Cloud AI 100
Qualcomm announced its new chip — Cloud AI 100 – will speed up artificial intelligence (AI) work. Qualcomm will be aiming to diversify beyond its mobile phone chips business and into a fast-growing market dominated by Nvidia and Intel.
Qualcomm at an event in San Francisco said it plans to begin testing its new Cloud AI 100 chip with partners such as Microsoft later this year, with mass production likely to begin in 2020.
Qualcomm Cloud AI 100 offers:
# More than 10x performance per watt
# 7nm process node bringing performance and power advantages
Analysts believe chips for speeding up inference will be the largest part of the AI chip market, Reuters reported.
Nvidia has released special chips for the inference task and Intel is working with Facebook on one that will be released later this year. Cloud computing vendors such as Amazon.com’s Amazon Web Services and Alphabet Inc’s Google Cloud unit are also making their own inference chips.
Cristiano Amon, Qualcomm’s president and the chief of its chip division, said the San Diego based company is taking a different approach by aiming to serve the smaller, simpler data centers that are proliferating.
In order to serve smaller “edge” data sites, Qualcomm is focusing on AI chips that consume small amounts of electricity and generate little heat – a specialty it developed when making chips for mobile phones, which run on small batteries and live in pockets.
Rivals such as Intel and Nvidia make more powerful chips that dominate in centralized data centers that suck up electricity and need complex cooling systems.
Qualcomm earlier failed in its efforts for selling CPU chips based on mobile phone chip technology to take on Intel.
Qualcomm Snapdragon 730, 730G and 665 Mobile Platforms
Qualcomm also launched reference designs associated with its Snapdragon 7 and 6 series targeting the mid-range segment of the smartphone market.
Qualcomm revealed that commercial devices based on Snapdragon 730, 730G and 665 are expected to be available in mid-2019.
Improved imaging and advanced AI implementation are key features in the new reference designs. The Snapdragon 730 features the fourth, and latest, generation of Qualcomm AI Engine, which was previously only available in Qualcomm’s top-line Snapdragon 855 chipset targeted at flagship smartphones.
# 2x the power of its predecessor
# provides up to 4x overall power savings
# runs at 20% less power than Open GL ES
# up to 35% performance uplift for each core
# up to 15% faster graphics rendering
Qualcomm demonstrated its leadership position in edge AI processing through its presence in most flagship smartphones, including Samsung’s Galaxy S and Note series, Google’s Pixel series, and Xiaomi’s Mi series.
Lian Jye Su, principal analyst at tech advisory ABI Research, said: “Qualcomm’s intention now is to democratize the use of advanced AI applications, notably for improved imaging, computer vision, secure authentication, and depth sensing.”
ABI Research said shipments of smartphones powered by AI engines will exceed 455 million in 2023.
Qualcomm is not unique as other competitors, including Hisilicon, with its Kirin 970 chipset, and MediaTek, with its Helio P90 chipsets, have already enabled users of mid-range smartphones to enjoy AI features that were previously only accessible to those who could afford flagship smartphones.
Qualcomm is likely to lead AI democratization in smartphones and become a key strategic chipset supplier for many OEMs since Hisilicon is a captive vendor for Huawei smartphones and MediaTek Helio P90 shows a weak AI ecosystem.