Friday, December 5, 2014

High Tech "State of the Union" - Q3 2014 Supplement #3


I met a traveller from an antique land
Who said: — Two vast and trunkless legs of stone
Stand in the desert.
 Near them on the sand,
Half sunk, a shatter'd visage lies, whose frown
And wrinkled lip and sneer of cold command
Tell that its sculptor well those passions read
Which yet survive, stamp'd on these lifeless things,
The hand that mock'd them and the heart that fed.
And on the pedestal these words appear:
"My name is Ozymandias, king of kings:
Look on my works, ye mighty, and despair!"
Nothing beside remains: round the decay
Of that colossal wreck, boundless and bare,
The lone and level sands stretch far away. - Percy Bysshe Shelley, "Ozymandias"

Thirty years ago I stopped in London on my way to vacation with my mother's family in Bologna, Italy. I arrived in early June and found in my first visit to the great city that it was foggy and freezing (which I understand is not too unusual for that time of year.) 

Despite the inclement weather, the side trip was an excellent one and included many wonderful discoveries. For instance: the hamburgers I wolfed down made the absolute best on offer back home taste like cardboard by comparison. A few pub stops revealed a selection of astonishing beers which I found superior to anything I've had even in Germany (side note to my German readers: before you send me a storm of hate mail, you have to admit that the Brits are exceptional brew masters.) Finally, I had the opportunity to try Indian cuisine for the first time in a quiet little restaurant near my hotel - a very unpretentious place with impeccable service, where the tablecloths were changed for every sitting. The experience shattered all my preconceptions of what fine food was like and demonstrated to me that even the best French and Italian cooking was like boiled mud by comparison.

Of course, only a minor selection of my experiences were on the culinary front. While in London, I decided to pay a visit to the British Museum. One of the great cultural institutions of the world, there was one exhibit in particular that burned itself into my memory. In one of the halls I beheld an Assyrian relief carved into some sort of black stone - obsidian or marble, I'm not sure which. The relief was massive, depicting scenes from Sennacherib's sundry victories over neighboring cities and tribes. In one scene, the leaders of a defeated force of the kingdom of Elam in southern Iran pleaded for mercy from Sennacherib; an inscription proclaims that he "cut their throats like sheep." Though such displays of cruelty were traditional amongst warring parties in the region in those days and were in fact considered 'de rigueur' as demonstrations of strength, the relief's carvings and depictions collectively left an impression upon me of truly terrible power - and by that I mean 'terrible' in the old sense: a feeling that combines awe and fear.

The state of Assyria, at the northern end of the Fertile Crescent in the Near East, was founded on the city of Assur (or Ashur) in 2600 B.C. For the next 16 centuries, the Bronze Age Assyrians experienced alternating periods of subjugation to their neighbors - Babylonians, Akkadians, Hittites, Sumerians and others - interspersed with periods of independence as a regional power. It wasn't until the beginning of the Iron Age in the early 900's that a militarized Assyria became overwhelmingly dominant, with hegemony over a vast area that encompassed the entire kingdom of Egypt, the states of the Levant, central and eastern Turkey down to the northern limits of the Arabian peninsula, and east over the Zagros mountains to the western plateau of Iran.

For three centuries, Assyria stood at the apex of the world as the most advanced, sophisticated and powerful civilization on the globe. A standing professional army, paved roads reaching to all corners of the realm that both facilitated troop movements and expanded opportunities for trade, a pony express system for accelerated communications and an efficient civil administration all relied on the Assyrian mastery of sophisticated organization and leading technology of the era. It seemed an invincible empire, wealthier than any nation seen before and an unstoppable, merciless scourge to its enemies.

Yet fall they did, to a resurgent and implacable Babylonian power to their south in 605 B.C. The energy and discipline of the Assyrians in their last decades never wavered as they implemented a continual stream of tactics and maneuvers to resist the Babylonian onslaught, but it was to no avail. Their undoing was, in the end, their own fault. Though the foundation of Assyrian society as a military state imposed limits on their ability to generate sufficient wealth for their war machine or instill the necessary loyalty and commitment from their subject peoples to respond to Babylon's challenge, they were unwilling to change their ways.

Some of you are at this point undoubtedly saying to yourselves "Arright, baldie, thanks for the history lesson. Now please tell me - what does any of this have to do with High Tech?"
It's a fair question, and I'll cut to the chase.

His resolve is not to seem, but to be, the best. - Aeschylus, "Seven Against Thebes"

Today we have our own High Technology empires that have carved out the story of their achievements and victories in stone - not from quarried marble, but fabricated silicon. These are companies with deep and broad IP portfolios that develop SoCs for the leading systems markets of the 3C's - communications, computing and consumer - as well as mobile computing (tablets and smartphones.) They are at the very pinnacle of the semiconductor value chain, with a mastery of both hardware and software technology so great that they have actually captured significant value away from their systems customers.

In "State of the Union" financial summaries from previous quarters, I put Broadcom, Qualcomm and Xilinx together as representative of this group, which I termed "The Stone Masons." Based upon requests from the readership and the evolution of my own thoughts on the subject, I stripped Xilinx out and put them into a new category along with Lattice and Altera, which I call "The Vanara" (supernatural beings from Hindu myth of great power that are reputed to be shape-shifters.) These companies were reviewed in the November 21st blog post. 

