Saturday, September 26, 2015

Heri Et Hodie Et Cras

Botticelli, "The Banquet in the Pine Forest", 1482-1483 (source: wikipedia)

"Si fortuna juvat, caveto tolli;
Si fortuna tonat, caveto mergi."

If fortune favors you do not be elated; if she frowns do not despond. - Ausonius

Dear Readers,
I am preparing several storylines involving Big Data - a vast topic which crosses almost all of the technology trends we've been following together on this blog for the last 18 months. The first chapter in the Big Data discussion was published here on 7/31/15 and I'll be providing a second installment on 10/9/15.

Today, though, we'll break with the blog tradition of exploring a single issue in depth and briefly cover several topics - a sprinkling of subjects we've visited before along with the completely new. In no particular order, we will lightly peruse both technical issues and business events in High Technology.

The M&A Bubble

Boethius, "The Consolation of Philosophy", 1460 (manuscript from the Bibliotheque Nationale, Paris)

"olat ambiguis mobilis alis hora; nec ulli præstat velox Fortuna fidem."
The shifting hour flies with doubtful wings; nor does swift Fortune keep faith with anyone. - Seneca the Younger

We've already reviewed the $37B Broadcom-Avago and $16.7B Intel-Altera deals intensively (on May 29th and June 6th, respectively.) The M&A activity in High Tech this year continues to be frenzied. Some highlights:
 - Cypress bought Spansion for $4B
 - NXP purchased Freescale for $11B
 - Pericom picked up Diodes, Inc. for $400M
 - Microsemi absorbed Vitesse for $389M
 - ISSI was acquired by China's Uphill Investment, beating out Cypress with a $731M bid
 - Dialog grabbed Atmel for $4.6B, winning against Intel and CEC (China Electronic Corporation, wholly owned by China's central government)

Some deals have gone sour - the most publicly visible one being Tsinghua's $23B bid for Micron, which the boys in Boise turned down based on fears of national security - based objections from Washington D.C. Nonetheless, M&A fever still has the High Tech industry in its grips. TI seems to be preparing to enter the fray and Qualcomm is reported to be studying AMCC as a potential target.

Another phenomenon which is attracting far less attention but which is no less significant is the divestiture of product divisions for strategic and/or financial reasons. NXP has already sold off its RF power products group for $1.8B to China's Jianguang Asset Management and may be preparing additional business unit sales. AMD is purportedly considering once more a move to split the company up along CPU and GPU lines or perhaps selling the GPU group outright.

Whether acquiring, breaking up or selling off companies and business units, the consequent reorganizations always create casualties in the workforce. For instance, HP's ongoing split into enterprise and consumer firms will leave 30,000 current employees without a paycheck when it's all finished.

High Tech has seen M&A activity throughout its history. Yet the current binge is completely unique, both in scale and in terms of economic conditions (which are poor compared to previous M&A peaks.) Why is all of this happening now? 

With regards to liquidating product groups or selling whole firms outright, this is a particularly propitious time in part because of one very deep-pocketed customer: China. Beijing announced last year its intention to become self-sufficient in semiconductors and seems poised to spend between $120B-$200B over the next decade to realize this goal. As can be seen from the above list of acquisitions, that effort is already well underway.

But there is more to it, of course. A combination of macroeconomic and technology forces have converged to create this particularly difficult and tumultuous environment. One can summarize them in the following list:

1. The death of Moore's law. It's a rather inescapable reality at this point that 28nm was the last node that intrinsically delivered 3P (power, performance and price) improvements together. Every node after that delivers at best two of the three and requires additional engineering effort that only adds to cost problems.

Granted, the semiconductor sector is striving mightily to squeeze the last drop of capability from silicon thru pursuit of technologies such as monolithic 3D-IC and FD-SOI. Yet though all the work being poured into such capabilities is clearly praiseworthy, it is equally clear that these will be the last optimizations for silicon. It's simply reached its physical limits. 

There are fascinating and promising substitutes under development worldwide - GaN, graphene, exotic combinations of elements such as molybdenum & sulfur, an 'aromatic ring' configuration of silicon called silicene and a number of others - including a potential return to germanium (perhaps in combination with ferroelectric technology now being explored for memories and re-purposed for general logic.) But they are all still far away from general deployment and will likely remain so for the next 7-10 years.

2. A dearth of semiconductor startups. The current $150M+ projected up front costs for chip ventures is by now a well known fact in the industry and has crushed the once spectacular silicon startup sector, along with all the new technology and market segment breakouts that these startups initiated. There are ways to drastically shrink costs and greatly increase the chance of market success thru Lean Startup concepts applied to hardware (a very detailed methodology for doing this can be viewed on EE Journal, where I wrote a 3-article series about 2 years ago.) A few working groups in the industry have made some gestures towards exploring Lean Startup methods for the chip sector, but they all seem to lose momentum and stagnate after making a press release or two because the problem looks a bit too challenging for their tastes. 

This brings to light a particularly distasteful issue blocking the forward progress of the industry - the adventurous entrepreneurs of the 70's, 80's and 90's appear to have been replaced by highly risk-averse technocrats who prefer the safety and security of industry standard specifications and regular product release schedules in established markets. To put it more plainly, the very people who are in a position to lead the industry forward into a new era of experimentation and invention apparently lack the guts to do so.

