Friday, November 7, 2014

High Tech "State of the Union", By the Numbers: Q3 2014

Source: commons,

If you can look into the seeds of time,
And say which grain will grow and which will not;
Speak then to me. - Shakespeare, "Macbeth"

When I first began this blog on March 30, 2014, one of my objectives was to report on the business climate of High Tech and identify trends which might indicate threats or opportunities for growth. To that end, I selected ten companies in the semiconductor, systems and software spaces that could be reasonably classified as leaders in their respective technology spheres. This spread of firms also provides a barometer of relative health for the markets served by High Tech offerings - the three C's (communications, computing and consumer), ISM (industrial, scientific and medical), mobile computing, automotive and military/aerospace.

In order to maintain continuity for the reader and provide relevant background information, previous analyses can be found at the links below.

Q1 2014 (published in April and May:)

Q2 2014 (published in August:)

With Q3 numbers now in hand, we should have a clear picture regarding 2014. The Q3 numbers are particularly important, in that they reflect the buildup to the holiday shopping season and are thus both an indicator of actual health sweetened by the industry's level of optimism. So, dear readers, let us return to the slopes of Mount Parnassus, where the Pythia awaits us in the Temple of Apollo, ready to serve us as oracle.

Please note: I am not an investment advisor, have no equity interest in any of the companies under review and am not retained by any of them to produce an opinion. As far as I'm concerned, the various financial markets for securities and equity are nothing more than casinos and, since Las Vegas is just a two hour drive for me, I prefer the brazen honesty of the Vegas Strip to any "investments" offered by Wall Street.

The Big Iron

Heaven makes sport of human affairs, and the present hour gives no sure promise of the next. - Ovid

The big news for HP last quarter was it's announcement of a split between its Enterprise and PC & Printer groups:

The motivation is easy to discern - HP divided its business by margin. It would not surprise me in the least if the PC & Printer group is eventually sold off to a concern in China (or perhaps India, which would be making a shrewd move in bringing higher levels of the technology chain over to their own shores.) I don't know if this means the two separate groups will be reporting financial results independently; if they do, it may force changes in my own reporting and analysis. We will have to wait and see.

IBM has been in the news as well, having finally concluded a deal for its fabs with GlobalFoundries:

Putting it plainly, IBM will pay GF $1.5B over three years to take over Big Blue's foundry operations. The fact that IBM needs to pay GF to take the fabs off its hands suggests that IBM's chip operations have severely bloated overhead costs and woefully insufficient manufacturing volume. 

IBM also plans to spend $3B over the next half decade on basic semiconductor research. As discussed in previous posts, IBM's chip technology is integral to the performance and quality of its computing systems. I sincerely doubt that $3B over 5 years will be adequate for IBM to keep its chip know-how (including process technology with its models & data, EDA, IC packaging, embedded IP and all the rest) sufficiently up to date to continue satisfying the extraordinary needs of its Server group. Again, time will tell.

Cisco has also undergone extensive internal changes. The story I'm hearing is that its 15 product divisions have been consolidated into 4. The following link tells part of the tale:

This change has been long overdue and should result in profound improvements to Cisco hardware and software QA&R, as well as eliminating a great deal of administrative overhead. However, I've been on the receiving end of layoffs myself and have great sympathy for the hard working folks at Cisco who are drawing the short straw in these re-orgs.

With calendar Q3 2014 financial reporting included, here is the updated chart for the Big Iron going all the way back to Q1 2008:

There has been speculation in the trade press that both Cisco and IBM have been particularly hard hit in their overseas revenues by last year's revelations about the NSA tampering with computing system hardware to plant snooping technology. One would imagine that this had some effect, and the post-Q4 2013 Cisco dip strongly suggests so. 

Nonetheless, the numbers tell a story about the fortunes of each of these firms and the enterprise networking & high performance computing segments of High Tech on a global basis. The long term HP and IBM trend lines are still on a downward slope and Cisco's growth continues to be mediocre. If the High Tech sector is experiencing a turnaround, it is not evident at the enterprise system level.


