Digital Economy Goes Green: Taking Over Energy-Efficiency Management

Abstract

Digital economy is one of the main themes of the 14th Five-Year Plan (2021–2025). The rapid development of new technologies—such as cloud computing, artificial intelligence (AI), and Internet of Things (IoT)—can support increased demand for cloud services and digital transformation at company and government department levels. These technologies can also boost the development of 5G base stations, data centers and the construction of other new technology infrastructures. The Fourth Plenary Session of the 19th CPC Central Committee in October 2019 has, for the first time, made it clear that data has become a new production factor in the era of the digital economy. We believe as the quantity and quality of the data both improve in the future, data can boost output growth in certain industries that are closely related to the digital economy and it can also provide computing power to support R&D of new technologies, serving as a source of technological innovation. New technologies represent a strong driving force for China’s digital transformation, and electricity is also a key in supporting the development of the digital economy. We believe advanced technologies can fuel the growth of computing power, and stronger computing power, in turn, can help improve technologies and support sustained growth of digital economy. In our opinion, the process of digitization can also make the conversion of electricity into computing power be more energy efficient in the future, thereby reducing CO₂ emissions and boosting the growth of the digital economy. This chapter consists of four parts. In part one and two, we analyze how technologies help cities and industries improve energy efficiency and reduce CO₂ emissions. Then in the third section, we analyze how AI and other new technologies help the technology sector become more energy efficient and achieve carbon neutrality. In the final part, we summarize the challenges that are likely to be faced throughout the process of digitization.

11.1 Digitalization to Help Cities Improve Energy Efficiency and Reduce Carbon Emissions

In digital economy, we believe AI, 5G and other new technologies will empower multiple industries and enrich our lives. This prompts us to think about the following questions: How far are we from living in the digital lifestyles? And how can we achieve carbon neutrality in the digital lifestyles? We think smart cities that integrate digital technologies and utilize data into city governance and management will improve the way cities operate, thereby enhancing the experience of citizens in their daily lives.

The concept of “smart city” has drawn worldwide attention since it was proposed by IBM in 2008. In our opinion, smart city is more than just a concept but also represents the idea of incorporating digital technologies into city governance. We foresee that the utilization of new technologies such as AI, IoT, 5G networks, and cloud computing will transform the current society into a mature technology ecosystem that can make cities smarter and solve issues throughout the process of urbanization. A research organization Markets and Markets estimates that the total market size of the global smart cities’ development will increase from US$308bn in 2018 to US$717bn in 2023, implying a CAGR of 18.4%.

What are the application scenarios of smart city? As the idea of “smart city” is becoming increasingly popular and new smart city projects are under construction globally, we expect a new city governance model will be created and promote sustainable development of the entire society. Smart city projects have been carried out in many areas, including environmental protection projects such as pollution monitoring and waste sorting projects; transportation projects such as cooperative vehicle infrastructure systems (CVIS), smart parking lots, and smart transport planning; and projects in other sectors like smart factories for manufacturers, and big data-enabled city governance projects.

While smart city projects are able to make citizens’ lives easier and more enjoyable, digital technologies can also contribute to enhancing energy efficiency and reducing CO₂ emissions in cities. How will smart city projects in various application scenarios utilize technologies to cut energy consumption and reduce CO₂ emissions? And how much carbon emission reduction can this process ultimately achieve?

In order to answer these questions, we calculate the net reduction of CO₂ emissions enabled by CVIS, smart airports, and smart cargo transport, comparing the CO₂ discharged by facilities after utilizing digital technologies and the amount of emission reduction by these projects. In addition, we calculate the drop in CO₂ emissions per 1kWh of electricity to analyze the “value for money” provided by digital technologies.

11.1.1 CVIS Projects: Enhancing Vehicle Allocation Efficiency and Reducing Fuel Consumption

CVIS, enabled by sensor detection, edge computing, and autonomous driving technologies, can obtain and share information about road and vehicle conditions via roadside units and vehicle-mounted terminals. Such systems, based on traffic flow, accident and road condition forecasts, can enable the coordination between vehicles (and also between vehicles and infrastructure), thereby accelerating crossroad vehicle throughput, reducing vehicle fuel consumption, and increasing transportation safety and redundancy.

