AI Prefabricated Modular Data Center Design Questionnaire

The explosive growth of AI computing power is fundamentally reshaping data center construction logic. Traditional on-site construction methods involve long cycles, high risks, and struggle to rapidly respond to the ultra-high power density demands of GPU clusters. AI Prefabricated Modular Data Centers (AI Prefab Modular Data Centers) integrate complete data center functions into transportable modules through standardized factory prefabrication and rapid on-site assembly. This approach significantly shortens delivery timelines while more precisely matching the requirements of high-density heat dissipation solutions such as liquid cooling.

To ensure the design solution aligns with the customer’s real operational goals, site constraints, and technical boundaries from the outset, we need to systematically collect key input information. The accuracy of the questionnaire directly determines module dimensions, power architecture, cooling system selection, redundancy levels, and final energy efficiency performance. Incomplete or significantly inaccurate information may result in modules that cannot be installed upon arrival, insufficient cooling capacity, or major rework at later stages. The following questionnaire is organized strictly according to the actual project design process, with detailed explanations provided for each major section and sub-item to help both parties achieve efficient alignment.

1. General Project Information

General project information forms the starting point and boundary for all subsequent design work. It clarifies the project’s official identity, responsible parties, geographic constraints, and schedule, directly influencing production planning, logistics, and on-site delivery. Only when these foundational details are clear can technical design proceed with defined priorities and constraints.

• What is the official project name? The official project name serves as the unique identifier for all design, production, logistics, and acceptance documents. It is used not only for internal document management but also directly linked to contract terms, customs declarations, insurance, and final delivery certificates. An unclear name or frequent later changes can lead to version confusion, difficulties in responsibility tracing, and even compliance issues.
• What is the official name of the end user? The end user is the entity that will actually operate and maintain the modules. Their management processes, security policies, and monitoring permission requirements often differ from those of the contracting customer. Clearly identifying the end user enables us to reserve the correct account systems, remote access interfaces, and operational permission boundaries during the design phase, avoiding secondary modifications after delivery due to mismatched permissions.
• What is the specific geographic location of the project? The precise geographic location determines the local climate zone, grid standards, seismic fortification intensity, road transport height and weight restrictions, as well as local building and fire codes that the modules must satisfy. Incorrect or vague location information can directly result in conflicts between structural strength, protection ratings, or interface standards and on-site regulations, preventing successful acceptance upon arrival.
• What is the target date for the project to officially enter operation? The target ready-for-service date is the core anchor for reverse-scheduling the entire project. It determines factory production sequencing, procurement of long-lead equipment, FAT testing windows, shipping schedules, and allocation of on-site installation resources. An unclear or frequently changing date causes production and logistics plans to lose their baseline, increasing rush costs and quality risks.

2. Site Conditions & Infrastructure Status

Site conditions are the prerequisite for whether prefabricated modules can be successfully deployed and operate stably over the long term. Since modules complete most integration in the factory, their dependence on on-site space, power, water, and ambient temperature is actually higher. Any missing or misjudged critical conditions may force major solution adjustments or even redesign of module dimensions.

• What are the specific details of currently available site space? The type, dimensions, clear height, and load-bearing capacity of available space directly determine module layout, lifting paths, temporary storage areas, and whether multi-layer stacking or secondary on-site structures are required. Insufficient space information often forces reduction of module sizes, inadequate aisle widths, or inability of lifting equipment to access the site, severely impacting installation efficiency and safety.
• What is the readiness level of the site’s power infrastructure? Existing or obtainable power capacity, number of incoming circuits, and whether generator or energy storage interfaces are already available determine the starting point of the module’s internal power distribution architecture. Insufficient power readiness or delayed information forces the addition of temporary generation solutions, adjustment of transformer capacity, or redesign of incoming switchgear, leading to higher costs and schedule delays.
• What are the site’s water source conditions and supply capacity? The abundance, quality, supply pressure, and drainage conditions of the water source directly determine whether the liquid cooling system adopts open evaporative, closed-loop, or pure dry cooling. Incorrect water information can cause cooling system selection failure, resulting in insufficient water volume, severe scaling, or inability to meet WUE targets, affecting long-term operational stability.
• What are the typical ambient temperature and annual extreme values at the site? Annual average temperature and yearly extreme high/low temperatures form the fundamental boundary conditions for cooling load calculations and outdoor cooling source selection. Deviations in temperature data can lead to undersized or oversized cooling equipment capacity—mildly increasing annual PUE, or severely causing server overheating protection or system shutdown during extreme weather.

3. Power, Redundancy, & Architecture

The power system is the core lifeline of an AI prefabricated modular data center. High-density GPU loads place extremely high demands on power quality, switching speed, and fault tolerance. Any deviation in power architecture can directly reduce overall availability or limit future expansion.