The phoenix hope, can wing her way through the desert skies, and still defying fortune's spite; revive from ashes and rise. - Miguel de Cervantes

As noted in last quarter's report, the PC Old Guard is endeavoring to shift their technology focus away from the decaying personal computer market into server and mobile applications. Microsoft is starting to see positive results from this strategy and is being rewarded for its persistence, as sales of its Surface tablet are now just shy of $1B and the product line is on the verge of profitability:

Considering that Microsoft achieved this while worldwide tablet sales are slowing as the market saturates makes the achievement that much more impressive. It indicates the value of Surface is such that the product is clearly beginning to command greater consumer interest.

Intel is continuing its campaign to dethrone Qualcomm with the release of four new Broadwell devices targeted at tablets:

The effort to break open the mobile computing market will continue with Skylake, which is currently scheduled to be released roughly one year from now. Significantly, Intel is putting primary emphasis on answering power consumption and power efficiency concerns in its early Skylake promotional efforts:

As I wrote last quarter, 2014 is a pivotal year for both MSFT and Intel, as they try to shake off the chains of their legacy PC business. Let's see what the revenue numbers tell us.

The long term growth trend for Microsoft is very evident over the last six years and I reaffirm my forecast that the company is the most likely High Tech firm to outperform the market. After 3 years in the doldrums, Intel's last two quarters suggest it is poised for a breakout of its own. I am now cautiously optimistic about the company. 

From a technology standpoint, Intel seems superlatively well positioned, with leadership in ultra-deep submicron process technology, 3D-IC packaging with TSV that integrates multicore hyperthreaded CPUs with GPUs and system memory, along with potential integration of ARM cores (assuming it becomes necessary.) If Intel can begin to genuinely convince mobile computing OEMs that it can run ARM applications more or less as well as Snapdragon devices, it will be all over for Qualcomm, Mediatek, Samsung and the rest in the mobile computing applications processor market and Intel will add $20B or more to its annual revenues. However, I won't be firmly convinced that the company is on the path to a strong future until I hear news that Qualcomm is losing smartphone and tablet accounts to Intel.

The Stone Masons

If we must fall, we should boldly meet the danger. - Tacitus

This portfolio of companies covers every High Tech industry segment in depth. Qualcomm is the dominant force in mobile computing; Broadcom is (arguably) the inventor of the SoC and plays primarily in enterprise networking and storage, with some additional products in consumer and wireless; and Xilinx covers nearly the full range of market segments (the three C's, ISM, automotive and mil/aero) with particular strength in wireline networking.

The revenue chart shows some very interesting changes from last quarter:

At first I was a bit surprised by the Qualcomm result, but upon reflection, there are indeed logical reasons for the Q3 revenue decline. As stated before, the smartphone and tablet markets are saturating this year, robbing Qualcomm of its primary growth driver. Compounding this is the fact that Apple - who develops their own ARM-based applications processor - has eroded the market share of its Android-based competitors and thus undermined Qualcomm's own customer base.

I believe Qualcomm has now entered into crisis. It's been a very strong company for decades and has possibly the most impressive technology portfolio in the semiconductor industry, likely surpassing even that of Broadcom. Yet with its major markets entering stagnation and Intel having painted a bullseye on it, Qualcomm must find some means of pivoting on its technology base and branching out in new directions - and it needs to do so with urgency.

Defending its revenue base would require putting sufficient pressure on ARM to develop a multicore CPU with hyperthreading that could efficiently support Windows applications - or perhaps signing an architecture license with ARM and doing the work internally (a very expensive and resource-intensive endeavor.) Failing that, perhaps a die sourcing agreement of Intel-compatible CPUs from AMD and incorporating such die into 3D-IC packaging might be made to work (though there will be some daunting technical and business obstacles to overcome.) If Qualcomm hasn't already been working covertly on some sort of solution along these lines already, it bodes ill for the long term viability of the company.

Finding alternate sources of growth will be just as significant a test of the company as defending against Intel. Considering Qualcomm's strengths, perhaps a direct foray into the IoT might be the smartest move - not just as a chip supplier, but a peripherals developer as well. The day Qualcomm announces some sort of MCU-equivalent of Snapdragon would be the best indicator that they have come to such a conclusion.