Efficiency enhancement: Smart traffic lights, tidal lanes, and autopilot cars can help reduce traffic congestion and make the traffic system safer and improve efficiency. Smart traffic lights can adjust light cycling time based on the amount of traffic to reduce the waiting time for both drivers and pedestrians. Tidal lanes allow cars to travel in a direction that can help improve traffic flow during rush hours by increasing the number of lanes for the side with heavier traffic. Autonomous vehicles, with vehicle-mounted cameras, radars, and other devices, can analyze conditions of vehicles and their surrounding environment before braking, taking turns, changing lanes, or adjusting their speed.

Emission reduction: Smart vehicles can utilize the simultaneous localization and mapping (SLAM) technology to locate their positions and formulate a map of their surroundings. This is achieved based on the information collected by vehicle-mounted cameras and the time-of-flight (ToF) distance measurement function of their laser radars. The SLAM technology helps vehicles improve route planning and reduce carbon footprint. In addition, trucks that use vehicle to everything (V2X) communication devices can travel in a fleet to mitigate wind resistance, share traffic light information to arrange braking time, and reduce fuel consumption thereby.

To what extent can autopilot enhance efficiency? According to a research from the University of Michigan, autopilot can reduce energy consumption by 19% compared with a traditional vehicle because they can improve route planning and optimize braking efficiency through the inter-vehicle-communication (IVC) system. In addition, we assume that the total number of private cars will drop by 1% following the wide adoption of autopilot technologies. Based on these assumptions, we estimate that by 2024, the annual carbon emission reduction of vehicles will be around 118mnt empowered by the autopilot technologies. Suppose that the amount of data generated by an autopilot test vehicle can reach around 10TB (terabytes) with an Inspur high-end full flash storage of HF18000G5 server. Then there would be about 25.66mn servers needed to support the computing process with 82mnt CO₂ emissions throughout the year. In total, we estimate that self-driving technologies can reduce CO₂ emissions by 42mnt per year, which means that the CO₂ emissions reduced by servers for self-driving cars is 1.51 kg/kWh (Table 11.1).

Table 11.1 Net reduction in CO₂ emissions from self-driving technologies per year by end 2024

11.1.2 Smart Airport: Optimizing Aircraft Taxiing Distance with AI-Enabled Precision Calculation

Through accurate prediction and management, Shenzhen Airport has reduced the taxi time of each aircraft by 1–2min, cutting oil consumption by 10–20L per aircraft on average. According to the Civil Aviation Administration of China (CAAC), there are around 10.25mn aircraft takeoffs and landings in China every year, and we estimate that smart airport projects will reduce carbon emissions by about 363,100tonnes per year (Table 11.2). Additionally, we estimate that data centers for smart airports will discharge 0.24mnt of CO₂ per year, based on the number of data centers and cabinets for the 241 airports in China as of February 18, 2021, and an additional 3% data storage required by AI- and digitization-enabled smart airports. Overall, we expect that smart airport projects can net reduce CO₂ emissions by 0.12mnt per year. This means the CO₂ emissions reduced by smart airport projects would be 1.52 kg/kWh.

Table 11.2 Net CO₂ emissions net reduced by smart airports

11.1.3 Smart Logistics: Internet Platform-Based Logistics Companies to Help Reduce the Empty Running Ratio

We estimate that the current empty running ratio for cargo trucks in the logistics industry is around 40% in China. An internet platform-based logistics company operating under the model of technology companies such as Uber would notably reduce this ratio. According to the Ministry of Public Security, China had 29.44mn cargo trucks as of June 2020. We estimate that CO₂ discharged by trucks will fall by 69.5bn kg, if the empty running ratio drops to 20% (from 40%) and the empty mileage falls 147.2bn km per year. Assuming this logistics company requires 0.1mn data center cabinets, it would discharge about 8.5bn kg of extra CO₂ per year. Therefore, we estimate that internet platform-based logistics companies can reduce CO₂ emissions by 60.99bn kg per year in total or 8.14kg/kWh (Table 11.3).