• What are the input voltage, phase, and frequency of the main power supply? Input voltage, phase, and frequency form the starting point for selecting all electrical equipment inside the modules. Grid standards vary significantly across countries and regions. Confirmation errors in specifications will require complete redesign and reprocurement of transformers, circuit breakers, and busbars, directly disrupting production schedules and generating substantial sunk costs.
• What is the target redundancy level or corresponding Tier standard? Redundancy level defines the tolerance capacity and switching logic of the power and cooling systems under fault conditions. It determines UPS configuration, busbar segmentation, number of cooling source backups, and final availability metrics. Setting the level too high significantly increases investment; setting it too low fails to meet business continuity requirements. Alignment with business objectives must be strictly established at the early design stage.
• Are relevant project technical documents already available? Existing tender specifications, site master plans, single-line diagrams, and similar documents are critical materials for reducing redundant surveys and quickly identifying interface conflicts. Without these documents, the design team can only start from assumptions, making it easy to discover incompatibilities with existing systems later, leading to interface modifications or even module returns to the factory.

4. High-Density IT Load & Cooling Requirements

High-density IT load and cooling represent the core differentiators of AI prefabricated modular data centers compared with traditional solutions. Current mainstream GPU server rack power densities commonly reach 80–132 kW or higher. Traditional air cooling can no longer cope, making liquid cooling the standard configuration. These parameters directly determine cooling piping, pump sets, heat exchanger capacity, and internal thermal management layout of the modules.

• What is the planned total IT load capacity? Total IT load is the capacity ceiling of the entire project. It directly determines the number of modules, power module specifications, total cooling capacity of the system, and scale of supporting infrastructure. Large deviations in load estimation either waste initial investment or cause the system to quickly hit bottlenecks as business grows, losing the flexibility of modular expansion.
• What are the specific models and quantities of servers/GPUs planned for deployment? Different GPU models have significantly different power curves, thermal design power, recommended cooling methods, and liquid cooling interface standards. Accurate model and quantity information enables precise matching of cooling medium, flow requirements, and quick-connect specifications, avoiding compatibility issues that prevent normal server installation or fail to achieve adequate cooling performance.
• What is the design power density per rack? Single-rack power density is the core input for fluid dynamics calculations, pipe diameter selection, pump head design, and hot-aisle layout. Higher density imposes stricter requirements on cooling uniformity, local hotspot control, leak detection sensitivity, and emergency drainage capability. Any underestimation will result in overheating risks under full load.
• What are the design inlet and outlet liquid temperatures for the server liquid cooling system? Design inlet and outlet liquid temperatures determine the temperature difference utilization efficiency and free cooling potential of the cooling system. Temperature settings must simultaneously satisfy the server manufacturer’s allowable range and the optimal efficiency range on the cooling source side. Improper settings either prevent full utilization of free cooling or cause internal condensation or insufficient heat dissipation in the servers.

5. Efficiency, Compliance, & Additional Needs

Energy efficiency targets and compliance requirements determine the long-term sustainability and deployability of the solution. PUE and WUE are key metrics for measuring infrastructure efficiency, while certifications and special requirements relate to whether the product can successfully pass reviews and meet the customer’s long-term operational strategy.

• What is the target PUE value set for the project? The PUE target directly drives the selection of cooling architecture, power path efficiency, and degree of free cooling utilization. Setting the target too low significantly increases investment in precision equipment; setting it too high fails to demonstrate the energy efficiency advantages of liquid cooling and modular design. It must be comprehensively balanced against actual business load characteristics and local climate conditions.
• What is the target WUE value set for the project? The WUE target determines whether the cooling system is allowed to use evaporative cooling or must adopt a fully closed-loop or dry cooling solution. In water-scarce regions, WUE is a hard constraint. An unclear target can force the original cooling design to be overturned later due to water conservation requirements, resulting in major changes.
• Which mandatory certifications or regulatory requirements must be met? Certification and regulatory requirements determine component selection, structural strength, fire compartments, testing standards, and factory acceptance procedures. Omitting critical certifications (such as UL, CE, local seismic or fire codes) can lead to rejection by regulatory authorities at the delivery stage, preventing operation or requiring costly rectification.
• Are there any other special constraints or integration requirements? Special requirements include energy storage integration, specified brand preferences, interfacing with existing DCIM/BMS systems, appearance customization, noise limitations, and other non-standard needs. These requirements are often raised only in later project stages yet have major impacts on internal module layout and interfaces. They must be fully collected and feasibility-assessed at the early design stage.

6. Additional Engineering Guidelines

When completing the questionnaire, please ensure the accuracy or reasonableness of the rack power density and server liquid cooling inlet/outlet temperatures. These two items are core inputs for high-density liquid cooling system design and directly determine fluid calculations, pipe diameters, pump selection, and precision thermal management strategies. Significant deviations may result in insufficient cooling capacity or over-designed systems.

If certain parameters have not yet been finalized, please provide a reasonable estimated range (for example, rack density of 80–120 kW, inlet temperature of 35–45°C). The engineering team will use this information to reserve appropriate safety margins and future expansion headroom, reducing the risk of major design revisions later due to parameter changes.

A complete and accurate questionnaire is the critical first step for an AI prefabricated modular data center to move from concept to a deliverable solution. It enables the engineering and architecture teams to produce a design that truly matches requirements in the shortest time, effectively reduces change risks, and maximizes the project’s long-term operational efficiency and return on investment.

If you need a directly fillable Word or Excel version of the questionnaire, or would like to arrange a one-on-one technical discussion for a specific project, please feel free to contact us at any time. We look forward to working with you to build high-density, high-efficiency prefabricated modular infrastructure for the AI era.

Please discuss with us for more information if you  need any further help.