Broadcom's revenue blip to the upside - a record high quarter for the company - is undoubtedly a welcome sign to investors and employees after relative stagnation over the last three years. On the earnings call, Broadcom cited strength in its broadband and STB offerings as the primary driver of its happy earnings report. Evidently the Netlogic acquisition was a very good one (congratulations to my ex-comrades in Netlogic from my MIPS and Xilinx days.)

Nonetheless, forward guidance is down. At least Broadcom recognizes they aren't out of the woods yet. However, it was also apparent from the conference call that Broadcom's executive management team believes 'staying the course' will take them forward. 

Evidently the company's management does not fully appreciate the difficulties that lie ahead of them in these new, more difficult times. Broadcom's margins will continue getting squeezed in stagnant markets as they find themselves increasingly constrained from implementing significant value-adding innovation with the arriving demise of Moore's Law. 

They will also find themselves unable to effectively break into new markets without freeing their employees from the straitjacket of behavioral and operational conformance of Broadcom's classic Silicon Valley hierarchical management style. Stated differently, the company's innately talented employee base and rich IP portfolio will avail them nothing if they continue demonstrating that they don't know how to apply them.

Furthermore, Broadcom will find itself dangerously vulnerable should Intel successfully attack Qualcomm's customer base. If Intel proves that they can efficiently and cost-effectively support ARM-based applications in the mobile computing space, there's nothing stopping the company from also going after Broadcom's ARM-based SoC business. Staying the course is the worst possible choice for Broadcom over the long term.

And now, to the final member of the Stone Masons.......
The Xilinx situation is at this point hopeless. It has been exactly four years since the company's quarterly revenue peaked at $620M. Since then, earnings have been in the doldrums. Xilinx seems to have largely missed out on the enterprise networking and datacenter growth spurt in the transition beyond 1Gbps transmission rates. In the same period, both Lattice and Altera experienced substantial growth (though for very different reasons.)

I analyzed the Programmable Logic segment in depth midsummer (see the link below.) When viewed from this perspective, it is evident that Xilinx is not really addressing the future of its business with anything like a long term strategic plan that encompasses a detailed understanding of its markets and consequent technical challenges.

I will maintain my watch on Xilinx revenue for the sake of continuity, but in the future will select either Lattice or Altera (or, if time permits, both) for revenue reporting and analysis in order to gather observations on a more dynamic part of the Programmable Logic sector and its implications to High Tech in general. Thanks to its Silicon Blue acquisition, Lattice seems to have taken the 'jelly bean' programmable logic space virtually unopposed and (as rumor has it) is taking active measures to participate in the budding IoT, while Altera has been apparently seeking out growth opportunities more aggressively and resolutely than Xilinx in the high end. I think we will learn much more by studying one or both of the other major players in Programmable Logic rather than Xilinx.

Area 51

Apple and Google are still the darlings of Silicon Valley, and deservedly so. They are, after all, more than mere innovators. These companies are inventors on a mission to change the world for the better.

It's rather plain to see that business is currently very good for both companies. The iPhone 6 and 6 Plus are helping Apple take market share back from competitors (Samsung in particular, it seems.) Google continues positioning itself for future success in mobile computing with the Android Wear OS, employing it themselves in a smart watch as its latest hardware product for the IoT after Glass.

Clearly, continuing financial strength is keeping investors and employees for both companies very happy:

The positive revenue trend for the Area 51 firms is obviously still intact. If anything, Apple's latest quarterly number tantalizingly suggests that the company might be poised for a breakout from the plateau that was clearly building in its profile. 

However, I wouldn't call one quarter's deviance an established trend just yet. After all, iPhone 6 sales are a growth spurt achieved at the expense of rivals in what is now a saturating smartphone market. As stated earlier, tablet sales are also rolling over this year. Thus, both of Apple's primary markets are entering a period of stagnation. Will the company be able to continue growing by taking further market share from competitors? The Q4 number will be extremely important for the short and medium turn outlook. The holiday spike will need to beat last year's number - and quite decisively.