Table 11.3 CO₂ emissions net reduced by smart cargo transport projects

11.2 Industrial Internet Empowers Enterprise Production to Achieve Cost Reduction and Efficiency Increase

Industrial internet projects are developing rapidly in China thanks to the wide deployment of 5G networks. As of March 2021, over 1,100 industrial internet projects were under construction in China. We note that the industrial internet has been applied in various scenarios by companies in multiple industries, such as Shanghai Commercial Aircraft Corporation of China, Shanxi Huayang Group, Sany Heavy Industry, and Xiamen Port. 5G technologies have facilitated industrial upgrades and transformation. The architecture of the industrial internet platform is displayed in Fig. 11.1.

Fig. 11.1

Source Industrial internet platform whitepaper released by Alliance of Industrial Internet in 2017, CICC Research. Note SaaS refers to software as a service, PaaS refers to platform as a service, IaaS refers to infrastructure as a service

Architecture of the industrial internet platform.

How will the industrial internet contribute to energy conservation and emission reduction? The industrial internet utilizes software platform and big data analysis technologies to process data collected by sensors in order to help industrial companies realize automatic control and smart corporate governance. Therefore, companies can use the industrial internet to improve manufacturing efficiency, reduce energy consumption and cut CO₂ emissions.

What are the implications of the industrial internet? According to the Ministry of Industry and Information Technology (MIIT), digital transformation of the economy is an inevitable trend. China should grasp the right development opportunities throughout the process of digitalization and informatization, given existing foundation and advantages to coordinate digital industrialization and industrial digitalization. In addition, China should also integrate a new generation of information technologies in manufacturing sectors, as well as mix manufacturing with other service industries and speed up the development of the digital economy. The ultimate goal of these fundamental pursuits is to promote the development of the real economy in China.

11.2.1 Industrial Internet: Enhancing Energy Efficiency Through Monitoring and Managing Energy Consumption Data

We are optimistic that the idea of carbon neutrality will promote the demand for downstream software applications, such as systems for energy monitoring, early warning, and other subdivisions. The increase in demand would bring business opportunities to downstream companies that provide digital solutions. For example, SaaS companies in energy management and control can help clients reduce energy consumption through analyzing energy consumption data collected via IoT and providing early warnings and solutions to clients. Companies that operate industrial cloud platforms can also help clients reduce cost and enhance efficiency, on the back of their cloud platforms and industrial apps.

State Grid released its carbon neutrality action plan in March 2021. According to the action plan, the firm will transform the power grid into an “energy internet” by strengthening the innovation and integration of technologies such as AI, IoT, big data, cloud computing and mobile internet technologies in the energy and power industry. This plan will also improve the interconnectivity of different energy sources and support alternative energy-based electricity and energy storage projects. State Grid also plans to accelerate the establishment of information collection, perception, processing, and application projects in order to promote data sharing on different energy sources and tap the value of these energy sources. According to the action plan, the firm aims to establish a world-leading “energy internet” by 2025.

• Smart energy management: Enesource Inc. provides industrial internet SaaS services to help clients reduce energy consumption and enhance efficiency. It has rolled out a series of solutions to improve energy efficiency in buildings, data centers, industrial companies, and commercial firms. Enesource can monitor energy efficiency online to help clients receive refined energy management services and reduce their CO₂ emissions. Specifically, it helps a national leading thermoelectricity company in Shanghai to develop smart energy services, reducing. Its energy costs by about Rmb15mn and cutting CO₂ emissions by around 44,200tonnes per year.

