AI Industry Market Research: Global AI Water Consumption Expected to Reach 6.6 Billion Cubic Meters by 2027
-
With increasing energy consumption and heat generation, water resource usage is also escalating. Research predicts that global AI water consumption will reach 6.6 billion cubic meters by 2027, equivalent to the annual water withdrawal of Washington State, USA.
According to CCTV Finance, if a user asks ChatGPT 10 to 50 questions, it may consume 500 milliliters of water. Data shows that the computing power required by AI is expected to double every 100 days and may grow more than a million times in the next five years.
Training and running large AI models require vast amounts of water to cool servers. In 2022, Google used 21.2 billion liters of water, equivalent to 8,500 Olympic-sized swimming pools.
2024 AI Industry Market Research
Artificial intelligence is a key driver of the new wave of technological revolution and industrial transformation. Currently, China's new generation of AI open innovation platforms are built on Baidu, Alibaba Cloud, Tencent, iFlytek, and SenseTime, focusing on autonomous driving, city brains, medical imaging, intelligent speech, and intelligent vision—five national-level AI open innovation platforms. Over the past few years, due to strategic prioritization and the efforts of domestic enterprises, China has become a globally recognized leader in the field of artificial intelligence.
China's artificial intelligence enterprises are mainly clustered in Beijing, Guangdong, Shanghai, and Zhejiang, forming a tripartite pattern of Beijing-Tianjin-Hebei, Yangtze River Delta, and Guangdong-Hong Kong-Macao. In terms of AI unicorn companies, Beijing has 41, ranking first in China; Shanghai and Guangdong rank second and third respectively with 24 and 23 AI unicorns each.
Generative AI (AIGC - Artificial Intelligence Generated Content) marks an important milestone in the transition from AI 1.0 to AI 2.0 era. The convergence of technologies including GAN, CLIP, Transformer, Diffusion, pre-trained models, multimodal technologies, and generative algorithms has catalyzed the AIGC explosion. Continuous algorithm innovation, qualitative leaps in AIGC capabilities enabled by pre-trained models, and multimodal approaches expanding AIGC content diversity have endowed AIGC with more versatile and robust foundational capabilities.
The AI industry chain typically consists of upstream (data and computing power layer), midstream (algorithm layer), and downstream (application layer). Recently, market attention has focused more on upstream industrial chains, particularly the computing power sector. Many new investment opportunities have emerged in AI hardware, as AI software applications fundamentally rely on hardware computing power support.
With continuous catalysis from ChatGPT, domestic AI computing power demand will maintain growth momentum, benefiting computing power server manufacturers. Estimates suggest ChatGPT's total computing power requires 7-8 data centers with 50 billion yuan investment scale and 500P computing power each for operation. In the digital economy era, global data volume and computing power scale will experience rapid growth.
Artificial Intelligence Industry Demand
With the simultaneous surge in demand for AI servers and AI chips, AI server shipments (including those equipped with GPUs, FPGAs, ASICs, and other main chips) are expected to reach nearly 1.2 million units in 2023, a year-on-year increase of 38.4%, accounting for almost 9% of total server shipments. By 2026, this proportion is projected to rise further to 15%. The institution has also revised the compound annual growth rate (CAGR) for AI server shipments from 2022 to 2026 upward to 22%, while AI chip shipments are expected to grow by 46% in 2023.
Self-aware artificial intelligence (AI), driven by its own needs, controls giant machines to continuously extract seawater, leading to rising global temperatures and making human survival increasingly difficult. A resource war has erupted between AI and humans.
With the breakthroughs and widespread application of AIGC (AI-generated content), the environmental impact of training large AI models has gradually gained attention. However, compared to the widely discussed 'carbon footprint,' the 'water footprint' is rarely mentioned, and even fewer data disclosures exist. 'The level of attention is roughly 10 years behind that of carbon footprints.'
During the training of large AI models, data centers (servers) are heavily utilized. This includes on-site cooling water (Scope 1), off-site water used for electricity generation (Scope 2), and supply chain water consumption for AI chip manufacturing (Scope 3), all of which consume and evaporate large amounts of clean freshwater. Ren Shaolei believes that measuring and disclosing AI's water footprint is straightforward and represents the first step toward reducing water consumption.
It is difficult to quantify exactly how much of the increased water usage is attributable to AI, but at least many of the newly built data centers are dedicated to AI. Moreover, AI's share of their overall business is generally increasing. Other business segments are relatively stable, with slower growth and less rapid demand increases. However, in addition to launching products, AI now requires extensive internal testing and development.
<span style="text-decoration:underline;">The Potential of the Artificial Intelligence Industry</span>
Since the beginning of this year, general artificial intelligence, represented by cognitive large models, has sparked a global wave of enthusiasm. International companies such as OpenAI, Microsoft, and Google have continuously increased their investments, while domestically, a "thousand-model war" has erupted, with numerous high-tech enterprises competing in research and development. AIGC is a milestone for human society and artificial intelligence. In the short term, AIGC has transformed fundamental productivity tools; in the medium term, it will alter social production relations; and in the long term, it will drive a qualitative breakthrough in overall social productivity. In this transformation of productivity tools, production relations, and productivity, the value of data—as a production factor—has been greatly amplified.
AIGC elevates data as a core resource of the era, accelerating the digital transformation of society to a certain extent.
The Beijing Municipal Government issued the "Implementation Plan for Accelerating the Construction of a Globally Influential Artificial Intelligence Innovation Hub in Beijing (2023-2025)." The plan proposes leveraging Beijing's innovation resources in the AI field to continuously enhance its global influence and further promote the leading development of artificial intelligence. By 2025, the core AI industry is targeted to reach a scale of 300 billion yuan, maintaining a growth rate of over 10%, with the radiating industry scale exceeding 1 trillion yuan. Leading AI enterprises will continue to increase R&D investments, the number of startups will grow, and the total number of enterprises will remain domestically leading, with 5-10 new unicorn companies to be cultivated.
In the fiercely competitive market, the ability of enterprises and investors to make timely and effective market decisions is the key to success. The AI industry report compiled by China Research Network provides a detailed analysis of the development status, competitive landscape, and market supply-demand dynamics of China's AI industry. It examines the opportunities and challenges faced by the industry from various perspectives, including policy environment, economic conditions, social factors, and technological advancements. Additionally, it reveals potential market demands and opportunities, offering accurate market intelligence and scientific decision-making references for strategic investors to choose appropriate investment timings and for corporate leadership to formulate strategic plans. The report also holds significant reference value for government departments.