Qualcomm once predicted 2019 would be a big year for 5G cellular. In its dreams.
Last December, Qualcomm staged 5G demos for press and analysts using an Ericsson base station and prototype handsets from Samsung and Motorola. They said lots of services were going to switch on and many OEMs were going to ship handsets in just a few months.
The reality is only 1% of the smartphones that ship this year will support 5G, according to the latest forecast from Gartner. What’s worse, the report projects total smartphone sales will fall 2.5% this year to 1.5 billion handsets — even lower than the 1.7-1.8 billion Qualcomm stated in the lowered forecast from its quarterly report this week.
Several factors drive the slowdown. Smartphones are mature with few changes generation to generation, so replacement cycles are lengthening, Gartner said. On the OEM side, no one wants to be the last to release a 4G handset, and only a few want to be the first with 5G as carriers roll out services at a stately pace, a few cities at a time.
Clearly 5G will be big, but it’s clear this wave won’t really get started until next year. It will be interesting to see what percentage of 2020 phones support 5G, presumably most of them. So, the wave is coming, but the dynamics suggest it was not so important to have 5G chips ready this year.
Speaking of waves, I wish I had a crystal ball to predict what percentage of 5G handsets will support millimeter wave bands in 2020. Qualcomm rushed a mmwave module to market last July that was a marvel of miniaturized wireless engineering.
Almost no one used it. Seven months later, the cellular chip giant announced a smaller version of the module which will largely wait on the shelf for handset sockets appearing next year.
Much of the 2020 smartphone story for EE Times readers will be about which RF front-ends OEMs chose for which bands. Those decisions will have a big impact on the handsets’ budgets for power, real estate and bill of materials.
A 5G mmwave RF front end could cost an OEM twice what it paid for a 4G version, according to Jon Erensen, a Gartner analyst for mobile and consumer chips. “The positive is memory is in the low phase of the price cycle, freeing up some wiggle room,” he said.
In some conditions, mmWave bands will deliver multi-gigabit streams of data. But to keep the data flowing, users will have to learn how to point their handsets at the nearest base station — if they can see the diminutive 5G small cells. This behavior likely will become the subject of some funny YouTube videos in 2020.
Even engineers at AT&T and Verizon are still getting the hang of dealing with signals and their reflections at 28-GHz and up. Frankly, this is a solution in search of a problem. I regularly watch Netflix movies on my LTE phone. Why do I need more data?
Like many people, I’m curious how long it will take for Apple to get its own 5G silicon into handsets. Even before it bought Intel’s cellular modem group, it was aggressively hiring cellular engineers in San Diego, Qualcomm’s home town.
I expect an Apple baseband or RF chip could emerge in the second generation 5G iPhone in 2021. But a whole baseband-to-RF chain probably will take longer. Thus, I expect the “multi-year” deal Apple struck with Qualcomm tapers off dramatically starting at year three.
It’s not clear how much RF expertise and IP Apple bought from Intel. The x86 giant was one of the pioneers (with arrows in its back) of WiGig, so it has plenty of mmWave experience with 60 GHz.
Evensen of Gartner is a bit skeptical.
“The number of companies that have tried to do modems and failed is staggering and several were internal projects. You need to hit at one time everything that leading vendors have learned over multiple years and each generation is a moving target,” he said.
Besides 5G basebands and RFFEs, watch for new and bigger image and AI processors in 2020 handsets, he added. A handful of aggressive IP providers might will vie for a shot at a big design win here, given how important the camera is to consumers.
But you can forget about being the next big chip inside a smartphone at least until the 5G transition is over. Given the demands of 5G silicon, there just won’t be any room to put another chip or any power left to run it. But who cares, smartphones are declining anyway, right?