• Smart manufacturing: Cloudiip, a subsidiary of Boncloud (300166.SZ), is a leading industrial internet platform in China. Cloudiip supports multiple industrial application sub-platforms such as the big data platform for steel and iron companies, and the cloud service platform for boiler companies. At the same time, it has also launched a series of mobile apps to help industrial companies reduce energy consumption and improve manufacturing efficiency. For example, the big data platform for steel and iron companies can increase the value created by a blast furnace by Rmb24mn per year; the cloud platform for wind power companies can help reduce maintenance expenses by 30% per year; and the cloud platform for boilers can cut energy consumption by 15% per year.

11.2.2 Application of Big Data Technologies Brings Accurate Energy-Saving and Efficiency-Improving Solutions in Power and Water Sectors

The wide application of big data technologies makes it possible to process massive amounts of data. Big data platforms based on data analysis algorithms can visually monitor the data collected by sensors, make precision estimates, and improve the decision-making process. These platforms can help power and waterworks companies monitor data pertaining to electricity and water usage, thereby improving electricity and water management.

China Southern Power Grid has utilized big data technologies to provide electricity data services in Guangdong province. The company, via the electricity big data platform, can monitor electricity data at factories and identify high energy-consuming devices. For example, China Southern Power Grid has helped a manufacturing company identify the energy-intensive, fixed-frequency equipment in its factories and reduce energy consumption by 0.5mn kWh per year after installing frequency converters in these devices. China Southern Power Grid has also helped a group of companies reduce electricity consumption since 2019. According to China Southern Power Grid, electricity cost at its clients can drop 10–20% thanks to its electricity big data platform.

Xi’an Waterworks Company utilizes the smart waterworks platform to monitor information about water treatment and coordinate production and management and enhance the process of fault handling. This process enhances the timeliness of equipment maintenance, and improves planning, scheduling and decision making of the company. With this platform, the company can coordinate production data, videos, management, equipment, and employees. The smart platform has helped the firm reduce its labor cost by 3.6%, equipment breakdown ratio by 15%, suspension ratio by 20% and energy consumption by 20%.

11.3 AI and Other New Technologies Help the Technology Industry Reduce Energy Consumption and Enhance Energy Efficiency

In previous sections, we analyzed how technologies help reduce energy consumption and cut CO₂ emissions in different industries, and calculated the net reductions in CO₂ emissions facilitated by technologies. However, the technology industry itself is also energy intensive. In the following section, we analyze how AI and other new technologies have been utilized in the telecommunications industry to improve energy efficiency, and how 5G base stations and data centers have adopted multiple measures to reduce energy consumption.

11.3.1 How AI Technologies Help Companies Reduce CO₂ Emissions

The AI algorithm can be used in the operation, maintenance and management of internet data centers (IDC) to optimize their power usage. AI technologies have been utilized in equipment breakdown forecast and analysis, system fine-tuning, and internal services. Companies, via these technologies, can monitor the environment around data centers and then they can use a real-time cooling system to allocate resources and reduce energy consumption. Many companies—such as Google, Global Data Services (GDS), and Zhongxing Telecommunication Equipment (ZTE)—have already utilized AI technologies in their business operations.

• Google: In 2016, Google began to use the DeepMind AI system in its data centers. This system has reduced energy consumption by more effectively controlling their servers, cooling systems and other modules. Cooling expenses at Google data centers have dropped 40% since then.

• GDS: GDS is one of the first domestic companies to utilize AI technologies in their data centers. The development and design of its weak current system and software should ensure that AI technologies can be applied, so as to make AI algorithm-based temperature control possible. It has also utilized robots to patrol data centers and detect equipment breakdowns, thereby reducing equipment maintenance costs and improving early warning systems.

• ZTE: The firm has developed an AI algorithm system based on its own AI Explorer platform. This system can collect data about the cooling and electricity systems before conducting data governance on the AI-enabled platform. It also utilizes AI algorithms to improve the operation strategy, which can reduce energy consumption by 15%.

