In the era of intelligent computing, data centers, as critical infrastructure, are experiencing ever-increasing energy consumption. How to measure and optimize their energy efficiency has become a global focus. Against this backdrop, Power Usage Effectiveness (PUE), a core metric, has emerged and quickly become one of the most widely used indicators in the industry. It successfully simplifies complex energy consumption data into an intuitive ratio, enabling operators, investors, and regulators worldwide to evaluate, benchmark, and promote data center energy conservation under a unified standard. Whether pursuing commercial benefits, fulfilling social responsibilities, or coping with increasingly stringent "dual carbon" regulations, the continuous attention to and optimization of PUE has become a key pivot driving data centers and the entire digital industry toward a more efficient and sustainable future.
Based on this, MIR 睿工业 has conducted a study on the key PUE policies of eight countries—the United States, China, Germany, Malaysia, Thailand, Saudi Arabia, Indonesia, and Vietnam—to provide insights into global data center PUE policies.
01. Basis for Country Selection and Current PUE Status of the Eight Countries
This study primarily selects countries with top-ranking data center capacities globally for analysis:
According to the "Global Digital Economy Development Research Report (2024)" by CAICT, in terms of IT load by region, the United States continues to maintain the first position in global data center capacity, accounting for 51%, Europe accounts for 17%, and China accounts for 16%;
The IT load of data centers in Germany is expected to reach 1,690 MW by 2025, accounting for a relatively high proportion in Europe;
Saudi Arabia accounts for 86% of the data center capacity in the Middle East, which is expected to reach 300 MW by 2025;
The scale of data centers in Thailand is expected to reach 1 GW, with the scale at 350 MW in 2024;
The IT load capacity of data centers in Vietnam is approximately 169.8 MW, lower than that of Indonesia (268 MW) or Malaysia (827 MW).
By comparing the average PUE of the aforementioned eight countries in 2024 with the global benchmark, MIR 睿工业 found significant divergence in PUE performance among the countries.
Average PUE of Various Countries and Global Benchmark PUE in 2024
Source: Global benchmark PUE is from the Uptime Institute 2024 report, analyzed and compiled by MIR 睿工业
Top Performers (PUE < 1.56): United States (1.40), Germany (1.43), China (1.46). The average PUE of these countries is already below the global benchmark, and their overall energy efficiency levels are at the forefront globally.
Performers Needing Improvement (PUE > 1.56): Malaysia (1.58), Vietnam (1.78), Indonesia (1.7), Saudi Arabia (1.82). The average PUE of these countries is higher than the global benchmark, leaving significant room for improvement in energy efficiency levels.
02. Detailed Analysis of PUE Policies in the Eight Countries
Overview of Data Center PUE Policies in the Eight Countries
Source: Collected and compiled by MIR 睿工业
2.1 United States: Abolishing Hard Targets and Shifting to a Triad of "Strong Monitoring + Technology + R&D"
At the federal level, the United States has not set a national PUE target. Its PUE policy has undergone a shift from setting targets to abolishing them. The current PUE regulatory framework takes the form of mandatory monitoring, technological guidance, and R&D-driven initiatives working in tandem.
Historical Evolution:
2016: The US Office of Management and Budget (OMB) released the "Data Center Optimization Initiative (DCOI)," setting targets for data centers: new builds with PUE < 1.4, and retrofits with PUE < 1.5.
2018: A federal executive order revoked the aforementioned unified PUE targets. The Federal CIO considered that PUE is significantly affected by factors such as climate, geographical location, measurement time, and redundancy design, making it unsuitable as a mandatory comparative and assessment indicator across agencies and regions.
Current Regulatory Framework:
Mandatory Monitoring: The "Data Center Optimization Initiative (Memorandum-19-19)" DCOI (M-19-19) issued by the US Office of Management and Budget (OMB) requires agencies to install metering systems in large data centers (>100kW) and continuously report PUE data to the OMB, without setting an overall PUE target.
Technological Guidance: Agencies such as the Federal Energy Management Program (FEMP) have released the "Best Practices Guide for Energy-Efficient Data Center Design," providing specific technical pathways for energy efficiency improvement (such as adopting natural cooling and optimizing incoming air temperature) as voluntary guidelines for the industry.
R&D-Driven: NREL (National Renewable Energy Laboratory) and ARPA-E (Advanced Research Projects Agency-Energy) are revolutionizing cooling technologies at their root by funding cutting-edge projects such as COOLERCHIPS, aiming to reduce data center power and cooling energy consumption in the long term.