There are greater concerns regarding Apple's long term growth potential. Its first entry into the IoT fray - the Apple Watch - is very conventional and extremely unlikely to make any serious contributions to Apple's growth (see previous editorials on the IoT for an explanation.) To make matters worse, the wearables market may turn out to be something of a fad. A significant percentage of smart watch customers have already walked away from their purchases:

A re-think of Apple's approach to the IoT is in order. One can already see a trend towards convergence between the smartphone and tablet (currently referred to in some corners of the trade press by the singularly horrible term "phablet.") It is my firm belief that Apple's future lies in the direction where a combined smartphone/tablet becomes a Personal Processor and everything else associated with it becomes a peripheral. This would position Apple to be at the center of the IoT universe.

Apple's vision should include a conceptualization of the Personal Processor as a device that can accept wireless plug-ins from a much larger touch-capable display screen, mass storage device, internet access point and full function keyboard, obviating the need for a desktop, television, home router and game console. If the storage brick, keyboard and display were mechanically shrinkable so that they could become easily portable, a separate laptop then becomes superfluous. Everything else - game controller, headset, smart glasses, smart home controller, etc and so forth - would also be a peripheral, with the PP as the center of an individual's digital universe.

Until Apple communicates such a vision, however, they risk being reduced by public perception to a mobile computing version of a late 1990's Intel - dependable but ultimately no longer compelling. Tim Cook needs to demonstrate that he's more than just an executive showman for iPhone rollouts.

To realize such a vision of a Personal Processor as detailed above, Apple will need to gain technical proficiency at the chip level in leading wireless standards that support P2P protocols and multiple channels such as BT 4.0, Wireless USB and Wi-Fi Direct, along with 3D-IC, TSV and, in the long run, graphene. It is an observable fact that the company continues to recruit very experienced chip developers. Apple also has a history of M&A activity in this sector with its purchase of PA Semi some years back. It would not surprise me in the least if Apple were to make another such acquisition in the next 2-3 years.

Google's situation is no less worrisome from a long term strategic perspective. Q4's number will be critical - if it doesn't top the Q4 2013 high, then Google will have entered a period of revenue stagnation. This could become a very stressful circumstance for Google's executive management. The company has already come under criticism for the level of its applied R&D expenditures which to date have produced no significant revenue generating products, especially in hardware. If revenue starts to drift sideways, those critiques will become heated.

The Temple of Apollo

As we bring our sacrificial offering and a few gold coins to the Pythia, what can we anticipate that she is likely to say to us? Assuming the priestess has access to the same data we do, we can reliably posit that she will agree with us on the following:
1. The Stone Masons and the Big Iron are failing to demonstrate any significant vigor in the market.
2. Microsoft continues to follow its own path and achieve growth that outperforms much of the rest of High Tech.
3. Apple has recaptured some market share from rivals. Whether they can sustain this remains to be seen.

The inescapable conclusion is that despite the furious labor of its participants, High Tech is stuck for the time being in first gear from a growth standpoint. The only bright spot in the sector appears to be the memory chip business, where DRAM and Flash prices are once again robust due to supply constraints. (Side note: considering that this is the third year in a row where suppliers are proclaiming a lack of sufficient output to keep up with demand, yet the wider market is not showing energetic growth, the situation justifiably arouses a healthy skepticism and suspicion.)

What could be keeping High Tech relatively stagnant? A look at the CRB (Commodity Research Bureau) global index might be enlightening in this respect.

This index tracks market pricing of 23 commodities across industrial, textile and agricultural markets - copper, wheat, iron, cotton, etc etc..... There are some useful observations to be made from this chart:
1. Though this is not directly pertinent to the current discussion, note how the index began a precipitous climb with endemic and severe volatility after 1971 - the year in which the US government broke the Bretton Woods agreement and definitively abandoned the gold standard for the USD. Interesting, isn't it?  ;-)
2. The index crashed in 2008-2009 and took another severe dip in 2011 from which it has not yet recovered. 2014 appears to be another weak year, as the index has dropped throughout the year.

