Get email delivery of the Cadence blog featured here
Steve Mollenkopf, the CEO of Qualcomm Incorporated, gave one of the keynotes at CDNLive here in Silicon Valley. Just a little background in case you know little about Qualcomm. They started in satellite communications. Then they pioneered CDMA technology that was initially used in the US, Korea, Japan, and other places around the world in 2G phones. They went on to develop technology that underlies later generations as well. They made CDMA technology available for other companies to license, but to ensure that the technology got traction they developed the baseband and infrastructure chips themselves, and initially even had a joint-venture handset company with Sony. As 3G and 4G came along, they continued to develop modems and also APs, integrating them as SoCs and grew dramatically to where they are today, the #1 fabless semiconductor company in the world, and #4 overall (behind Intel, Samsung, and SK Hynix).
In the early days of CDMA, outside of the rich world relatively few people had a phone. Now, thanks to a combination of Moore's Law and increasingly high volumes, the industry is shipping close to 2B phones per year, most of them smartphones, and pretty much everyone has access. In developing countries, especially in Africa, they are bypassing the whole landline stage and going straight from nothing to cellphones. Qualcomm alone shipped nearly 1B chips for phones last year, under their Snapdragon brand. Steve predicted that there will be 8.5B smartphones shipped in the 2015-2019 era. There are only 7B people in the world, so now you know why pants come with at least two pockets. Smartphones were the first truly mass-market SoCs.
However, these volumes are about to be dwarfed by the Internet of Things (IoT), which will ship in 10s of billions and grow in time to 100s of billions of installed devices. The CEO of almost every company now has one question on their mind: "What happens when electronics disrupt my business?"
The semiconductor industry existed before smartphones, of course. But the dynamics of the industry have changed. In the PC era, companies tended to specialize, with different companies supplying the CPU, the graphics, the memory, the networking, and so on. But mobile needs many technologies at the same time since they need to be on the same chip. IoT will be like that, too.
Qualcomm will need to continue innovation in multiple areas beyond CPU and modem—with GPU, ISP, audio codecs, and more. People will need experiences that are seamless. At MWC in Barcelona, Mark Zuckerberg predicted that virtual reality is going to be a really immersive environment and will drive next-generation cellular technology (since Facebook owns Oculus who make the Rift VR goggles, that might not be a totally objective viewpoint!). But VR requires integration of high-quality audio, accurate and responsive motion tracking, huge pixel quality, and quantity.
Some of the biggest market segments have heavy government involvement and regulation: healthcare, automobiles, education. This has made them slow to change but they are also so large that even a small change in efficiency is a few percentage points of GDP. They represent both a huge opportunity for the electronics and semiconductor industries, but also a challenge since, unlike a cell-phone modem, there are a lot of requirements for safety since products are life-critical.
One area Qualcomm has been investing in is drones, with a product called Qualcomm Snapdragon Flight. This reduces seven large circuit boards with a total area of 189cm2 to a single small 23cm2 board. This doesn't just add more technology, it reduces weight, which in turn makes the drone able to stay in the air longer.
Another area where Qualcomm has been investing is in the datacenter. They are one of the significant partners of ARM attempting to disrupt the Intel domination of the datacenter. The potential profits are huge but it is also a long-term strategy with no guarantee of success.
The next generation of mobile phones is known as 5G although that makes it sound like all the standards are in place, which they are not. The bandwidths are going to be so high that this will be "fiber to the curb" without the fiber, in much the same way as few of us bother to plug in an Ethernet cable to our laptops any more (my MacBook Air doesn't even have a socket for one, which caught me out once when I stayed in a hotel that only had wireless access in the lobby but all the rooms had patch cables). 3G and 4G have been characterized as connecting people to the internet. 5G will connect people too, of course, but it will also have a massive IoT component connecting things.
Slides courtesy of Qualcomm Inc
See also: Tom Beckley's CDNLive Keynote: Addressing Complexity and Safety Challenges
See also: CDNLive Keynote: Expanding Your Opportunities with System Design Enablement
Previous: Jim Hogan and the Early Days of Virtuoso
Next: TI and UI: Texas Instruments' Experience with the Common User Interface