The Stone Masons now have two new members: MediaTek and Nvidia. In terms of technology and talent, this new lineup represents the cream of the crop in the semiconductor sector worldwide.

The Stone Masons have certain basic characteristics in common. When they started out, they were true technology pioneers, staking their claims on territories that were once dominated by ASICs. Conventional wisdom dictated that no single company could develop products that could encompass the many unique hardware requirements of OEMs or have the technical sophistication to build software stacks and efficiently integrate them into chip hardware architectures to the satisfaction of systems houses. Beginning with Broadcom, the SoC firms proved the pundits wrong as they developed chips that captured 80% or more of the hardware needs for wireline networking, broadband, storage and wireless customers while providing software distributions containing OS's, firmware, middleware and applications I/F's along with tools that gave system developers the flexibility they needed to create value-adding personalization thru software. They were, in other words, real frontiersmen and daring adventurers who opened vast new horizons for microelectronics.


Today, SoC product development is heavily governed by industry standard specifications - storage, networking, wireless and multimedia. Control plane processing is centered on individual licensing choices for ARM CPUs, whereas data plane processing tends to be a homegrown thing, such as Qualcomm's Adreno, Broadcom's VideoCore and NVidia's own GPU IP. There is some additional third party licensing for items such as DSP cores, which both Broadcom and MediaTek source from CEVA. Nonetheless, the great majority of the IP portfolio possessed by each member of the Stone Masons is home made. 

These companies are also ferociously driven by the scheduling demands of their markets - in particular, consumer applications with their very short system development, release and obsolescence timelines. For example: the development treadmill for segments such as HDTV are the most brutal and unforgiving sectors of High Tech. It has been a normal thing over the past two decades to observe a half dozen startups launch with promising new video technology and find only one survivor after 3-5 years. A single defect in a chip release could result in a missed market window and potentially spell doom for the fledgling firm.

As a consequence of the operational straitjacket imposed by market schedule demands and industry standards, SoC companies tend to be very sternly regimented with a quasi - military management style. Naturally, the intensity with which that administrative form is applied will vary somewhat from company to company and even between departments in the same firm. Nevertheless, as a general rule such work environments are hostile to initiative, creativity, going against the grain and challenging authority.

With this in mind, let's take a look at the quarterly revenue numbers, from q1 2008 to q3 2014. The MediaTek numbers are normally reported in New Taiwan Dollars; thus, to keep the results directly comparable, I've used an NTD:USD conversion rate of 30.77:1 (valid when I collected this data in the third week of November.)

Qualcomm and Broadcom were reviewed in the November 7 editorial and their status is unchanged. As a refresher, you can read the analysis here:

Like Broadcom, Nvidia has also been stagnant for many years. Unlike Broadcom, though, Nvidia did not experience a growth spurt from Q1 2009 to Q4 2010 - hence the roughly $1B difference between their quarterly numbers. 

There are more fundamental differences between the two companies besides revenue levels. Broadcom is a broad line supplier of extraordinarily sophisticated hardware and software solutions to large, established SoC market segments in the 3C's. Nvidia also offers exceptional technology, but relies on most of its revenue coming from the stagnating desktop and laptop computing segments. The Tegra mobile applications processor venture into smartphones showed early promise but has been soundly defeated, primarily by Qualcomm. The rest of Nvidia's segments are all exploratory in nature - industrial design, medical diagnostics, supercomputing, data analytics and other such initiatives. All of these expeditionary efforts are clearly intended to stretch Nvidia's visual and parallel computing technology to the limits with the hopes of triggering growth thru breakthrough applications that give rise to nascent markets dominated at the outset by the company.

One can only applaud Nvidia's courage in devoting so many resources to what are, in effect, applied R&D efforts intended to provoke the birth of new markets for the company's offerings. Yet this nearly seven year stretch of essentially zero growth is disconcerting, to say the least. There are two primary causes of Nvidia's continuing lack of success - management practices and unrecognized opportunities.

The shaft of the arrow had been feathered with one of the eagle's own plumes. We often give our enemies the means of our own destruction. - Aesop's Fables, "The Eagle and the Arrow"

A critical reason for failure can be placed squarely at the feet of the firm's leadership. Nvidia is notorious in Silicon Valley for having an incredibly heavy-handed and even viciously authoritarian style of management. 

Callousness with regards to recognizing and rewarding employee efforts and contributions seems to be standard procedure except for a few managers and departments. In such a work environment, innovation and creativity are not only stifled from above but also withheld from below, as employees spend most of their energies avoiding punishment and additional stress thru activities that are not particularly vulnerable to criticism. Stated differently: in an organization where negative feedback is zealously practiced, employees do not stretch themselves or stick their necks out to achieve great and noteworthy things but instead keep their heads down in order not to attract the attention of the whip.

Though Nvidia's executive management team would undoubtedly argue energetically to the contrary, the revenue numbers speak for themselves. What is especially tragic is that the basic elements to position the company for spectacular growth and leadership in the IoT are already in place, but there seems to be no executive recognition of that fact (see the discussion of the Jetson development platform from the November 14 editorial: .) 