In the 5G era, AI technologies can improve the adaptability of base stations to reduce their energy consumption. AI-enabled smart energy efficiency solutions can help networks under different standards reduce energy consumption. The AI system can use historical data to build learning models, and continues modifying the models based on the real-time data. ZTE began providing AI-enabled energy efficiency solutions for domestic telecom carriers in 2019. This product has been utilized in more than 100,000 residential communities, helping base stations reduce energy consumption by 10%–15%. We estimate that electricity consumed by every 1,000 base stations can be reduced by 1.5–2mn kWh per year.

• Recognizing energy-saving scenarios: The AI system can automatically monitor the energy efficiency of base stations in different residential communities. It can be connected to the interfaces of base stations and the Oracle Management Cloud (OMC) platform. Base stations can be operated under the energy efficient model in accordance with changes to traffic volume and other business indicators.

• Business forecast: The AI system can estimate the load of networks, based on AI algorithms, base station data, and the load model of the training business. With the load forecast, the system can take measures to regulate traffic volume and improve user experience.

• Choosing energy efficiency strategies: The system can improve learning algorithms to choose, fine-tune, upgrade, and execute proper energy efficiency strategies. Such efforts can also help companies achieve energy efficiency goals and KPI targets. The selection model of AI-enabled energy efficiency strategies is shown in Fig. 11.2.

  Fig. 11.2

  

  Source Research on AI-enabled energy efficiency technologies for 5G stations published on Application of Electronic Technique in 2019, CICC Research

  AI-enabled energy efficiency strategies.

11.3.2 Improving Equipment, Station Locations, and Networks to Help 5G Jointly Built and Shared Base Stations to Reduce Energy Consumption

Energy consumption of the telecommunication industry has been increasing in the 5G era. However, we think the total energy consumption in the construction of 5G networks will drop due to improved equipment and advanced technologies. According to China Communications Standards Association, no-load power consumption of main equipment in existing 5G base stations is around 2.2–2.3kW, and the full-load power consumption stands at around 3.7–3.9kW. We estimate that the average power consumption of 5G base stations will drop to around 3.3kW by 2025 (vs. 3.85kW in 2019), as equipment is likely to become more energy efficient and base stations will no longer operate under the full-load model thanks to the use of hibernation technology and other new technologies for the base station. In addition, we estimate that the overall power consumption of 5G base stations will be around 79.3bn kWh in 2025, as 2G and 3G networks will be gradually dismantled and the growth of new 5G base stations will be mild (Table 11.4).

Table 11.4 Power consumption by base stations of three major telecom carriers in China

11.3.3 Software and Hardware Technologies: The Trend of Upgrading Base Structures of Telecom Infrastructure

Base-station equipment: Enhancing energy efficiency of active antenna units (AAU) will notably reduce energy consumption. RF amplifiers are more energy intensive than other devices in base stations. As a result, we think improving technologies for key devices—such as power amplifier modules, digital intermediate frequency (DIF), baseband modules, and transmitter receivers—will noticeably improve the energy efficiency of AAU. For instance, we think the power consumption of base-station equipment will drop and the performance of 5G systems will improve, if companies can improve semiconductor technologies and upgrade semiconductors to make chips smarter and enhance their integration and processing capability.

Base-station locations: Using software and AI technologies to reduce energy consumption of base stations in residential communities. AI technologies can analyze traffic volume in different hours and in different locations, and then they can improve energy efficiency solutions for different base stations during different hours. Such solutions can help reduce energy consumption without weighing on user experience.

Networks: Multi-network coordination technologies can help existing networks achieve energy reduction targets. Given the use of 5G technologies, we think 4G networks using TDD/FDD technologies will coexist with 5G networks that utilize NR technologies. According to China Mobile Research Institute, multi-network coordination technologies can apply clustering and neutral network algorithms to improve the energy efficiency of 5G equipment. Such technologies can both be used in 4G and 5G networks. According to China Mobile Research Institute, companies have also developed the 4G multi-carrier energy system (MCES). Tests on existing networks show that the MCES can reduce the power consumption of 4G networks by more than 0.4mn kWh per 10,000 residential communities. We think the power consumption reduction will be even more notable, after base stations are connected to 5G networks. The plan for energy-saving technological advancements for 5G base stations is displayed in Fig. 11.3.

