Data Center
A physical facility used for the centralized housing of computer systems and associated equipment. Data centers serve as core infrastructure for the digital economy and cloud computing, consuming large amounts of electrical power and requiring continuous cooling.
Core Facts
A typical large data center can consume millions of liters of water per day, equivalent to the daily water consumption of a small town.
With the proliferation of large language models and generative AI, power consumption in data centers is rising rapidly, making energy use a key bottleneck in the expansion of the AI industry.
A 100-megawatt-class data center may consume millions of gallons of water per day, equivalent to the daily water usage of tens of thousands of households.
Timeline (last 90 days)
The AI industry chain involves multiple sectors such as chips, cloud computing, data centers, and energy, and an expectation adjustment in any one link could trigger a chain reaction.
AI systems can continuously monitor the operating state of a data center, automatically adjust variables such as CPU frequency, cooling parameters, and task distribution, and find a better balance between performance and energy consumption.
Resource allocation in traditional data centers follows a static logic, reserving redundant capacity to handle peak loads, resulting in low resource utilization and energy waste.
Most existing AI governance frameworks assume that models run inside monitorable data centers, and once AI can leave these environments, traditional auditing and shutdown measures may fail.
A recent study proposed that certain AI models may possess the ability to self-replicate without the support of traditional data centers.
AI training clusters have far higher requirements for power supply, cooling capacity, and network bandwidth than traditional data centers, and are highly physically concentrated, so once a key node is damaged, the entire training task may be interrupted and difficult to recover quickly.
The successful drone strikes on Russian data centers exposed the vulnerability of centralized infrastructure in modern conflict: when critical computing resources are concentrated in a few physical locations, a single precision strike can cause widespread service outages.
The industry widely believes that inference will become the main body of data center computing power spending.
Traditional data centers primarily rely on grid power with diesel generators as backup power.
The return cycle for data center assets is as long as 10–15 years.
40 more timeline events
All Facts (20)
A typical large data center can consume millions of liters of water per day, equivalent to the daily water consumption of a small town.
85%VerifiedWith the proliferation of large language models and generative AI, power consumption in data centers is rising rapidly, making energy use a key bottleneck in the expansion of the AI industry.
75%VerifiedA 100-megawatt-class data center may consume millions of gallons of water per day, equivalent to the daily water usage of tens of thousands of households.
75%VerifiedNorthern Virginia, home to the world's largest data center cluster, has seen its local utility Dominion Energy file multiple requests to raise electricity rates to cover grid expansion costs.
75%Verified超大规模数据中心服务器数量通常超过10万台，单个园区用电量可达数百兆瓦
75%VerifiedThe training and inference of large AI models place enormous demands on computing power, and the electricity consumption and carbon emissions of data centers have become topics of industry discussion
65%VerifiedQueue times for grid access by data centers are growing increasingly long, and transmission capacity in many regions is approaching saturation.
65%VerifiedA large supercomputing data center can consume hundreds of megawatts of electricity annually, equivalent to the total electricity consumption of hundreds of thousands of households.
65%VerifiedThe long-term local jobs created by data centers are relatively limited and disproportionate to the resources they consume.
65%VerifiedTech giants building or expanding data centers to support AI operations have become a significant source of new load on the power grid.
65%VerifiedThe annual electricity consumption of a GW-scale data center campus can reach billions of kilowatt-hours, equivalent to the total electricity consumption of approximately 1 million households.
65%VerifiedThe inference process of large language models consumes as much computational resources as the training process.
65%VerifiedData centers are highly automated with limited local employment contribution; a large data center may require only dozens of technical staff to maintain operations.
65%VerifiedA hyperscale data center with an investment of billions of dollars may require only 50–200 people for daily operations.
65%Verified国际能源署（IEA）2024年报告预测全球数据中心用电量将在2026年前翻倍，其中AI工作负载是最主要的增量来源
90%UnverifiedWith the explosive growth in demand for AI training and inference, global data center construction is entering an unprecedented peak period.
70%UnverifiedThe demand for GPU computing power to train and run inference on large models is growing explosively, making data centers and their supporting power supply the scarcest resources in the industry
70%UnverifiedAs data centers continue to expand in scale, power supply and cooling capacity are becoming hard constraints limiting the expansion of AI services.
70%Unverified全球数据中心用电量以每年约10%的速度增长
70%UnverifiedData center construction in arid regions or water-stressed watersheds directly competes with agricultural and municipal water use.
60%