Low-cost smartphones like the Moto G5 introduced a few months ago are shipping with soon-to-be-outdated chips from Qualcomm, which has announced successor chips.
Qualcomm on Monday introduced the Snapdragon 630 and 660, which are massive upgrades to chips used in low-cost smartphones introduced over the last six months or so.
And in a few months, you'll be able to buy low-cost smartphones with these new chips, with prices starting at US$200. The handsets will feature LTE download speeds equivalent to that in the iPhone 7 and have graphics processors capable of capturing 4K video.
Additionally, the low-cost smartphones will charge up faster than ever. The chips support Qualcomm's Quick Charge 4, and smartphones can charge up to 50 percent in just 15 minutes.
The Snapdragon 660 is like a tweaked version of the Qualcomm 821 chip -- used in top-tier phones like the Samsung Galaxy S7 -- for low-end phones. It is based on the Kryo architecture, which is also in the 821.
Smartphones with the Snapdragon 660 will have 2560 x 1600 pixel screens. The chip has about a 20 percent faster CPU than its predecessor, the Snapdragon 653. Its Adreno GPU is about 30 percent faster.
Some of the big upgrades with the new chips are best highlighted by the low-end Snapdragon 630, which succeeds the 625, which is used in the Moto G5 Plus. Phones with Snapdragon 630 will have Bluetooth 5 and 802.11ac, compared to Bluetooth 4.1 and 802.11 b/g/n in the Moto G5 Plus.
The Snapdragon 630 also supports USB-C 3.1, while the Moto G Plus has an older micro-USB 2.0 charging port. The eight-core Snapdragon 630 chip is not based on the Kryo architecture, but on standard ARM Cortex-A53 designs. That is similar to the Snapdragon 625, but the 630 has faster clock speed.
The phone with Snapdragon 630 will be able to grab 4K video, while the Moto G5 Plus is capable of 1080p capture. The GPU in Snapdragon 630 is about 30 percent faster than the Snapdragon 625/626. The CPU is roughly 10 percent faster.
A secondary feature of the new chips is support for artificial intelligence via APIs like Facebook's Caffe2. Developers can write applications where cameras can recognize images in real time. Smartphones with the new chips will take advantage of the CPU, GPU, and digital signal processor for machine learning.
This technology could bring natural language processing to smartphones. While the use of multiple processing units isn't the most efficient way to create neural networks for machine learning, it's a good start.
Also, the chips are made using Qualcomm's 14-nanometer process. That's not quite as advanced as the top-line Snapdragon 835 -- used in the Samsung Galaxy S8 phones -- which is made using the 10-nm process.
The new chips are also designed for tablets. Some vendors may choose to use the chips in drones, robots, and internet of things devices.