Advertisement

Blog

We’ve Let the High-Tech Genie out of the Bottle

When Bill Shockley and his colleagues at Bell Labs in northern New Jersey developed the first transistor in the late 1940s, a djinn was released that would spread its magical influence for the next six decades to every corner of the globe. Successive waves of technological transformation, each one larger than the last, swept across the face of the earth, leading us to a world of today that was the stuff of only the most fantastical science fiction just a generation ago.

From those first crude examples of field effect and bipolar junctions in germanium arose the current wonderland of our interconnected world. The aggregate knowledge and activity of humanity is transmitted across a plethora of wired and wireless channels, cutting across social, ethnic, physical, cultural, and political boundaries, even triggering revolutions and collapsing governments.

Many changes stemming from the invention of the transistor have been wonderful. Whole new industries have been created. Information that used to take months or years to find and gather is now accessible in minutes. A farmer eking out a living in the Sahel can send a free text message to his brother making a living as a cab driver in Minneapolis. A smartphone system design team in Korea can have a real-time multimedia conference call with an SoC verification team in San Diego, a group of software developers in Bangalore, or an embedded IP supplier in England or Israel. A consumer anywhere in the world with the current model of that Korean smartphone can shop online, search for a local store or restaurant, pull up customer reviews to help him/her decide on a choice, and get driving directions instantly. Even Shockley's djinn must be mesmerized at all the change that has been wrought since his breakout into the modern era.

Yet there has been a dark side increasingly evident in the pervasiveness of technology in our lives. A dozen separate instances of retail and financial companies finding their secured credit card accounts getting hacked en masse, intelligence and law enforcement agencies worldwide accessing private, corporate, and government data in ways that range from borderline to egregiously illegal and the rising threat of belligerents remotely bringing down national electrical grids, communications networks, or financial institutions are all part of the Pandora's Box of dangers and ills that the unleashing of the high-tech djinn has brought to our day and age.

What is also becoming starkly evident besides the good or bad nature of any particular facet of our new high-tech epoch is how the industry itself is changing. In stark contrast to the last four decades of increasingly heady prosperity, we who are direct participants and contributors to the high-tech sector cannot help but feel that the current holiday season and our industry have something fundamental in common: both are entering into winter.

As an example: Virtualization has been a hot area for the past few years. Enterprises are looking for help in employing Hypervisor, Hadoop, Docker, or other virtualization architectures for their datacenters. SDN has arisen as the network counterpart to server software abstractions. There are security reasons for this, of course. But the greater impetus for growth in the virtualization market is the interest of enterprises in maximizing the value of their fixed infrastructure investments thru increased efficiency. Stated differently — the corporate/business sector isn't building out server farms and LAN layouts in anticipation of future growth, but is much more interested in keeping their costs under control by getting the most out of what they already have. This is a very telling nuance in corporate spending patterns.

Some scoff at the idea of any kind of high-tech slowdown — after all, Apple released the iPhone 6 and sold 10 million in the first weekend, the Internet of Things is all over the news, and lots of Web 2.0 companies have soaring Wall Street valuations. Clearly, the composite high-tech industry is perfectly healthy and 2015 will see the industry break out into explosive growth — right?

Let's take a look at this quantitatively. The semiconductor industry is the barometer for all of high tech. If systems houses are selling routers and servers, service providers distributing STBs & base stations and consumer products firms are shipping HDTVs and smartphones, the microelectronics industry will be busily designing and manufacturing chips to meet the system level demand. The two graphs below show how semiconductors have been doing over the last six years (2008-2013.) The first graph breaks out semiconductors by broad categories (note: MPU aggregates CPU, MCU, and DSP) and the second graph is a sum of chip industry sales as a whole. Revenues are in $B.

There are some interesting things to observe in the above two charts. Every product category took a hit in the 2009 financial crisis and grew substantially in 2010. However, starting that same year, only Logic and Optical products grew to any significant extent. All the other categories were either essentially flat or tending down. The odd thing is that industrial, scientific, and medical (ISM) and automotive were supposed to be growth markets for silicon during these last several years — yet the analog, sensor, and discrete segments in particular do not reflect such prognostications.

The second chart shows that the total semiconductor industry growth for the past six years has been quite light — a compounded 3.5% per year. This means that chips have been growing at a pace that has barely stayed ahead of even the most optimistic inflation estimates.

When combined with the saturation of mobile computing (smartphones and tablets) along with the stagnation of consumer and computing markets (HDTV, STB, home gateway, desktop, and laptop), the conclusion is inescapable — high tech has entered its first winter of falling or zero growth.

Is this winter a permanent phenomenon of a now mature industry, or is it an event that is part of a broader cycle? We'll examine this in several installments over the following weeks, beginning with the actual quantifiable status of high-tech companies across the spectrum. From there, we will peruse what engineers and developers are doing to drive their offerings forward over the short and medium term, concluding with a look at what's on the long-term horizon.

