Replace (11/25/2020): The article under might have been written in 2016, nevertheless it nonetheless stands up as a postmortem of what went incorrect with Intel’s cellular efforts — with one crucial omission. Again in 2016, we didn’t know Qualcomm had been ruthlessly implementing licensing and buying phrases that made it successfully inconceivable for producers to supply Intel-based cellular gadgets. I bear in mind questioning why Intel couldn’t discover a single US firm to supply a telephone round its platform for love or cash when the unique Xolo X900 in contrast properly sufficient towards a then-current iPhone.
Intel nonetheless made numerous errors with Atom, as this text discusses, however the truth that Qualcomm had a stranglehold available on the market behind the scenes clearly had an impression on what sort of success Intel was ever going to attain.
I genuinely preferred the Xolo X900 gadget I examined all these years in the past, and the Bay Path tablets I had circa 2013 have been nice gadgets. Atom’s cellular efforts will all the time stay an attractive might-have-been. The US third Circuit Courtroom of Appeals has since thrown out the antitrust findings towards Qualcomm. It needs to be famous that the FTC disagreed with that call. Intel continues to fabricate 5G base stations, however the scope of its 5G enterprise has shrunk considerably after promoting its 5G modem to Apple.
Authentic story under:
In Half 1 of this two-part collection, we mentioned the distinction between Intel and TSMC’s foundry fashions and the way these variations made it extraordinarily tough for Intel to compete in cellular. In Half 2 we’ll discover the particular choices Intel made, the rise and neglect of Atom, and why the corporate’s superior foundry know-how wasn’t sufficient to beat the market.
The common explanation for why Intel misplaced the cellular market is that its x86 cellular processors both drew an excessive amount of energy or weren’t highly effective sufficient in contrast with their ARM counterparts. Intel’s choice to promote its ARM division and XScale processor line in 2006 has been extensively derided as a crucial error. It’s a easy, commonsense clarification with only one flaw: It errors signs for trigger.
All of This Has Occurred Earlier than
Intel’s struggles within the cellular market didn’t start with Medfield, Moorestown, and even the choice to promote its ARM enterprise and XScale chip division ten years in the past. As EETimes reported in 2006:
Intel (Santa Clara, Calif.) spent greater than $10 billion to enter the communications enterprise through the years, however the microprocessor big misplaced its shirt — if not thousands and thousands of dollars within the area. The reported communications-chip sale is claimed to be a part of Intel’s plan to overtake the corporate. Intel can also be set to incorporate the layoff or redeployment of 16,000 staff, based on hypothesis from one Website.
Change “communications” to cellular, alter the variety of fired staff, and that paragraph may’ve been written right now. Intel’s issues in cellular aren’t new; Santa Clara has been struggling to enter new markets for practically 20 years. Different articles from 2006 emphasize that XScale gross sales had been pretty low, as had income from Intel’s networking and communications division.
From Intel’s perspective, promoting XScale made sense. Constructing a cellular processor enterprise round ARM cores would have restricted Intel’s potential to leverage its personal IP and experience in x86 manufacturing, whereas concurrently reducing into its earnings (Intel would have owed vital royalties to ARM if such a design ever grew to become fashionable). Atom was already properly into growth in 2006 and Intel determined to guess by itself experience and software program growth abilities.
Atom and the Rise of x86 In all places
Opposite to fashionable perception, Intel wasn’t caught fully off-guard by the rise of smartphones or the recognition of small, Web-connected gadgets. Atom growth started in 2004; the Silverthorne core that Intel debuted in 2008 had a TDP of simply 2-3W at a time when most cellular Core 2 Duo processors have been caught in 35W territory.
Intel’s unique Atom.
Most individuals bear in mind Atom because the chip that launched a thousand netbooks, however that wasn’t Intel’s unique plan. The corporate thought nettops and netbooks could be a distinct segment marketplace for Atom, not the chip’s main platform. Atom and its successors have been supposed to launch an armada of Cellular Web Gadgets, often known as MIDs.
Gadgets just like the Gigabyte MID M528, proven above, look hopelessly quaint right now, however Intel was clearly desirous about the way forward for cellular computing. The corporate envisioned an ecosystem of netbooks and MIDs pushed by its personal customized x86 structure, a purpose the press dubbed “x86 all over the place.”
There have been clearly executives at Intel who understood how crucial cellular could be to the corporate’s long-term future and pushed for aggressive positioning and product ramps. Sadly, these efforts have been stymied by others who have been involved concerning the impression Atom and the low-cost gadgets it was purported to allow would have on Intel’s main enterprise. MIDs and later netbooks have been purported to be bare-bones, low-cost gadgets, helpful as secondary machines and for fundamental duties, however no extra.
Intel was engaged on Atom SoCs in 2008, however its first absolutely unified chip wouldn’t ship till 2012.
