I have gotten a fair amount of feedback on my notes from earlier in the week. I am poking holes in a lot of growth stories, which has provoked a fair amount of good discussion.
After publishing some of my recent materials, I realized there is an unspoken link between two themes that I have been writing a lot about lately. In my piece on semiconductor consolidation, I say that there is no growth left in the industry. And then a few days later I say that the Internet of Things (IoT) is not going to be a monolithic giant market, but rather a host of niches. These two things are closely related, as many people in the semiconductor market think that IoT is going to be the next big wave. Semis have had twenty years of great growth from PCs, then cell phones and smartphones, and now many people hope that IoT can be the next leg, to keep the party going.
I think this is unlikely. Maybe a better way of putting it is that the industry will continue to grow, but at a much slower pace than it has enjoyed in the past. I am not saying that innovation will end, nor I am saying that there is never going to be a growth spike from some amazing new product category. But at the very least, IoT is not going to provide the same sort of growth that PCs and smartphones have.
There are two ways that I think about this – dollar content continues to shrink and competitors continue to enter the fold.
Let’s do some rough math. The PC industry sold about 250 million units last year (desktop + laptop). The silicon content in a PC is (generously) $300 per device – including CPU, memory busses, etc. That works out to about $75 billion in revenue (Intel did $56 billion in revenue last year, so pretty close for rough math). The math for mobile phone works out to 2 billion units, but only $75 in silicon content, working out to $150 billion. By contrast, let’s assume we live in a world where we ship 20 billion connected devices for IoT every year. The ASP decline on this figure is likely to be even worse, down to $7 per device. That equates to a $140 billion market, nothing to sneeze at, but actually a bit smaller than smartphones. And remember, I am comparing last year’s market for real products with some future market for imaginary products. By the time IoT reaches real volume, all these numbers will have eroded away through Moore’s Law.
|Annual Units (m)||Silicon content (ASP)||Total ($ m)|
Put simply, IoT can not wholly replace the smartphone market. Semiconductors can continue to grow, but that growth will not feel the same as they heady growth to which we have been accustomed.
This is the math underpinning the cautious tone of my big IoT report (D2D #13). From where I sit today, it looks like the overwhelming majority of IoT systems will not need big complicated processors, nor will they need lots of memory. Some industry verticals may go for something beefier, but if we want to talk about 50 billion IoT units (let alone 500 billion), these are going to have to be super cheap devices. All of this points to massive compression in the bill of materials (BOM) for IoT relative to smartphones. Yes, there will be lots of cool things we can do when we embed memory into IoT sensors and build widely dispersed databases, but those will only be deployed a very small set of circumstances anytime soon.
There is also a second reason why we can expect big compression in IoT dollar content. There is no big new technology or radio standard on the horizon. IoT systems will be built using Wi-Fi, Bluetooth and possibly Zigbee. Those are standards which are fully mature. We can build perfectly awesome IoT systems using these standards as they are today. And what do we call components where we have already reached ‘good enough’? We call them a great market from low-cost up-and-coming chip companies in China. I think there is a huge opportunity for Chinese companies to use IoT as their entry into the global leagues of chip markets.
To be clear, I am not some prophet of doom and gloom. The semiconductor industry is not entering some horrible contraction. But people should be prepared for things to change. The industry will have a very different landscape in just a few years.