laitimes

Is the AI model an "electric swallower"?

author:CBN

(The author is Associate Professor Li Zhiguo, School of Management, Fudan University)

Who will decide the future of AI?

Today, when everyone is keen to discuss computing power, another hidden problem has quietly emerged.

By 2027, the electricity consumption of global AI data centers will be comparable to that of the Netherlands, Sweden and other countries.

Electricity may become a constraint on computing power.

Sam Altman, founder of Open AI, said: "The future of AI depends on clean energy breakthroughs. ”

Nvidia founder Jensen Huang pointed out: "The end of AI is photovoltaic and energy storage. ”

As artificial intelligence technology moves towards its "peak", the power industry will also usher in new great changes.

What impact will AI technology have on China's industrial economy, and how can we seize the opportunity through industrial reform?

01 The end of AI is new energy

Artificial intelligence has driven a new round of scientific and technological revolution and industrial transformation, and more and more industries and fields will use artificial intelligence to achieve digital and intelligent transformation, which will continue to promote the upgrading of trillions of traditional industries. Artificial intelligence continues to create new market demand, which will continue to promote the emergence of trillions of new industries. From professional design, text creation to intelligent driving, it is becoming a new application scenario in the era of artificial intelligence.

In the future, the development process of artificial intelligence will be non-linear, and the potential will be beyond imagination. By combining with applications in many vertical fields, the new infrastructure of artificial intelligence will continue to reduce costs and increase efficiency for the industry, and continue to create new demand, new business models and new economic growth points. In a sense, artificial intelligence is a "new power" that is transforming major industries such as healthcare, transportation, entertainment, and manufacturing, enriching the lives of countless people.

For China's industrial economy, AI technology can greatly improve production efficiency, especially in the manufacturing industry, in terms of intelligent production, quality control, and supply chain optimization, AI can reduce production costs, improve capacity utilization, and improve yield. At the same time, AI technology will play an important role in new product development, process innovation, business model innovation, refined management and intelligent decision-making.

In the process of AI technology development, as large models "bloom", the problem of energy consumption growth has gradually become prominent. Every aspect of natural language processing is inseparable from power support, including hardware devices such as GPUs, data storage and transmission, etc., which require a lot of power.

At present, the resources and energy consumed by human beings in production and life have exceeded 30% of the entire earth. There is a saying that "if everyone on the planet lived the life of a European, it would take 3 Earths, and if everyone lived the life of an American, it would take 5 Earths." ”

Artificial intelligence has brought human social economy into the fourth industrial revolution, and we have passed from the era of mechanical power to the era of energy drive, and then to the era of information drive. In the past, oil was the most important factor of production, and field data resources became the most important factor of production. However, looking back, I found that behind the information drive, energy support is still inseparable.

Transportation, energy, and communication will form a new paradigm matrix, with transportation pointing to power, energy pointing to speed, and communication pointing to fusion, which will form the development paradigm of the fourth industrial revolution, which requires a new energy system to match.

Constrained by resource capacity and climate change factors, humanity must move towards green development. The core of green development is to rebuild the energy system, especially to build a clean, low-carbon, safe and efficient power system, that is, a new power system with new energy as the main body.

In the future, we will build supercomputing centers, data storage, and data transmission ...... As long as the entire industrial chain of AI-related equipment is logically the first to achieve greening. The significant rise in AI-fueled energy consumption is likely to accelerate the reform of the entire energy system.

02 Smart Grid Accurate Prediction of "Peak Shaving and Valley Filling"

So far, in order to maintain the stable operation of the power system, whether it is the production power generation end, the transmission and distribution end, or the demand side, it is objectively necessary to focus on thermal power. There are two reasons for this: one is stability, and the current power system itself is built on the basis of thermal power. Second, it is controllable, which is convenient for "peak shaving and valley filling" to adjust the balance of electricity consumption.

In the case of photovoltaic power generation, although the cost has been greatly reduced at the power generation end, the cost is still very high at the energy storage and transmission end. At the same time, photovoltaic power generation also has intermittent and fluctuating shortcomings in power supply, which will have a certain impact on the operation of the power grid system and is not conducive to stable power supply.

But on the other hand, AI itself can help future-proof the power system. In the long run, there are three main development trends of new energy, smart grid, and energy Internet.

One is smart grid management. Artificial intelligence technology can improve the efficiency and stability of power grid operation, and can achieve accurate prediction and dispatch optimization of power load, including analysis of electricity consumption, electricity consumption habits, seasonal changes and other factors of various regions, enterprises and users, so as to more accurately arrange power generation planning and energy distribution.

In this way, it can not only improve energy utilization, but also solve the problems faced by the new power system, such as adjustable capacity, reduced inertia, and weakened ability of the system to imbalance.

The second is to strengthen the supervision of the electricity market and carry out electricity trading. For example, the transaction data of the electricity market can be monitored and analyzed in real time to mine patterns from massive data, identify abnormal patterns, ensure fair competition and stable operation, and optimize the transaction matching mechanism through intelligent algorithms to ensure low transaction costs.

The third is energy efficiency management and energy conservation and emission reduction. Through intelligent sensing equipment and big data analysis, real-time monitoring of energy consumption at the industry, enterprise, and factory levels can identify energy waste and inefficient energy utilization problems.

03 The key to the era of "wind and photovoltaic" is energy storage

The era of large models has brought important opportunities for China's renewable energy industries such as photovoltaic and wind power. By applying AI technology, these industries can better predict and manage energy production, optimize equipment maintenance, improve grid operation efficiency, facilitate market-based transactions, and promote green and low-carbon transition.

Artificial intelligence technology not only helps to improve the utilization efficiency and sustainable development of renewable energy, but also plays a positive role in promoting the transformation of China's energy structure and the replacement of traditional energy with clean energy.

China is a global leader in clean energy technology, especially in the photovoltaic industry, and we have been involved in the formation of the industry since its formation and led the industry to grow rapidly, and now it has matured and even entered the "red ocean" stage.

At present, the entire industrial chain of China's photovoltaic industry has covered more than 70% of all links in the world, and some links even exceed 90%. Our capacity is expanding faster than the growth of market demand. Whether it is silicon wafer production, diamond wire cutting for monocrystals, or technical application frameworks, China has a clear advantage in the world.

As for the wind power industry, China has an advantage in the processing and manufacturing of key components, and our power generation capacity is rapidly increasing through cost reduction and large-scale production.

In addition, hydrogen energy, as one of the ultimate energy sources in the 21st century, has the characteristics of clean and high energy density, but there is still obvious room for progress in technology and safety, and the cost reduction and transmission of hydrogen energy are issues that need to be solved in the future development of the industry.

Whether it is photovoltaic, wind power or hydrogen, energy storage technology is the biggest common challenge in the current clean energy industry.

Nvidia founder Jensen Huang has previously publicly said: "The end of AI is photovoltaic and energy storage, don't just think about computing power, if you only think about computers, you need to burn 14 earths of energy." ”

Although energy storage technology is constantly improving, the cost is still very high at this stage, which is the main reason why photovoltaic and wind power cannot completely replace thermal power.

In the future, China has great potential for new energy development, especially in terms of improving energy efficiency and optimizing energy distribution. The development of smart grids and smart energy systems will have a positive impact on the operational efficiency and stability of the entire power system, thereby improving overall energy efficiency.

The views expressed in this article are solely those of the author.

(This article is from Yicai)

Read on