Intel’s post-launch perspective in the direction of Atom is greatest summarized as benign neglect. Whereas the chip went by way of a number of revisions to combine parts and cut back prices, Intel refused to commit the sources that may have made Atom a best-in-class participant within the cellular market. From 2008 to 2013, Intel launched a cost-reduced model of its Nehalem structure, the Westmere 32nm die shrink, a brand new structure with built-in graphics (Sandy Bridge), a high-end fanatic platform (Sandy Bridge-E), a brand new 22nm CPU with FinFET know-how (Ivy Bridge), one other architectural refresh (Haswell), and a second-generation fanatic platform (Ivy Bridge-E). That’s two full tick-tock cadences for Intel’s big-core enterprise, whereas Atom didn’t even make the soar to 32nm till 2012. Its single architectural refresh up to now arrived in 2013, simply after the launch of Ivy Bridge-E.
Regardless of being initially starved for sources, 32nm Atom chips have been aggressive within the midrange cellular market. With Medfield, Intel appeared to have turned a nook, however the firm’s designs typically failed to seek out a lot traction available in the market. Solely Intel’s contra-revenue technique gained the corporate vital pill market share, and people beneficial properties have been solely sustained by way of heavy monetary losses.
Intel’s cellular and communications income and losses from Q1 2013 by way of Q2 2014.
Atom wasn’t the issue — Atom was the resolution Intel didn’t have the center to chase.
The Powerful Choices Intel Didn’t Make
Intel failed to achieve traction in cellular as a result of it wasn’t keen to danger upsetting the financial mannequin that had reworked it right into a titan of computing. The corporate’s fabs, manufacturing methods, and sources have been geared in the direction of giant, costly processors, not churning out big numbers of low-cost cellular cores. Prioritizing Atom over Core would’ve required the corporate to retool no less than a few of its fabs to emphasise throughput and decrease prices to be able to compete with the ARM processors constructed at Samsung and TSMC. It could’ve meant decrease gross margins and fewer revenue per unit bought.
Intel tried to hurry issues up finally, however each its 14nm course of and the next-generation Goldmont CPU core have been delayed.
Intel did take steps to enhance its aggressive standing vis-a-vis ARM and ARM’s foundry companions, nevertheless it hardly ever took them shortly and infrequently didn’t observe by way of. Intel purchased Infineon Wi-fi in 2011 for $1.four billion, however to today all of its publicly introduced wi-fi merchandise, together with the XMM 7480 modem, are nonetheless constructed on 28nm at TSMC. Smartphones and tablets have all the time used SoCs, however Intel didn’t launch its first Atom-based SoC till 2012 — 5 years after the iPhone launched and 4 years after Atom’s personal debut.
The SoFIA partnership with TSMC raised eyebrows, however not income.
One factor we need to stress right here is that Intel’s choice to guard its core (Core) enterprise and product margins might have been incorrect, nevertheless it wasn’t loopy. Refitting fabs, constructing experience in SoC design, and porting modems from TSMC would have required giant money infusions and take vital quantities of time. If Intel had launched Atom with an aggressive plan to place the chip in smartphones by 2010, issues might need performed out very in a different way. By the point the corporate woke as much as the menace it confronted from ARM and service provider foundries, it was too late to make up the hole.
Why Intel’s Foundry Tech Couldn’t Save Its Cellular Enterprise
Intel’s course of know-how management couldn’t save the corporate’s cellular division as a result of it wasn’t designed to take action. Smartphone and pill OEMs wished gadgets with built-in LTE radios; Intel didn’t have them. Even the SoFIA partnership with TSMC by no means got here to market, apparently as a result of Intel couldn’t safe sufficient quantity to kickstart manufacturing.
Intel’s 14nm issues delayed its next-generation pill processors from 2014 to 2015. Its 10nm node, as soon as anticipated to safe huge economies of scale over TSMC, has been pushed to 2017 as properly. I don’t suppose these delays performed an enormous function in Intel’s choice to go away the cellular market, however they might have influenced it. In 2012, Intel nonetheless anticipated to be on 10nm by 2016 with EUV (excessive ultraviolet lithography) ramping in the direction of full manufacturing. The now-canceled 450mm wafers weren’t anticipated in-market fairly this quickly, however Chipzilla anticipated to acknowledge vital value financial savings from shifting to the bigger wafers within the 2018 to 2020 timeframe — value financial savings that might’ve additional improved its standing towards Samsung, TSMC, and GlobalFoundries.
If Intel had begun reorienting in the direction of Atom when it launched the chip in 2008, it would’ve weathered these delays and cancellations with out a lot bother. Failing to take action left its beleaguered cellular enterprise going through higher-than-expected prices and minimal income.
Intel didn’t lose the cellular market as a result of Atom’s efficiency and energy consumption didn’t compete with ARM; analysis and analysis confirmed that Atom was able to matching ARM efficiency in a number of market segments. It misplaced the cellular market as a result of it didn’t make the adjustments that may have allowed it to compete on value with merchandise manufactured at TSMC and Samsung. The exacting guidelines and distinctive layouts that drove Intel to the highest of 1 market couldn’t be simply tailored to others, and Intel was unwilling to danger its place on the high of the standard x86 marketplace for a dangerous payoff in cellular. There’s no proof that conserving XScale or creating ARM merchandise would have modified that — if something, the ARM division would’ve been beneath much more strain to make sure it by no means grew to become a menace to the x86 enterprise.