MADISON, Wis. — As the IoT World gets its mojo working in San Francisco this week, I find myself bombarded with tweets from the show floor. And I hear the devil on my shoulder saying, “Damn, girl! You’re missing all those stories.”
As a reporter, that’s a horrible feeling. After all, I’m not there. I’m on my way to catch a flight to China.
But long trips come with opportunities. First thing, I can catch up on all the reading I never seem to have time for.

On top of my stack is an Internet of Things (IoT) Report recently put together by McKinsey & Company and the Global Semiconductor Alliance (GSA).
This crisply written report is not just educational, but it confirms — and answers, to some extent — the skepticism about IoT and its challenges, which have worried chip industry insiders for the last few years, although few have expressed their misgivings in so many words.
What distinguishes this report from other market research firms’ IoT predictions is its sharp focus on the semiconductor industry. Rather than being just an IoT cheerleader, the report offers some surprisingly honest and insightful views on IoT from executives at GSA member companies.
As much as all of us want to believe in Cisco’s famous forecast that 50 billion devices and objects will be connected to the Internet by 2020, we also recognize our struggles in figuring out how to get from here to there.
Unresolved issues
IoT has unresolved issues that could not only trip up the nascent IoT segment, but also come back and haunt the chip industry for a long time.
As the McKinsey-GSA report points out, a major challenge is the security and privacy of user data. If that’s not a big enough headache, the report listed many other issues. They include the difficulty of building customer demand in a fragmented market, a lack of consistent standards, a fragmented marketplace with many niche products, the challenge of extracting more value from each application, and technological issues that affect the IOT’s functionality.
Such findings are no longer surprising. But I have a feeling I’m not alone sensing that a few prescient practitioners in the chip industry have begun tamping down the financial market’s high hopes for an IoT gold rush.
A good example is how NXP Semiconductors’ CEO Richard Clemmer responded to a financial analyst’s question during a recent earnings’ call.
The analyst said that some competitors are already starting to break out IoT in their fiscal numbers. When he prodded Clemmer to speculate on the trajectory of IoT, Clemmer responded: “It sounds like the dot-com phenomena.”
He added that at NXP, “We prefer to call it the smarter world as opposed to the Internet of Things.” He added that “smarter world” is a term “a little bit more descriptive than just the IoT, which is still relatively undefined.”
Well, that’s just anecdotal evidence.
Ambiguity
Almost by necessity, McKinsey-GSA’s IoT report is full of anecdotal evidence. They compiled it by interviewing 30 GSA members who are senior executives at semiconductor companies or at companies in adjacent industries that are part of the IOT ecosystem, such as network equipment and industrial automation. They also surveyed 229 semiconductor executives at GSA member companies.
As the report points out, their interviews “revealed some ambiguity about whether IoT would be the top growth driver for the semiconductor industry or just one of several important forces.” It said, “Our survey of executives from GSA member companies also showed that they had mixed opinions about the IoT’s potential, with 48 percent stating that it would be one of the top three growth drivers for the semiconductor industry and 17 percent ranking it first.”
Many chip industry executives downplay a lack of consistent standards as a major roadblock to IoT. We all live in the world of multiple standards, they say.
But as shown in the chart below, included in the report, there are myriad competing standards for low-range and medium-low data rates. They could “hinder growth for many IoT applications,” the report pointed out. Aside from missing interoperability and consortia battles, there are “additional incompatibilities in higher communication layers, such as 6LoWPAN vs ZigBee.”

Crowding the space marked as 2 in the chart above, there is standard white space for low-data-rate, low-power, high-range applications such as smart grid. In that specified area, Wi-Fi and LTE have high power consumption, while alternatives with low power and wide range, such as LTE Cat. 0, 802.11ah, Sigfox, and OnRamp, “are in very early stages and compete against each other,” the report pointed out.
The McKinsey-GSA report suggests, “Given the current uncertainty, semiconductor players should pursue a hedging strategy — in other words, focusing on selected standards that are likely to gain widespread acceptance but planning for alternative scenarios.”
Next page: Value chain




>> classifying IoT devices into archetypes based on their specifications. The companies can then create a single platform to cover each archetype,
Some archetypes i can think of: bluetooth accessory. Low-power long-range(or mesh) wireless mcu for sensing in the factory.Smart, High compute(cortex-m4) sensor node with wireless. High accuracy sensing mcu. Self-powered chip.
It's not different from what the mcu business did before the IOT.
>> I don't want to be a party pooper here, but chip companies' desire to do things other than chips is admirable on its face but aspirational at best.
I think ARM(with the mbed) and Atmel(with arduino) did some good wrk, that at the very least helped them reduce support costs and make their platforms more sticky. That's valuable financly. And i can see the mbed-os(to be released later this year) making some money for ARM.
And embedded designers today choose chips alot due to software and support, so it's not like chip companies have a choice.