14 comments on “We’ve Let the High-Tech Genie out of the Bottle

  1. Ariella
    December 30, 2014

    “The second chart shows that the total semiconductor industry growth for the past six years has been quite light — a compounded 3.5% per year. This means that chips have been growing at a pace that has barely stayed ahead of even the most optimistic inflation estimates.”

    It's true that moderate inflation rates hover around 2-3%. But It take it that we're talking about increase in growth rather than just revenue, so it's not necessarily a sign of stasis. 

  2. apek
    December 30, 2014

    Hi Peter,

                  While IoT startups have been flourishing, why is Analog and sensor development not good. Is it that now there is only mostly software side of technological development with IoT and no progress with hardware development? Can only software development succeed while ignoring hardware progress? When there was a market dip in 2009 because of stock market crash in 2008, do you see 2014 as a replication of 2007 highs? Could you let me know if you are able to make any correlations between 2007 market highs and 2014 highs based on sales of semiconductors?

    Best regards,

    Apek

  3. Himanshugupta
    December 30, 2014

    3.5% growth is very moderate if we consider the semiconductor industry to be the base of all the development happening in electronics and related industry. The problem i see is that the hardware is already too far ahead of the software and while there are improvements in the semiconductor chips year-on-year the same is not true with the software upgrades. I want to back my claim with the hotly emerging fields such as IoT or embeded systems or connected devices. We have got all the hardware available at a very cheap price but not many companies have been able to draw a reliable software architecture.

  4. Eyetalian Stallion
    December 30, 2014

    An excellent observation, Himanshugupta. I fear, however, that SW may catch up to HW, and perhaps too soon. With 28nm representing (for all practical purposes) the Last Node, hardware innovation is already grinding to a halt. This will eventually limit the 'canvas' upon which SW developers can create, which is a very bad thing all around.

  5. Eyetalian Stallion
    December 30, 2014

    Hello Apek,

    The data that I cited in the editorial is from the SIA. I do not have data before 2008 and unfortunately cannot research the possibility of the trend which you suggest. If you do find a way to access the earlier data, I think it would prove to be very interesting. 🙂

  6. Eyetalian Stallion
    December 30, 2014

    Ariella – consider what kinds of growth rates used to be normal for the semiconductor industry. If you look at the history of semiconductors up to the dotcom bust of 2001, the years before that showed 14-16% growth, even including dips every 3-4 years. The current rate of growth is nothing short of pathetic by comparison. What makes it even more worrisome are two factors: (1) the de-facto end of Moore's Law, and (2) the saturation of the mobile computing market (tablets and smartphones.) Mobile computing has been the primary growth driver over the last 5-7 years. I think we are entering the hardest period ever experienced by High Tech. How we break out of it is something I will be editorializing about in future posts.

  7. Ariella
    December 30, 2014

    @Peter are you taking those years before the bust as the benchmark of what it should be?

  8. Himanshugupta
    December 31, 2014

    Agreed Peter below 28nm the price per chip is high so there might not be real benefits but 45nm and 65nm are still the workhorses for chip makers and that might continue for sometime unless there is real need to go to lower nodes. I think building reliable softwares is a tedious task. I remember having an argument with a colleage about how cheaply one can buy hardware component and with a little knowledge of the embedded software one can really start some kind of product development with little investment.

  9. ahdand
    December 31, 2014

    @himanshugupta: Well there are many for different prices but remember there is something called QUALITY too 

  10. Taimoor Zubar
    December 31, 2014

    “While IoT startups have been flourishing, why is Analog and sensor development not good. Is it that now there is only mostly software side of technological development with IoT and no progress with hardware development? Can only software development succeed while ignoring hardware progress?”

    @apek: I think the revenues greatly depend on the competition in the industry and not necessarily on the quality of the output. The hardware companies may have been doing well but the revenues declined or remained steady because there was a lot of competition within the industry that made prices low and hence pushed the revenue down.

  11. Taimoor Zubar
    December 31, 2014

    @Peter: Over the past decade, how has the trend been between hardware and software industries? Has the growth been co-linear? Has software been able to dominate there as well and leave hardware behind? Anything major that's happening in this decade which wasn't there in the last one?

  12. FLYINGSCOT
    January 3, 2015

    I was surprised to hear that semicon is growing so poorly when one considers how prevalent new gadget sales are.  I wonder where these data came from and if there are other sources available to do a sanity check.

  13. Eyetalian Stallion
    January 3, 2015

    The data is from the SIA.

  14. Eyetalian Stallion
    January 3, 2015

    I honestly don't know, Taimoor. It is a subject worth researching, without question. 🙂

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.