Internet of Things Creating Opportunity Bubble for Analog

I've written before about the impact the Internet of Things (IoT) will have on electronics sales including the special role played by analog devices in IoT. Most recently, in my Planet Analog blog entitled Digitizing Analog Sensor Data for the IoT, I argued that because of all the sensors, integrated packages including Analog Front Ends for sensor signal conditioning are likely to see growth. Here I want to take some of the numbers out there and bracket what the opportunity might be.

Last year IC Insights noted that global unit shipments of semiconductors could exceed one trillion devices in 2016. They also said in a mid-year update that Analog IC shipments were outstripping all other categories in terms of growth, forecasting nearly 9% CAGR (Compound Annual Growth Rate) for Analog IC sales from 2013 to 2018. Market data firm IHS said in March that all Industrial semiconductors, of which Analog is the largest category, would grow by nearly 10% a year to over $55B annually by 2018. Separately, IHS reported total semiconductor sales in 2014 were $355B; these figures put Industrial ICs at about 12% of total semiconductors. I think that all Analog IC sales are as much as 16% of all semiconductor sales. Using these figures my rough estimate is that Analog could be between $46B and $62B in 2015 growing at a 9% to 10% rate. For the sake of analysis, let me assume ASP (average selling price) of $0.5, and conclude that Analog IC unit shipments will be around 100B units in 2015.

Ron Exler, Senior Strategist with Saugatuck Technology (a Connecticut-based subscription research / advisory and strategy consulting service) said in a recent post: “The Industrial Internet is a term coined by General Electric. It involves the massive systems that make up society’s infrastructure. These systems include energy generation and distribution, transportation management, healthcare diagnostics, and manufacturing monitoring. Industrial Internet is the IoT applied to industrial applications with analytics to provide context-aware intelligence.” (Source: Saugatuck Technology Inc.)

In another post proving some of the latest findings, Mr. Exler notes: “It’s exciting and innovative. Yet IoT also has unrelenting concerns of security, connectivity, data processing, analysis, and manageability.” Referring to Figure 1, the Saugatuck article points out that Industrial IoT is likely to be on internal networks versus the public internet. The figure summarizes Saugatuck's findings that 55% of companies plan to address infrastructure to manage IoT in the next 3 years. The implication I draw from all of this is that a bubble of Industrial IoT spending is going to occur, or, looking at it the other way, a bubble of opportunity will occur for Analog semiconductor companies.

Figure 1

Figure 1.  Saugatuck Technology findings from their Infrastructure Survey.  The last category is IoT Management which would include most industrial applications, since they may not be exposed on the public internet, but connected to private infrastructure.  Figure and content Copyright, Saugatuck Technology, used with permission.

Figure 1. Saugatuck Technology findings from their Infrastructure Survey. The last category is IoT Management which would include most industrial applications, since they may not be exposed on the public internet, but connected to private infrastructure. Figure and content Copyright, Saugatuck Technology, used with permission.

Coming at this another way, Cisco estimates the Internet of Everything (IoE) had 10B connections in 2013, and will have 50B connections in 2020 (see Figure 2). Cisco's growth curve lines up with the bubble of opportunity over the next 3 years; you can see that by around 2018 things will be leveling off.

Figure 2

Figure 2.  Cisco estimates connections for IoE to grow from 10B in 2013 to 50B in 2020.  Figure from Cisco's IoE Internet of Everything (IoE) whitepaper.

Figure 2. Cisco estimates connections for IoE to grow from 10B in 2013 to 50B in 2020. Figure from Cisco's IoE Internet of Everything (IoE) whitepaper.

Using Cisco's growth predictions, I estimate there are about 11B connections in 2015, growing to 30B in 2018. That is adding 19B connections from 2016 to 2018. If I conservatively consider an IoE node to have at least one Analog IC for the wireless section, one for power, and one for the sensor, the 19B connections represent 57B Analog ICs. Going back to the market data, using a growth rate for Analog IC shipment value of 9.5% per year, there will be roughly 330B Analog ICs shipped over these same 3 years. In other words, the IoT (or IoE if you prefer) could represent over 15% of Analog IC unit demand from 2015 to 2018.

So there you have it. The Internet of Things could well a harbinger of a new golden age of Analog. These applications will definitely drive ICs related to RF (wireless), power/energy harvesting, and signal conditioning (including A to D converters and others). What do you think?

