backtop


Print 3 comment(s) - last by p05esto.. on Jan 18 at 9:32 PM

Despite struggles with divide focus, production issues, Intel remains strongly profitable

After yesterday's pleasant surprise from Taiwan Semiconductor Manufacturing Comp., Ltd. (TPE:2330) (TSMC), many analysts were hoping for something similar from Intel Corp. (INTC), the U.S. chipmaking veteran that for decades has enjoyed a dominant position in the personal computer industry.  Instead, Intel missed analyst estimates offering yet another indication of the lessening importance of the traditional PC and the growing role of mobile devices, a sector in which Intel struggled.
 
I. Swing and a Miss
 
In Q4 2013, Intel pulled in revenue of $13.8B USD, narrowly better than the $13.7B USD or $13.72B USD that analysts surveyed by Bloomberg and Thomson Reuters I/B/E/S (respectively) estimated.  Revenue was up 3 percent on a year-to-year basis (from Q4 2012).
 
The miss came in net profit, which came in at $2.63B USD, or $0.51 USD/share.  Analysts had expected $2.68B USD ($0.52 USD/share), according to both Bloomberg and Thomson Reuters I/B/E/S.  Net income was up 6 percent on a year-to-year basis.
 
Most of that gain came from the data center group, which grew revenue 8 percent.  The traditional PC Group was flat in revenue on a year-to-year basis.

14 nm SoCIntel CEO Brian Krzanich brags about his company's next generation upcoming SoC product at IDF 2013 in Sept. [Image Source: WCCF Tech]

The "Other Intel architecture operating segment", which includes Intel's smartphone processor efforts grew even more -- 9 percent in a year-to-year basis.  But that revenue growth did not necessarily grant profit -- growth in past years has come at the cost of increasing net losses from the unit, one of Intel's three core businesses.  The quarterly gain was also not enough to stop Intel's yearly revenue total from dropping for a second year straight on mobile struggles.

Intel

Intel continued to share the wealth with shareholders, paying $1.1B USD in dividends.  It also spent $528M USD to repurchase 22 million shares.
 
II. No XP Bump
 
Analysts were disappointed to see that the upcoming April end-of-life for Microsoft Corp.'s (MSFT) still widely used Windows XP wasn't driving fresh sales. 


Alex Gauna, an analyst at JMP Group Inc. (JMP), comments to Bloomberg:

The prevailing view coming in was that things are getting better.  The quarterly result and forecast says that things are still not very good.

Chief Financial Officer Stacy Smith tries to cast a more cheery spin on the results, while admitting to PC declines, stating:

[Q4] was a return to financial growth for the company.  Specific to the PC market, it is getting better. But, consistent with third-party estimates, we’re expecting a small, single-digit decline [in PC sales].

In a comment with Reuters, the Intel exec tried to reassert that PCs were still a "strong" market, blaming the U.S. government for a decline in orders.  He comments:

The PC market was a little stronger than we thought. What's driving PCs right now are the innovative form factors we've been working on... We saw a tapering off in order patterns across certain customers in certain segments at the end of the quarter, and we think that was driven by the government shutdown and the uncertainty around the debt ceiling.

In terms of outlook, Intel plans to cut its capital spending to $11B USD this year and expects revenue to be flat at $48.3B USD. It also expects its gross margin -- a measure of profitability -- to drop from 62 percent in Q4 to 60 percent, on average in 2014.  Intel expects a $1B USD quarter-to-quarter (QoQ) drop in revenue in Q1 2014, and also warned of a $200M USD restructuring charge, which should damage the net profit.
 
III. Intel's Process Lead May be Disappearing
 
Eroding Intel's margins in Q4 were higher than expected capital spending costs of $4.8B USD.  Intel is currently in the process of transitioning its desktop and mobile chips to the 14 nm node.  After the traditional tapeout and die fixes Intel hoped to proceed towards a volume release in Q3 2014 of Broadwell, the world's first 14 nm consumer CPU product.  Instead, the die fixes didn't work as well as expected and Intel was forced to make further corrections and bump Broadwell to Q4 2014.

Intel breaking fab
Defects at 14 nm forced Intel to delay one of its processors. [Image Source: Intel]

TSMC is putting pressure on Intel, transitioning to 20 nm this quarter.  Intel still has some advantages thanks to its 3D transistor technology, but TSMC is actually expected to produce processors on a smaller node than Intel for nearly four quarters, although the actual sales window will be shorter, given OEM integration times.  Still it is likely we will see 20 nm TSMC product before we see 14 nm Intel product.
 
TSMC also had plans to tapeout 16 nm, 3D transistors next year.  And Samsung Electronics Comp., Ltd.'s (KRX:005935) (KRX:005930) is expected to tape out 20 nm this year, reaching volume production next year.