Fig. 11.3

Source CMRI, CICC Research

Timeline of energy-saving technologies for 5G base stations.

11.3.4 Data Centers: Cloud Computing and Advanced Refrigeration Technologies Help Improve Power Usage Effectiveness (PUE)

Data center architectures have evolved from centralized computing in 2000 to distributed computing in the PC era and then back to centralized computing in the new era. In the 1960s, computing and storage resources were stored in mainframes, which served as the computing power supply centers for companies. Only business customers could afford these expensive devices. As a result, computing resources were concentrated in companies in the era of mainframes. In 1979, IBM launched private computers (PC) and started the PC era. Computing resources began to be operated separately, shifting from mainframes to PC. Since 2000, the computing model has become centralized again thanks to data centers that could operate more than 1,000 servers.

Data centers are the underlying infrastructure in the era of the digital economy; they play a crucial role in converting electricity into computing power. Large data centers implement centralized management of computing and storage resources (Fig. 11.4), and users can share resources through the internet. This model is becoming increasingly popular, and the concept of “cloud computing” is gradually developing (Fig. 11.5). Data centers represent the upstream of the IDC industry value chain. They provide server trusteeship and other infrastructure services for cloud companies, internet companies, and government customers. Data centers convert electricity into computing power, as their machinery equipment and temperature & humidity control devices consume electricity.

Fig. 11.4

Source Juimg.com, CICC Research

Large data centers manage computing and storage resources.

Fig. 11.5

Source Sangfor.com, CICC Research

Cloud computing concept diagram.

Collective computing power can notably reduce energy consumption at data centers, contributing to the process of achieving carbon neutrality. According to Pike Research, the use of cloud computing technologies reduced the energy consumption of global data centers by 38% in 2020. AWS believes that collective computing resources can notably enhance computing efficiency. Microsoft estimates that Azure collective data centers are 72%–98% more energy efficient than traditional data centers, as they can enhance the IT maintenance efficiency, IT equipment efficiency, data-center infrastructure efficiency, and regeneration capacity.

Standard and modular electromechanical devices can help data centers reduce cost and enhance efficiency. We think the structure of the machine room system is the key to the competence of operating IDC. Arranging the structure of this system requires the capability to predict the IT load capacity, the electromechanical layout, and the heat-sinking conditions to support higher power density in the future.

Electricity and refrigeration resources represent the core electromechanical device services provided by IDC. IT and refrigeration devices account for around 85% energy consumption of IDC. The use of IDC cooling measures should follow local conditions. In northern China, devices can be cooled via indirect evaporation as the environment is drier and colder, thereby reducing the use of cooling water. In southern China, IDC utilize the immersive liquid cooling technology, using liquid to reduce the heat generated by central processing units (CPU), memories, and other IT devices. IDC also utilizes distributed energy supply, high-voltage direct current power supply, and modular UPS technologies to reduce power losses and improve PUE.

Sharing electromechanical resources and using modular base stations can also help reduce energy consumption and enhance efficiency. Prefabricated modular data centers represent a pre-engineered product. The infrastructure units in such data centers are pre-assembled in factories before being re-assembled at construction sites. Data centers that utilize the modular design can be constructed rapidly. They share cabinets and other infrastructure in order to save the space in machine rooms and reduce electricity cost. Modular data centers also enjoy policy tailwinds in China. The Three-Year Action Plan for Cloud Computing (2017–2019) released by the MIIT encouraged companies to improve technologies and products for environmentally friendly and modular data centers. For example, Chindata provides modular cooling devices and power distribution devices for data centers, thereby shortening the construction period of data centers by 30%–40%. US companies also use the modular design to make the allocation of machine room resources more flexible. As a result, a machine room can operate in conformity with standards for tier-2, tier-3, and tier-4 data centers, cabinets with different unit power dissipation can expand their capability, and companies can provide data centers for wholesale and retail customers. We think modular technologies can show companies’ competence in utilizing IT equipment technologies. However, the iteration cycles for IT and network equipment technologies last shorter than that for infrastructure. We think modular data centers are key to improving the planning for the structure of machine rooms so as to make electromechanical modules more applicable and more affordable and enable such modules to work in conformity with standards for the rapidly-changing information and communication technologies. In our opinion, enhancing the efficiency of modular and standard equipment means that companies need to keep the entire structure in their mind. The detailed advantages of modular data centers are listed in Table 11.5.