My forecast for Nvidia is to expect further stagnation in its revenues. Until the company's leadership recognizes that its future lies in the talent and initiative of its employees and pushes most of its power and authority deep down the chain of command, its potentially stellar future will remain unrealized.


One of several companies spawned from Taiwan's foundries over the last two decades, MediaTek came into being in 1997 when UMC spun off its consumer electronics design group. The company has made a variety of acquisitions over the years to bolster its presence in Wi-Fi, DTV, DVD and STB markets, as it clearly intends to be the master of the Digital Home. 

MediaTek has also become a power in mobile telephony, particularly in low to medium end feature phones and smartphones for China. Qualcomm has eroded MediaTek's position in application processors, but the company has an increasingly strong offering in baseband processing and is contending aggressively with Marvell in the China market.

Starting from q1 2013, MediaTek seems to have put itself on a steady and healthy growth streak. From an IP standpoint, MediaTek has a world class portfolio of wireless and multimedia properties. The MStar merger was finally completed in early 2014, further reinforcing the company's positioning for the Digital Home. It seems distinctly within the realm of possibility that MediaTek will eclipse Broadcom as the world's largest broad based SoC vendor sometime in 2015.

For seeing they saw not, and hearing they understood not, but like shapes in a dream they wrought all the days of their lives in confusion. - Aeschylus, "Prometheus Bound"

Yet despite my optimism regarding the company's long term prospects, I remain somewhat guarded for the near and medium term. The mobile computing market has saturated and consumer electronics is stagnating. The latest reported health in revenues is likely due to the same anomalous spike in STB and home gateway segments observed by Broadcom. In fact, both companies are projecting business to be flat for q4 2014. 

What is particularly surprising is that MediaTek doesn't seem to have much of anything to say at all regarding the IoT - a market for which it is pre-eminently well positioned thru its highly talented workforce and rich IP library. It is, in fact, mystifying why MediaTek isn't demonstratively touting its innate capabilities for the IoT sector. 

The company did launch its LinkIt initiative back in June, with the MT2502 running an optimized OS targeting wearables and other IoT applications. MediaTek Labs was formed in September to shepherd the company's involvement in the IoT and support development by the 3rd party community.

But all of this is being done rather quietly - in fact, almost surreptitiously - in stark contrast to the efforts of Qualcomm and Broadcom to promote their offerings for the fledgling IoT space. The entire organization of the MediaTek Labs mission comes across as very regimented and perfunctory, even cookie cutter in its sobriety. This should be setting off alarm bells across MediaTek's investor community, as it seems management does not feel the appropriate sense of urgency to find and develop new markets for its wares. 



Everything that is really great and inspiring is created by the individual who can labor in freedom. - Albert Einstein

When Zeus led the Olympians in revolt against Cronos and the Titans for control of Creation, some of the Titans defected to his side. Most notable of them all was Prometheus, who would afterwards thwart Zeus in his efforts to keep humans in abjection to his rule.

Things came to a head when Prometheus stole the gift of fire from Olympus and gave it to humanity. Wrathful Zeus had the rebelling Titan chained to a rock in the Caucausus mountains, where an eagle would feast upon his liver for all eternity.

The Stone Masons, having served a Promethean role for semiconductors in their creation of SoC technology, also find themselves victims of forces beyond their control or ability to resist:

The now chronic weakness in consumer discretionary income in first and second world economies torment the Stone Masons, suppressing and weakening their revenue growth. While consumers stringently manage their dwindling levels of disposable cash, fewer smartphones, tablets, high definition TVs, broadband routers and DVRs are purchased. This cascades downstream to the infrastructure providers for backhaul, edge routers, cable modems, networking equipment and datacenters, eventually reaching the SoC vendors, who find that falling demand and the approaching demise of Moore's Law constrain their ability to effectively add more features and value to their offerings and that their new and enhanced products have steadily fading leverage to command volume and defend margins.

However, while the fetters binding Prometheus to a rock were forged by Hephaestus at the command of Zeus, the chains and shackles restraining the Stone Masons are ones of their own making. Once daring and fearless in breaking thru the perceived confines of chip technology, the SoC companies are increasingly hog-tied by their own operational constructs that manacle employee initiative and curb the full flowering of their talent and creativity, channeling their energies into well worn paths of scheduled and controlled activity which once led to prosperity but now only reach stagnating pools and dead ends. 

Broadcom led the Stone Masons in bringing a Promethean gift to High Tech a score of years ago and in the process changed the semiconductor industry. Will they once again bring a new benefaction of technological, operational and creative fire to High Tech, or will it be one of their rivals? Then again, maybe the time of the Stone Masons is past, and like the Assyrians who clung to their ways and eventually fell to the Babylonians, perhaps the MCU houses will usurp the SoC sector and take the lead in the IoT market from them.

What indeed does the future hold for the Stone Masons? Perhaps a famous poet from old Attica said it best:

Time waxing old can many a lesson teach. - Aeschylus, "Prometheus Bound"