3. Cheap debt available from TBTF banks, thanks to the now 7 year old ZIRP policies of the US Federal Reserve and similar initiatives from all the major central banks. We've examined this extensively in the last two quarterly "State of the Union" series.

4. The end of the Mobile Computing boom. We've seen how the stagnation and decline of the tablet and smartphone markets affected Apple, Mediatek and Qualcomm last quarter (with Qualcomm in particular getting punched solidly in the crotch.) With consumer discretionary income declining (even at the high end) and middle class household incomes continuing to deteriorate, this market segment - the only one with real energy over the last 5-6 years - won't be springing back to life any time soon.

5. The semiconductor and systems markets have been very noticeably weakening throughout the year. A very telling indicator has been the ongoing decline in DRAM prices, vividly contrasting with the last two years of strong growth. Another is the total collapse in market forecasts. The SIA/WSTS and Gartner predictions of 4.9% and 5.4% growth at the start of the year have crumbled by 60%-70% already. Before the year is out, I anticipate they will further drop their projections to somewhere between zero and 1% for the year. We must hope that these numbers do not also prove to be optimistic.

All of the above, combined with the relentless pressure from shareholders and analysts to grow revenues and drive stock price performance, to provoke the executive management at High Tech companies into aggressively buying their way into new technology, market segments and revenue streams. There are those who think this trend will continue for some time:

Some say that the current M&A craze is driven primarily by companies positioning themselves for the nascent IoT market. Surely this is indeed part of it, but it's worth remembering that in the past most Technology companies would have done this on their own - there's nothing particularly outrageous about IoT in terms of design goals or industry standards for interfaces & communications. The strongest driver is, in this particular case, the widespread availability of abnormally low cost financing that permits financially healthy companies to quickly make the necessary acquisitions on the cheap rather than spend time and effort creating the necessary capability in-house.


"Ut varias usus meditando extunderet artis / paulatim."
Practice and thought might gradually forge many an art. - Virgil

We've spoken so far of yesterday (heri) and today (hodie) - but what about tomorrow?

Of the many technologies of the future that I've scrutinized, a very fascinating one has until now escaped my notice - that of Flexible Electronics. Combined with the growing capabilities of 3D printing, the potential range of applications is staggering.

Fortunately, one of VF's readers is up to his elbows in this field. Aniello Franzese is an electrical engineer winding up his Masters studies in conductive ink printing for electronic circuits. Based on a conversation we had about one week ago, Aniello has graciously provided a brief description of the technology, which I have included below. As such, Aniello is VF's first Guest Blogger.

One of the immediate applications for Flexible Electronics is in the realm of rapid functional prototyping. Research is under way to develop a cheap and rapid method for printing and testing circuit and board designs. The field is maturing quickly, along with parallel solutions for PCBs.

The technique used in this article is based on metal conductive ink. The ink is an organic solution containing an emulsion of nanometallic particles. After printing, the ink is further processed thru exposure to a flash lamp in a curing step which evaporates the organic material, improving conductivity. The procedure is called 'pulse forge photonic curing.'

Photonic curing is the high temperature processing of a thin film using pulsed light from a flash lamp. It is employed primarily for substrates such as plastic or paper which do not lend themselves (for obvious reasons) to conventional annealing in an oven. 

This particular technique for printing flexible circuits had to be researched to discover its boundary conditions for applications. Multiple 3cm width squares were processed with voltages ranging between 100V and 400V. The research revealed that at low voltages, conductivity is unaffected. At high voltages, though, the ink began to evaporate. Increasing the voltage to 600V actually damaged the plastic paper substrate.

The printing software used is Adobe InDesign. It is particularly suited for this application as the resolution of a printed line is identical to that drawn in the software. The best line resolution achieved was near to 10 microns. This can only be done horizontally; in the vertical dimension, the line is interrupted for very small widths, constraining the best case dimension to 70 microns. The fault lies in the printer, as the print cartridge moves horizontally. Because of this, vertical lines are deposited in a non-continuous manner.

"Felix qui potuit rerum cognoscere causas."
Blessed is he who has been able to win knowledge of the causes of things. - Virgil

Aniello is currently debating whether to immediately enter the workforce with some Technology company or to begin studies towards earning a PhD in quantum computing from Delft University in the Netherlands. For those of you who would like a genuine inventor on your technical staff, I suggest you move quickly to contact him.  ;-)


Illustration from a medieval manuscript of Chaucer's "Canterbury Tales" (source: 

"Casus ubique valet: semper tibi pendeat hamus, quo minime credas gurgite, piscis erit."
Luck affects everything; let your hook always be cast; in the stream where you least expect it, there will be a fish. - Ovid

These are turbulent times in our profession, what with consolidations, reorganizations, layoffs and stagnation in both technologies and markets. I hope that all my readers are riding out this difficult period well, with the Jolly Roger streaming at the top of the mast as you stand on the bow, surging thru the foam-flecked sea and laughing with gusto as Poseidon throws a fury of wind and wave at your barque.

Let us all remember as we toil thru the day and the week, keeping our tempers in check, our spirits buoyant and our resolution firm in face of today's trials, that we are here because we are exuberant believers in the future. We are all dreamers of a tomorrow that is better than today.

Please let me know if you'd like to hear more about anything covered today. See you all again in two weeks time.  :-)