to the risk it faced from ARM and merchant foundries, it was as well late to make up the gap.

Why Intel’s Foundry Tech Could not Conserve Its Cellular Organization

Intel’s system technology management couldn’t help save the company’s cell division since it was not developed to do so. Smartphone and pill OEMs wanted equipment with built-in LTE radios Intel did not have them. Even the SoFIA partnership with TSMC never arrived to sector, seemingly since Intel couldn’t safe sufficient volume to kickstart output.

Intel’s 14nm complications delayed its up coming-technology pill processors from 2014 to 2015. Its 10nm node, when envisioned to safe great economies of scale in excess of TSMC, has been pushed to 2017 as well. I don’t assume these delays performed a big function in Intel’s conclusion to leave the cell sector, but they might have influenced it. In 2012, Intel continue to envisioned to be on 10nm by 2016 with EUV (extraordinary ultraviolet lithography) ramping in direction of full output. The now-canceled 450mm wafers weren’t envisioned in-sector very this shortly, but Chipzilla envisioned to recognize considerable expense cost savings from moving to the larger sized wafers in the 2018 to 2020 timeframe — expense cost savings that could’ve further enhanced its standing towards Samsung, TSMC, and GlobalFoundries.

If Intel had begun reorienting in direction of Atom when it released the chip in 2008, it might’ve weathered these delays and cancellations with no much trouble. Failing to do so remaining its beleaguered cell business enterprise experiencing better-than-envisioned costs and minimum revenue.

Intel did not get rid of the cell sector since Atom’s functionality and electric power consumption did not contend with ARM research and analysis confirmed that Atom was capable of matching ARM functionality in multiple sector segments. It dropped the cell sector since it did not make the improvements that would have permitted it to contend on expense with solutions created at TSMC and Samsung. The exacting regulations and unique layouts that drove Intel to the top rated of 1 sector could not be quickly adapted to others, and Intel was unwilling to hazard its placement at the top rated of the typical x86 sector for a dangerous payoff in cell. There is no proof that preserving XScale or developing ARM solutions would have adjusted that — if something, the ARM division would’ve been below even much more strain to be certain it never grew to become a risk to the x86 business enterprise.

Now read through:

Leave a Comment

Your email address will not be published. Required fields are marked *