4 comments on “Internet of Things Creating Opportunity Bubble for Analog

  1. nhwank
    May 25, 2015

    Blaine's stimulating opportunity review reminds me of the gold rush. And the telecom bubble. As with any innovation, large numbers don't impress investors: check, yes it is BIG. Show me the use case and I show you the money. Just like the promise of high-speed fiber-optic internet in the late 90-ies: data travelled at Autobahn speeds but the toll road failed to provide on-ramps for the masses to participate. The same challenge surfaces by connecting Billions of devices. Blaine has already very accurately pointed out the challenges, which have been echoed in last week's Smartgrid IEEE conference held at Stanford: resilient communications, repeatable and scalable systems architecture, interoperability (hello Plugfest), and most importantly: (cyber)security. One of just many applications, energy, which falls into the laggard class of adopters who eventually will “cross that chasm”. It is a nuisance if your smartphone doesn't connect to your thermostat while traveling you end-up with a triple-high electricity bill. It is a mess when your wireless network is penetrated, your security system disabled, and you return from your well deserved vacation to an empty house. Although it would be easier to get those readings from your advanced meter reading device (AMI) to see if you are home. But it but could wreck your life if critical water infrastructure is breached (google Illinois Water Systems attack on Nov 8, 2011), and your tap water (exclude Californians who don't have any) gets potentially poisoned by terrorist cells. Fortunately the Illinois event only destroyed water pumps, but it clearly showed vulnerability. Optimistic as we entrepreneurs are, let's assume we get it (the technology and economics) right. As a continuous line of high-tech disasters (last weeks pipeline burst in Southern California) shows, each blessing has its risks. Can you live with this risk? Or are we turning a blind eye, hoping for the best instead of planning for the worst? 

    Do you still remember Denial-of-Service attacks? How do tiny (cheap) little micro-controllers swart against such attacks when Billions of machines interact? Back in 2003 when I was working on first broadband powerline modems at Texas Instruments, leading white goods suppliers stopped commercialization in the tracks due to the price point economics. There is no budget on the Bill-of-Material (BOM) for a Digital Signal Processor (DSP) required to process high volumes of data for high volumes of product. But that were only 15 million devices per year. Then in 2010 our start-up team invented a data decimation algorithm and architecture to bring health prognostics to critical power electronics infrastructure, transforming a DVD full of noise into a text message of actionable information. Nobody cared. The problem was solved before many knew there is indeed a problem. Shouldn't we learn from the past and first learn from the business processes, examining the telecom revolution (caused by the telecom bubble) before we techies get all excited about combing through peta-bytes (that's a lot of zeros!) of noisy data just because we can, and the money-tree is getting more greedy (again)? Or is it simply the return of the analog empire that felt overrun by the digital Gorilla during the last decade? After all, life is analog, but so much easier to manage when simplified by zeros and ones. With Moores law out-of-order for sensitive analog or power process technology, price and profitability pressures have been increasingly worse for analog. It just does not scale the same way. Physics doesn't compress. With many nuances of industrial applications, no one-size fits all. At least purchasing managers wouldn't by more (silicon) than they need. Structured ASIC you say? How do we build one in analog? Analog is custom. Let's see who has been there before, learn, adapt and iterate. I should know, I launched 14 highly successful power management product lines during the first 15 years out of my now 22-year career, and decided to get out before the semiconductor industry turns into a banking industry.

  2. eafpres
    May 25, 2015

    Hi nhwank.  Thanks for your on-point remarks and the history as well.  I was interested about your comments on power-line communications.  Oddly enough, though it did apear to die on the vine in most consumer applications, it is now built into most SoCs designed for Smart Metering.

    Regarding security, you are correct there are many reasons to be concerned.  However, in the blog I linked above on Smart Meters the SoCs I noted had hardware encryption processors along with the application processor.  So, if the power company properly implements security, then it should be nearly impossible for someone to play a man in the middle or similar attack.

    I think, and perhaps it is to this concern that you refer, that the bigger problem is the security of the data on the back end.  Possibly of even greater concern is that as the number of these big data systems grows, so does the number of people “on the inside”; i.e. the potential for insider attacks will only get larger.  As the criminals are very good at valuation of data sources, they will pay well for the best ones.

  3. raghooram
    May 27, 2015

    great post

  4. anvessh
    May 29, 2015

    nice one

Leave a Reply