Table 11.5 Advantages of modular data centers

11.4 What Are the Challenges Ahead?

There is a long way to go before reaching carbon neutrality. In our opinion, investors should be neither too optimistic nor too pessimistic. In order to realize the goal of net zero carbon emissions, it is necessary to be aware of potential risks and challenges ahead, and in particular we need keep an eye on the following challenges.

The penetration rate of the industrial internet remains relatively low. According to the Industrial Internet Industry and Economic Development Report (2020) released by China Academy of Information and Communications Technology (CAICT), the economic value-added scale of China’s industrial internet industry reached Rmb3.1trn in 2020, accounting for 2.9% of the total GDP. The industrial internet has been gradually used in petrochemical, steel & iron, electronics, and information industries but the penetration rate is still only 2.76% and there is still much room for improvement. We think this industry should consider the following questions. How to utilize data to shorten the commercial investment period and extend the investment return period? How to improve the inclusiveness of their platforms and the connectivity of data in order to utilize data as a production factor to the utmost degree.

The application of energy-saving systems in many cases remains at the data collection level. The energy conservation platform enabled by the industrial internet technologies consists of three layers—the edge layer is for data collection, the platform layer is for the industrial PaaS (platform as a service), and the application layer is for industrial apps. We note that in real applications, some companies only use this system to collect data, and the data they collect has not been utilized to help reduce energy consumption. We believe that the industry in the future should think about how to quickly realize data empowerment to shorten the commercial investment cycle and generate returns over an extended period of time. How to achieve platform compatibility and data connection, and maximize data production factors?

The costs of developing and maintaining AI systems are generally high. For a wide range of small and medium-sized companies, it is necessary to balance the trade-off between the benefits of higher energy efficiency and the cost of AI systems. The use of AI systems to reduce energy consumption is facing headwinds from the increased cost of system development, data storage and algorithm updates, as the computing power of such systems continues to increase. Meanwhile, the huge energy consumption brought by the increase in computing power should also not be underestimated. Small and medium-sized companies should be aware of the trade-off between the use of AI systems to improve energy efficiency and the energy consumed by such systems. How to reduce the cost of developing and maintaining AI systems, how to help companies in different industries to develop and adopt AI systems, and how to use AI technologies to reduce energy consumption and cut CO₂ emissions are the questions we need to consider.

The feasibility of using 5G technologies to reduce energy consumption remains to be seen. China Mobile Research Institute (CMRI) expects that by 2022, the use of 5nm process baseband chips will be realized, the penetration of GaN power amplifiers will also reach 90% with the development of process, semiconductor material, and RF system technologies. In the meanwhile, CMRI estimates that the overall power consumption will drop 8% YoY in 2022. In addition, through the sharing of baseband units (BBU) following the use of 5G networks will likely remove the need for hardware board configuration thereby reducing the power consumption. However, it remains to be seen whether using 5G technologies to reduce energy consumption will be effective since most of these developments are still developments. In addition to considering making technological breakthroughs, we suggest thinking about the following questions. How to improve the allocation of low-frequency resources in order to reduce construction of base stations? How to help telecom carriers, to the utmost extent, co-build and share base stations?

We see risks and challenges ahead in the journey towards carbon neutrality. We think companies and industries should be aware of these risks and challenges and stick to the ultimate goal of achieving carbon neutrality.