Last year, the advent of ChatGPT caused a global sensation, and this year, the spotlight has shifted to Sora. Both innovations originate from the OpenAI, an American open artificial intelligence research center, which boasts technology that is advanced enough to be indistinguishable from human output. In the previous year, AI was able to learn human language; in the current year, it has demonstrated the ability to understand human vision. Experts project that the realization of artificial general intelligence (AGI) may occur within the next two to three years. Nevertheless, as AI technology continues to advance at a rapid pace, the critical challenge of energy consumption has become increasingly prominent.
At the World Economic Forum, Sam Altman, CEO of OpenAI, cautioned that the next generation of generative AI systems will consume a significantly greater quantity of electricity than previously anticipated, potentially posing a significant challenge to energy systems. This perspective is shared by Nvidia CEO Jensen Huang, who emphasized that the future development of AI depends not only on computing power but also on addressing energy consumption issues.
Zhou Hongyi, founder of 360 Group, has repeatedly stated that the primary constraint on AI development is energy. Top AI companies are increasingly concerned about electricity shortages, highlighting the need for efficient cooling solutions provided by advanced Industrial air conditioning systems. Experts predict that while computers accounted for less than 2% of global electricity consumption in 2018, this figure has now risen to 10%, and by 2030, it could reach one-fifth of global electricity consumption. These forecasts underscore the crucial importance of energy independence in the ultimate phase of AI development.
On March 11, the A-share new energy sector experienced a surge, with CATL’s stock price soaring by 14.46%, driving up related stocks. This spike was attributed not only to Morgan Stanley upgrading CATLโs stock rating but also to recent discussions about AI’s energy demands. Although neither Jensen Huang nor Sam Altman explicitly stated, “The end of AI is photovoltaics and energy storage,” they highlighted the enormous energy requirements of AI technology.
In a speech, Jensen Huang mentioned that improvements in computer performance reduce the need for endlessly increasing the number of computers and data centers, thereby conserving significant amounts of energy. Altman also pointed out that the two critical “currencies” of the future are computing power and energy, and without breakthroughs in energy, achieving technological goals would be impossible.
In reality, the energy consumption of generative AI like ChatGPT is staggering. OpenAI’s ChatGPT processes about 200 million requests daily, consuming as much electricity as 33,000 households. The energy consumption of generative AI searches is four to five times that of traditional searches.
China has been proactive in addressing this challenge. In 2022, China’s infrastructure computing power reached 180 EFlops, with data centers consuming 270 billion kWh of electricity. By 2025, computing power is expected to exceed 300 EFlops, with electricity consumption reaching 350 billion kWh. In 2023, China added 292 GW of wind and solar power, and by the end of 2024, the total installed capacity of wind and solar power is expected to surpass the 2030 development target of 1.2 billion kW.
At the 14th China International Energy Storage Conference, experts unanimously agreed that accelerating the development of new energy storage technologies is essential for integrating high proportions of renewable energy and is a hot area of competition in the international equipment manufacturing industry. In 2023, the global cumulative installed capacity of energy storage reached approximately 294.1 GW, with new energy storage accounting for 30%. China’s cumulative installed capacity of new energy storage reached approximately 32.2 GW, representing a 196.5% year-on-year increase.
In conclusion, the future development of artificial intelligence is inextricably linked to energy. While AI technology progresses, the substantial energy consumption issue must be addressed, which will drive rapid growth in photovoltaic and energy storage markets. In the context of AI and new energy integration, early investment in the energy sector will be crucial for future technological advancement.