Another useful chart is the BDI (Baltic Dry Index), which tracks shipping rates for cargo haulers at the world's major ports:


Clearly this chart requires some explanation. :-)

There are several measurements drawn on this chart - the Relative Strength Index, 50 and 200 day Moving Averages and the Moving Average Convergence/Divergence (along with a complementary histogram.) These are measures employed by 'technical analysis' traders who subscribe to various linear mathematical modelling techniques for market data - Elliot Wave Theory, Dow Theory and the like. I am not a fan of any of those (believing instead that Mandelbrot and Taleb have it right, in that complex systems such as markets are inherently non-linear) and prefer to stick to the data as is.

The measurement intervals for the actual index - the dashed line in the image above - are weekly, from Q4 2012 to today. Notice how the index tends to peak right around the holidays. This is because of the shipping activity that accomodates consumer demand for all the things we want Santa Claus to bring to us. As a point of comparison, the BDI peaked at roughly 11,800 in May 2008. Since then, it's never crossed 5000. This is despite the fact that a considerable number of cargo haulers were retired last year due to chronic over-capacity driving down shipping rates.

In December 2012, the index peaked at around 1900. In December 2013, the peak was roughly 2400-2500. Currently it is just above 1400 and tending down.

The two charts are conclusive. The global economy is still at a near stall and has no appreciable upward momentum. The CRB index in fact indicates the possibility of another recessionary slowdown, which may be also reflected in the surprisingly poor BDI curve.

What this means for High Tech is that end users and customers of technology still have limited disposable income for discretionary expenditures. In other words, coaxing someone to spend $50-$200 on a new smart watch or signing a purchase agreement with an enterprise to install a datacenter expansion is considerably more difficult than it was just 7 years ago. 

In today's High Tech business environment, Cost Is King. The value propositions of technology offerings must adjust accordingly, either to directly address customer requirements for reducing total cost of ownership or to deliver a value prop so colossal that cost becomes a secondary issue.

Never let the future disturb you. You will meet it, if you have to, with the same weapons of reason which today arm you against the present. - Marcus Aurelius

Such news is hardly welcome for anyone, but especially for those of us in High Tech. We are all optimists to one degree or another, proud of the work we accomplish and the sophisticated technology which we help bring into being. We all want our efforts to pay off - and High Tech workers put an effort into the job like very few people in other industries ever experience.

Notwithstanding the work we do and the things we achieve, the time for a re-flowering of the High Tech sector is still in the future. People in our industry continue to work on wonderfully innovative ways of enhancing silicon capabilities, as well as materials that will one day replace it. Once these inventions and innovations are in confluence with major global economies that have finally shed their sovereign debt burdens and the unhealthy portions of their finance sectors, it will be High Technology that reignites growth and prosperity globally, and all our efforts will finally bear fruit.

I have received requests to cover some other companies besides the ten currently in the portfolio. To that end, I will be publishing a supplementary report sometime later in the month that will cover as many of them as I possibly can. These companies include:
- STMicro                                            - Infineon
- Mediatek                                           - Altera
- ARM                                                  - Lattice
- NXP                                                   - NVidia

Personal Note:
Though it may take a reader only 10 minutes or so to go thru an editorial, it actually takes me an average of 20 hours over a period of 4 days to compose these blog posts. I find that during that time I often rewrite sentences and sections, contemplate and reformulate my ideas and alter & refine them repeatedly.
I put this sort of effort into composing these articles because I view them as a conversation that I'm having directly with you. I want you to both enjoy the experience and be stimulated by it, as my goal is to make it a rewarding experience for you. With great modesty I hope that I am achieving this at least in part, and I thank you for your attention during what is almost certainly an already very busy day.


  1. Just a small point and I don't think it invalidates your conclusions but Qualcomm is already an arm architecture licensee and makes it's own arm cores. The snapdragon itself for example is a full Qualcomm design so they already have demonstrable processes, teams and IP in custom arm cores.

    1. Interesting - I did not know this, Matt. That helps, without question. Now Qualcomm has to figure out if they can rely on ARM to provide a Windows-enabled CPU or if they have to design one themselves.

  2. Great wrap-up of the quarter. Too bad about Xilinx, but the numbers don't lie. It will be interesting to see what the future holds for the company because historically semiconductor companies without growth just fade away through consolidation.


Feel free to comment or critique!