2.2 China: Multi-Level Regulation and Strong Enforcement
China's data center PUE regulation has formed a complete system characterized by clear hierarchy, explicit targets, and distinct rewards and penalties. Its core logic is: the national level sets PUE targets with clear timelines, while local governments formulate stricter standards on this basis and ensure the implementation of PUE targets through innovative economic and administrative means.
National Top-Level Design Sets Targets: Represented by the "2024 Special Action Plan," it is explicitly stipulated that the national overall PUE must be reduced to below 1.5 by 2025, with targets of ≤1.25 and ≤1.2 set for newly built large-scale and national hub node projects, respectively.
Local Execution Combines Incentives and Constraints: Major provinces and cities have further tightened the national standards (e.g., Beijing requires the PUE of ultra-large data centers to be ≤1.15), and strongly drives PUE compliance by combining "punitive" measures (such as differential electricity pricing in Beijing and Anhui, project approval restrictions in Jiangsu and Anhui, and mandatory elimination in Hebei) with "incentive" measures (such as construction and operation subsidies in Sichuan, energy consumption quota support in Shenzhen, and priority approval in Guangdong).
National and Regional PUE Policy Targets and Reward/Penalty Measures in China
Data Source: Official government websites, collected and compiled by MIR 睿工业
2.3 Germany: Legislative Hard Constraints, Dual Mandates of PUE≤1.2 + 100% Green Electricity
By early 2026, Germany has established a policy system for data center energy efficiency regulation centered on PUE (Power Usage Effectiveness) as the core metric, combining legal mandates with industry self-discipline. We analyze that the broader background for establishing this policy system is to respond to the EU's "Green Deal" and carbon neutrality goals.
Legal Core of the Policy System Establishment - The Energy Efficiency Act
Revised in 2023 and implemented in phases starting from 2024, the Energy Efficiency Act is the core legal basis for Germany's data center PUE regulation. Its key provisions include:
Section 18a: Explicitly includes "large data centers" (annual power consumption ≥ 1 GWh) in the regulatory scope of key energy-consuming units;
Mandatory Disclosure Obligation: Starting from 2025, all large data centers must submit annual energy efficiency reports to the Federal Network Agency (BNetzA), which must include data such as PUE value, IT load, cooling methods, and the proportion of renewable energy used;
Energy Efficiency Improvement Obligation:
Data centers put into operation before July 1, 2026: PUE must be ≤ 1.3;
Data centers newly built or significantly retrofitted after January 1, 2027: PUE must be ≤ 1.2, and 100% powered by renewable energy (which can be achieved through PPAs, green certificates, or self-built photovoltaics, etc.).
2.4 Malaysia: Tiered Guidance and Incentive First
According to the "Malaysia Green Data Center Construction Guidelines" released at the end of 2024, tiered regulations are made based on data center type, construction year, power supply, and capacity, reflecting the refinement and targeted nature of the policies. The flexible clause followed by all regulations, "verification required after baseline determination," indicates that the designed PUE targets mainly serve as design guidelines rather than immediate mandatory thresholds. The core reason behind this is that regulators are still completing the baseline摸底 (摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底 -> baseline摸底...... (Note: The word "摸底" was repeated in my thought process due to a glitch, I will translate it correctly as "baseline survey").
The core reason behind this is that regulators are still completing the baseline survey of actual energy consumption levels of data centers nationwide. Therefore, the initial implementation phase of the policy will focus on data collection and baseline establishment. However, this regulation has significant dynamics. Once the baseline is established, the target values in the guidelines will have the foundation to be transformed into specific compliance assessment criteria. At that time, based on more accurate data, regulators may calibrate the target values and shift to strict enforcement.
In addition, Malaysia has explicitly proposed the development goal of "green data centers" in its national digital economy plan, requiring the PUE of newly built large data centers to gradually move towards below 1.2.
Malaysia Data Center Classification and Corresponding PUE Design Targets
Data Source: "Malaysia Green Data Center Construction Guidelines"
In terms of incentive policies, the Malaysian Investment Development Authority (MIDA) provides multiple incentives such as tax exemptions, operational subsidies, and land preferences for newly built data center projects that adopt highly efficient cooling technologies (such as liquid cooling, indirect evaporative cooling, etc.). For example, hyperscale data centers that adopt liquid cooling technology and have a PUE below 1.2 can enjoy corporate income tax exemptions for up to 10 years; for technological R&D investment, a certain percentage of R&D expenses can be deducted from the tax amount.