IDF 2013
[Image Source: Jason Mick/DailyTech LLC]

Intel is trying to expand its contract fabrication abilities, but it isn't necessarily in as strong a position as some might think.  Intel recently announced it would only be upgrading two of its facilities for 14 nm production this year, upgrading several other facilities for 22 nm production.  At 22 nm, Intel's older fabs will have a leg up on TSMC's older 28 nm fabs, but will be behind TMSC's latest 20 nm fabs.
 
With only two capable of 14 nm production, supply will be limited.  Intel will be doing well to fully satisfy global PC demand; mobile efforts will likely stay stuck at the 22 nm node.  That's not to say that we won't see 14 nm smartphone CPUs from Intel in late 2014 or early 2015.  But even if we do see such chips, Intel is unlikely to be able to produce them at high enough volumes to serve top OEMs' flagship devices.
 
Intel promises it will move faster in the mobile market, with CEO Brian Krzanich stating:

Intel was slow to respond to the ultra-mobile PC trend.  We will move Atom even faster to our leading-edge silicon technology.

But how he's going to accomplish this with only two 14 nm fabs for the near future appears unclear.

Intel 14 nm
Intel's 14 nm yield is improving, but the company still is delaying the release of the new product a bit. [Image Source: Intel via ExtremeTech]

Krzanich, who was appointed last May, has a tough challenge ahead as he tries to preserve the declining PC business, while growing Intel's embattled entrants in the booming mobile markets.  As the fierce competition between Samsung and TSMC -- two companies solely focused on mobile -- has shown, it's hard enough to win market share even if you spend a lot and are solely focused on mobile.  For Intel, whose focus is divided, it's a daunting challenge.
 
Another long-term problem is the path ahead after 7 nm transistors, which Intel plans to release in 2018.  A die shrink or two more might be possible, but eventually you're running out of atoms to have a workable transistor.

hybrid memory cube
Intel's node-size expertise will be less valuable than companies' like Samsung and Micron's chip-stacking expertise. [Image Source: Micron]

At that point the clearest direction would be to build 3D circuits with multiple layers.  But companies like Samsung or Micron Technologies Inc. (MU) may suddenly gain the upper hand when that happens.  Where Intel's expertise is in node shrinks, Samsung and Micron have become the experts in stacked chipmaking as that technology is already driving the push of DRAM and NAND circuitry to greater densities.  Intel -- which also produces NAND -- has certainly spent some time researching the technology, as well, but has somewhat less experience with it.
 
At one time it would be simple to just say, "Oh, Intel will just spend more and eliminate this lead."
 
But in 2013 Samsung spent $22B USD on chipmaking capital expenses, precisely twice Intel's investment of $11B USD.  And the talent gap is nowhere near what it once was, with many of the best U.S. chip designers having migrated to Samsung or TSMC.

Sources: Intel, Bloomberg, Reuters



Comments     Threshold


This article is over a month old, voting and posting comments is disabled

14nm + 450mm
By Samus on 1/18/2014 12:35:23 AM , Rating: 5
You didn't mention the other side of the coin. While everybody else is at 20nm 300mm wafers, Intel will be at 14nm, 450mm wafers.

This gives Intel's two 14nm fabs the production capacity of 6 of TSMC's 20nm fabs. This, of course, assuming the fans have the same sq Ft of production space, and identical yields (best case for a new process likely being around 80%)

But you get my point. Intel isn't delaying 14nm because it isn't ready. They're delaying it because of the 450mm retooling.

Everyone is always going to be multiple steps behind Intel. They might catch up in one area (lithography) but there is always process, engineering, production capacity, quality control, etc.

TSMC still hasn't even announced how they are going to control leakage with their 20nm FitFet parts. Information is pretty sparse considering they are supposed to be going live with 20nm next quarter.




14nm + 450mm
By Samus on 1/18/2014 12:35:24 AM , Rating: 2
You didn't mention the other side of the coin. While everybody else is at 20nm 300mm wafers, Intel will be at 14nm, 450mm wafers.

This gives Intel's two 14nm fabs the production capacity of 6 of TSMC's 20nm fabs. This, of course, assuming the fans have the same sq Ft of production space, and identical yields (best case for a new process likely being around 80%)

But you get my point. Intel isn't delaying 14nm because it isn't ready. They're delaying it because of the 450mm retooling.

Everyone is always going to be multiple steps behind Intel. They might catch up in one area (lithography) but there is always process, engineering, production capacity, quality control, etc.

TSMC still hasn't even announced how they are going to control leakage with their 20nm FitFet parts. Information is pretty sparse considering they are supposed to be going live with 20nm next quarter.




Huh
By p05esto on 1/18/2014 9:32:16 PM , Rating: 2
Very interesting to see how quickly the competition has crept up on Intel the mighty giant. I really like Intel as a company, always have.... I will support them as a US company before I do Samsung and the others. I think maybe the competition will be good for us all. For a few years Intel just kind of coasted and enjoyed leading AMD. They still lead, but need to step up about right now. Those darn mobile chips caught everyone off balance.




"It looks like the iPhone 4 might be their Vista, and I'm okay with that." -- Microsoft COO Kevin Turner














botimage
Copyright 2014 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki