Facing the challenge of increasing labor costs and competition in manufacturing, China’s State Council launched the “Made in China 2025” initiative in 2015.The initiative aims to transform China’s manufacturing industry to one that is innovation-driven. It calls for more efficient and more intelligent manufacturing through the integration of production processes with the internet. A major force in achieving this goal is Internet of Things (IoT), which, generally speaking, refers to the connecting of smart devices to the internet. In fact, the IoT sector exceeded RMB 900 billion (USD 142 billion) in 2016. China’s IoT sector is continuing its rapid growth with a compound annual growth rate of over 25%. Because artificial Intelligence (AI) acts as the brain to IoT body, the two are natural partners. In July of 2017, the Chinese State Council laid out a new development plan with aims to become the world’s premier AI innovation center by 2030.
The top-level initiatives on advancing AI and IoT (AIoT) innovation have a profound impact on local-level participation. Private companies are spending more and more on technology research and business development. Baidu, for instance, established a new AI lab and received government funding to do so. Likewise, Alibaba plans to invest USD 15 billion in AI technology. Additionally, the company is upgrading its operating system (YunOS) strategy to better focus investment in IoT.
Investors have also demonstrated their increasing interests in AIoT. Qualcomm Ventures, the venture capital arm of the giant telecommunications products company, has invested in nine Chinese AIoT startups across different applications.
Around the year 2003, an upsurge in IoT in China was unable to make substantial progress. 2017, however, became the epoch of AIoT thanks to the above-mentioned top-down policies. Six emergent trends have pushed through and will continue to help secure large-scale AIoT development in China.
1. Reduction in sensor prices and sizes
Nobody can deny that sensors are the core technology of IoT. The good news is that they have become and will continue to become more affordable. In 2014, the average cost of sensors was USD 1.30 and in the year 2020, it is expected to decrease to USD 0.38, constituting a 70% reduction in cost. Uses of these sensors will continue to diversify with consumer and developer demand. Consumer Physics, for instance, allows users with its sensor-equipped SciO products to scan food and obtain information regarding freshness and healthiness.
Microelectromechanical systems (MEMS) miniaturize the size of sensors, and more compact sensors further reduce production costs. MEMS coupled with nanotechnology have been applied in various fields within physical, chemical, biological, and medical domains. MEMS have penetrated into consumer electronics, home appliances, recreation, and the automotive industry.
2. Cost decrease in data processing
Today’s data processing capacity has increased 30,000 times from that of 20 years ago. IoT data processing usually takes place on the on the cloud. Moore’s Law — the observation that the number of transistors in devices tends to double every two years — has been the accelerator of cloud-computing performance.
But it is now the post-Moore’s Law era. Intel has slowed the pace of rollouts of new generations of chips with smaller and denser transistors. Meanwhile, some startups such as DeePhi are now focused on distributing deep neural networks over both the cloud and end devices.
3. Increasing bandwidth and efficiency
Cisco predicts that approximately 50 billion IoT devices will be connected by 2020, an estimate that doubles the amount of connected devices we have now. The prospect poses a huge change to the internet. Inevitably, we will need sufficient bandwidth and speeds to move all of this data.
In the past 10 years, bandwidth has increased considerably. In 2005, the fastest speed of CPRS was 100kbps, whereas by 2015, 4G provided speed up to 3000 times faster at 300Mbps. 5G tests in 2018 are expected to provide maximum speeds of 1Gbps. It is optimistic to say, speeds might achieve a blistering 10Gbps, with the proper technology advancement.
Equally important is the fact that the cost of bandwidth is decreasing. From an efficiency standpoint, 4G is 4 times better than 3G and 5G will be 3 – 5 times greater than that of 4G. In terms of efficiency, the cost of 5G ends up being lower. Currently, China Mobile offers its more than 11 million IoT subscribers affordable monthly data plans from only RMB 6 for 2 megabytes to RMB 208 for 11 gigabytes.
4. Sophisticated AI technologies for various applications
AI technology propels IoT efficiency. AI technology, including voice recognition and image recognition, has matured significantly. Progresses in Natural Language Processing (NLP) is catching up as well. For example, Sogou voice input can correctly recognize 97% of speech. Unisound’s Far-Field Voice Input Processing helps users to locate a microphone 5 to 10 meters away. Tricorn equipped with NLP technology enables users to interact with devices using natural language for multiple speech interaction rounds. In 2017, Alibaba partnered with KFC China to debut facial recognition payments.
5. Rising labor costs spur automation
In 2016, average manufacturing wages in China rose to USD 3.60 per hour, considerably higher than other developing countries like Sri Lanka, where manufacturing wages average only $0.50 per hour. ‘Made in China’ isn’t cheap anymore. Instead, it has become a double-edged sword to encourage market solutions to increase productivity. Automation is a natural result. IoT, a key component of automation, has a proven record of increasing industrial manufacturing management efficiency. A McKinsey Global Institute report shows automation could raise productivity growth globally by 0.8-1.4% annually.
6. Burgeoning purchasing power
China is quickly moving towards a consumption-driven economy. Consumption-contributed GDP growth increased from 50.2% in 2014 to 71% in the first three quarters of 2016. China’s GDP per capita in 2016 was USD 8,123, a 5% increase over that of 2014. The growing demand for consumption amongst the middle and upper-middle class is a major driver for the proliferation of premium, smart AIoT devices.
With the lowering cost of hardware and IT infrastructure, improving technologies and growing purchasing power, AIoT may likely bring about economic growth and technological innovations. Equally important, we can expect it to bring significant profit, prosperity, and transformation to the manufacturing industry.
（Top photo from diot.com.cn）