2.5 Thailand: Tax-Based Tiering and Incentive First
Thailand currently has not set a national mandatory standard for data center PUE, and its regulatory model is "incentive-oriented".
The Board of Investment of Thailand (BOI) revised the investment promotion policy for the data center industry in May 2025. The core feature of this policy is: using PUE values as the main scale to distinguish investment incentive levels, providing greater tax incentives for high-efficiency data centers, thereby guiding energy efficiency improvement rather than imposing mandatory constraints through legislation. Specifically, the policy divides data centers into two categories:
High-Efficiency (Energy-Saving) Data Centers (PUE ≤ 1.3): Eligible for corporate income tax exemptions for up to 8 years. They must simultaneously meet the following conditions: ① IT load ≥2MW, ② at least 4 high-speed communication lines, ③ submission of a highly efficient water management plan, ④ provision of data hosting services with advanced computing capabilities such as GPUs, ⑤ submission of a project plan beneficial to Thailand (such as training and R&D cooperation).
Ordinary Data Centers (PUE > 1.3 or not meeting the standards): Corporate income tax exemptions for up to 5 years, with other conditions being the same as high-efficiency ones.
2.6 Saudi Arabia: Tier Certification for Neutrality, Energy Plan as a Hard Access Threshold
In August 2023, the Communications, Space and Technology Commission (CST) of Saudi Arabia released the "Data Center Services Regulations" (hereinafter referred to as the "Regulations"). These regulations aim to improve the service level of data centers in Saudi Arabia and provide regulatory transparency and a fair competitive environment for cloud service providers (customers) who wish to host and operate cloud services in data centers within the country.
These regulations define concepts such as data center services and impose requirements on service providers regarding operator neutrality, energy management, and sustainability plans.
In the regulations, data center services are defined as: hosting services provided by data center service providers to customers, including space, power, and cooling support for hosting servers, network components, storage devices, etc.
Regarding registration, the regulations require service providers to register at all development stages of the data center. All data centers holding CST-recognized Tier II construction certification (Standard) and CST-recognized Tier III construction certification (Advanced) should "maintain operator neutrality," which means allowing any network connection service provider to provide network connection services to customers inside and outside the data center without discrimination, and not being limited to the data center facilities of a single connection service provider.
Data center service providers of both Standard and Advanced categories should provide energy management and sustainability plans to reduce energy consumption, carbon emissions, and electronic waste. Service providers should proactively fulfill social and environmental responsibilities by practicing responsibility through reducing carbon emission impacts, improving energy efficiency, and reducing electronic waste. Service providers must keep CST updated with accurate information of their data centers at all times, including space, power capacity (used and available), other connection service providers, certification lists, etc. If the relevant service provider fails to sufficiently demonstrate that it meets the minimum technical qualification requirements stipulated in these regulations (such as minimum certification standards, energy plans, proof of experience, quality standards, financial resources, and other conditions), CST has the right to refuse any registration application at its discretion.
2.7 Indonesia: Macro Energy-Saving Targets + Government-Promoted Construction
Indonesia currently has not set specific data center PUE control targets. The core feature of its policy is national top-level planning setting macro energy-saving targets + government investment promoting digital infrastructure construction.
The Ministry of Energy and Mineral Resources (ESDM) of Indonesia has released the National Energy General Plan (RUEN), which stipulates cross-sectoral energy policy application and implementation plans aimed at achieving the goals of the National Energy Policy (KEN). The quantified target requires a 1% reduction in energy intensity annually from 2015 to 2025; and sets the final total energy consumption target: a 17% reduction by 2025 and a 39% reduction by 2050.
The Indonesia Investment Authority (INA) and Singaporean capital Granite Asia have cooperated to jointly invest over 1.2 billion USD in the technology and AI ecosystem, which includes support for specific projects such as the Batam DayOne Data Center Park. This shows that the government is actively promoting digital infrastructure construction, laying an industrial foundation for implementing stricter energy efficiency standards in the future.
Currently, the Indonesian government is drafting a new decree, which is expected to impose higher and more specific requirements on the energy efficiency and sustainability of data centers.
2.8 Vietnam: Drawing a "1.4" Hard Line First, Regulations and Standards Still Pending
Vietnam's regulatory framework for data center PUE is taking shape and becoming increasingly specific. Its characteristics are manifested as: establishing phased national energy efficiency targets and setting clear PUE numerical upper limits for newly built projects.
