Somebody actually REDUCED their IoT forecast?

Some analysts are starting to get the idea that their credibility is worth something. Research firm IC Insights has actually dialed back its latest IoT semiconductor projection through 2019, although still calling for what would be quite robust overall growth.

I’ll reiterate a position I’ve taken many times: when it comes to embedded markets with long design-in and deployment lifecycles, it is really hard to maintain a CAGR of over 20% for a period of several years. Big CAGRs are common in the beginning of a curve since the growth is projected over a small revenue base. However, once traction sets in and the market reaches a significant size, projections had better show the CAGR rolling off to be trusted.

What also typically happens is the design-in curve proceeds pretty much according to plan, but the deployment curve representing production orders slows down from the first projections. OEMs are an optimistic bunch before they run into the problems of obtaining actual end customer orders and qualifying integrated systems together, both of which can take much longer than just designing and building IoT sensors and gateways. This has played out time and time again in telecom infrastructure gear, and it is pretty much the same scenario for the industrial IoT.

IC Insights offers this latest curve, explaining they have pulled out some $1.9B from their forecasted revenue lowering the outyear CAGR from 21.1% to, coincidentally, 19.9%. (Before the conspiracy theorists start chirping, nobody paid me to write this – it just caught my attention.) I double-checked the math, their text says 19.9% and their chart says 19.1%, probably a typo.

Tearing down the revenue stack bears out my comments. IC Insights says they are seeing connected cities applications moving slower than previously thought, and it is the dominant term in the equation with semiconductor content at $15.7B in 2019. Connected vehicles is moving faster but on a much smaller base of revenue, $1.7B in 2019.

Wearables are an interesting part of this projection, and I’m guessing they ran the math on this before the Apple pivot to fitness this week. They see $3.9B in 2019, but I suspect that relies heavily on an assumption of Apple Watch 2 success. Connected homes account for $1B, and industrial IoT connections add up to $7.3B, both unchanged from the previous iteration of the projection. For the visually inclined, I charted the 2019 numbers quickly.
Most interesting to me as an IIoT pundit/researcher is the applications that get most of the ink – wearables, connected cars, and home automation – are only about ¼ of this forecast. Keep in mind this IC Insights projection is for the semiconductor content in these segments, not the device/system end value. My guess is industrial IoT gets bigger based on design wins I’m seeing, and consumer IoT gets smaller (or at least gets pushed out), but we’ll save that for another time.

This work is part of a much bigger IC Insights study on semiconductor content, titled IC Market Drivers – I don’t have access to the full report. The briefing for this IoT semiconductor forecast update is available online free of charge. It seems like IC Insights has a pretty solid methodology and isn’t afraid to trust their numbers and revise projections.