In September 2020, the Ministry of Industry and Trade of Vietnam led the formulation of the "Guidelines for Energy Efficiency Management of Data Centers," initially setting the direction for national energy efficiency improvement: the PUE target is 1.6 by 2025, and the target is reduced to 1.45 by 2030.
On October 29, 2025, the Vietnamese government further set an upper limit of 1.4 for the PUE performance of newly built data centers.
However, Vietnam's overall regulatory framework is still not perfect. Currently, there are no specific regulations targeting the Power Usage Effectiveness (PUE) of data centers, nor are there official national standards for green data centers. In practice, enterprises mainly rely on internationally recognized Tier levels or PUE standards for design and certification.
03. Comparative Analysis of PUE Policies in the Eight Countries
Source: Analyzed and compiled by MIR 睿工业
Summary of PUE Target Characteristics:
Countries that have set PUE targets include Germany, China, Malaysia, and Vietnam. Their policies have the following three characteristics:
Refinement of PUE Guidance Policies: They basically adopt the management idea of tiering and classification. For example, China classifies by region, existing stock, and increment, while Malaysia makes tiered regulations based on data center type, construction year, power supply, and capacity, reflecting the refinement and targeted nature of the policies.
Gradually Clear PUE Compliance Requirements and Timelines: Most regions (Germany, China, Malaysia, Vietnam) have set binding PUE compliance timelines.
PUE Policy Requirements and Timelines of the Four Countries
Source: Analyzed and compiled by MIR 睿工业
Quantitative Indicators Become the Consensus of Most Countries: Among the eight countries/regions in the survey scope, most countries have already taken PUE as the core quantitative indicator to measure data center energy efficiency, such as Malaysia, China, Vietnam, and Germany.
Summary of PUE Policy Commonalities:
There are significant differences in PUE-related policies among the eight surveyed countries, but the commonalities can be summarized into the following five policy routes:
1) Mandatory Legislation Route: Germany is the only country among the analyzed ones that has established a mandatory, executable PUE target system through legislation. The legislative lead is the Federal Ministry for Economic Affairs and Climate Action, the deliberation and approval are by the Bundestag and the Bundesrat, and the official release carrier is the Federal Law Gazette (published by the Federal Ministry of Justice), which has legal effect.
2) Technological Guidance Route: Represented by the United States, it emphasizes data transparency and technological innovation. There are no mandatory PUE targets at the federal level, but there are mandatory PUE data reporting requirements for large data centers (>100kW). It guides the industry to enhance technological innovation capabilities through technological practice guidelines and cutting-edge R&D, focusing on long-term industrial transformation.
3) Incentive Guidance Route: Malaysia, China, and Thailand utilize economic tools such as tax incentives (mainly corporate income tax), subsidies (such as energy consumption fees), and differential electricity pricing to implement rewards and penalties.
4) Regulatory Framework Route: The representative is Saudi Arabia. Saudi Arabia has not set specific PUE numerical requirements; the core lies in the regulatory framework for energy management and sustainability plans. In August 2023, the Communications, Space and Technology Commission (CST) of Saudi Arabia released the "Data Center Services Regulations" (hereinafter referred to as the "Regulations"). It aims to provide regulatory transparency and a fair competitive environment for cloud service providers (customers) who host and operate cloud services in data centers within the country.
5) Guidance First, Legislation Later Route: Indonesia and Vietnam both prioritize the release of national-level strategic plans or guidelines, setting macro targets (Indonesia's energy intensity target, Vietnam's PUE target) to guide the direction of industrial development, and then follow up with the formulation of specific laws, standards, or mandatory access regulations (Vietnam has set an upper limit for newly built projects, and Indonesia is drafting a new decree).
Conclusion
As global attention to data center energy consumption continues, MIR 睿工业 expects that more countries will improve their PUE regulatory systems. Policy development trends may include: transitioning from voluntary guidelines to mandatory standards, expanding from a single PUE indicator to a comprehensive energy efficiency evaluation system, and combining environmental indicators such as the proportion of renewable energy used with PUE requirements.
Overall, the status of PUE as the core indicator of data center energy efficiency has been established. The continuous improvement of policies in various countries will drive the global data center industry toward a more efficient and sustainable direction.
*Log in to the MIR DATABANK Manufacturing Professional Data Platform to get more manufacturing data information, company information, and research reports! https://www.mirdatabank.com