NVIDIA-Certified Associate

NVIDIA AI Power and Cooling Certification Exam

(NCA-APC)

What Is the NVIDIA AI Power and Cooling Certification (NCA-APC)?

NCA-APC is an entry-level certification that validates a candidate’s foundational understanding of the power, cooling, and mechanical systems supporting AI data centers and AI factories, with special focus on safe and effective cross-trade coordination.

Please carefully review our certification FAQs and exam policies before scheduling your exam.

To access the exam, you’ll need to create a Certiverse account.

If you have any questions, please contact us.

Exam Details for NCA-APC

Duration: 1 hour 

Price: $125 

Certification level: Associate    

Subject: AI Power and Cooling    

Number of questions: 50

Scoring: Pass/Fail 

Language: English

Validity: This certification is valid for two years from issuance. Recertification may be achieved by retaking the exam.

Credentials: Upon passing the exam, participants will receive a digital badge and optional certificate indicating the certification level and topic.

Who Is AI Power and Cooling Certification For?

This certification is designed for entry-level and experienced tradespeople, apprentices, helpers, installers, technicians, facilities staff, contractors, commissioning support, project personnel, and anyone needing introductory physical infrastructure awareness. This includes professionals working in: 

  • Electrical systems and power distribution
  • Mechanical systems, HVAC, plumbing, and pipefitting
  • Controls and monitoring
  • Installation and commissioning support
  • Facility operations and maintenance
  • Construction and project coordination

It is suited to those who need to recognize major infrastructure systems, understand how power and cooling support AI equipment, and communicate effectively with the teams responsible for those systems.

What Are the Recommended Qualifications for This Certification?

  • Experience: No prior AI data center experience is required. General technical or worksite familiarity is helpful, along with an understanding of major power and cooling components and basic electrical and heat flow paths.
  • Technical readiness and monitoring: Read simple diagrams, labels, and status indicators; recognize installation and commissioning readiness; and compare readings and alarms with supplied operating limits.
  • Safety, communication, and work boundaries: Follow site procedures, communicate observed conditions, and escalate concerns to authorized personnel.

    The certification validates foundational awareness and does not replace trade qualifications, role-specific training, or site authorization.

Certification Learning Path

Here is the recommended training to prepare for this certification exam.

Share

Electrical and Liquid Cooling Installation

This self-paced course teaches you how to safely install, connect, and start up electrical and liquid-cooling hardware. It focuses heavily on hazard recognition and knowing when to escalate issues.

MGX Physical Installation

This self-paced course gives you step-by-step guidance on how to physically assemble, rackmount, and integrate modular NVIDIA MGX™ systems.

NCA-APC Exam Preparation Topics and Recommended Reading

The information below is intended to help you prepare for this exam.
Topics covered, weight per topic, recommended training, and additional readings are provided.

AI Factory Fundamentals

Build a plain-language mental model of an AI factory, its supporting facility systems, the people who coordinate them, and the limits of an Associate credential.

AI Factory Fundamentals Potential Exam Topics:

  • Recognize the basic purpose of an AI factory and the role of supporting facility infrastructure.
  • Identify the major facility systems that support AI equipment: power, cooling, mechanical, controls, monitoring, and heat rejection.
  • Explain in simple terms why concentrated AI equipment increases the importance of capacity, cooling, coordination, monitoring, and availability.
  • Identify common trade and project roles and match a basic issue or question to the appropriate owner.
  • Recognize the limits of the credential and the difference between awareness, trade competence, work authorization, and design authority.

14%
Exam Weight

What You Need to Know

NVIDIA defines an AI factory as specialized computing infrastructure that creates value from data across ingestion, training, fine-tuning, and high-volume inference. At facility scale, dense AI infrastructure depends on coordinated electrical supply and distribution, air and liquid cooling, mechanical plant, controls, monitoring, networking, and heat rejection.

Think of the AI factory as one production system with two essential facility paths:

  • Power path: sources and backup systems move electrical energy through protection and distribution to the racks.
  • Heat path: air and/or liquid collect heat from equipment, move it through facility loops, and reject it outdoors.

Controls, sensors, dashboards, alarms, labels, records, and communications make both paths visible. A loss or constraint in one system can appear as a symptom in another, so trades must coordinate even when ownership remains separate.

Roles and Boundaries 

  • Electrical/facilities power: source, switchgear, generation, UPS/BESS, distribution, protection, capacity, and switching authority.
  • Mechanical/cooling: airflow, piping, pumps, valves, CDUs, heat exchangers, chillers, heat rejection, water quality, and leak response.
  • Controls/BMS/EPMS/DCIM/IT operations: points, status, alarms, trends, logs, monitoring paths, and approved control actions.
  • Commissioning: verification planning, checks, tests, issues, retests, records, and handover evidence.
  • EHS/security/project leadership/design authority/OEM: their exact responsibilities are site-specific. Match a question to the role named by the current responsibility matrix rather than assuming a universal owner.

Preparation Activities

  • Draw one box for compute and one for each supporting system; connect dependencies in both directions.
  • Explain the difference between awareness, qualification, authorization, and approval authority without using job titles as shortcuts.
  • Given an unfamiliar issue, identify what evidence you would capture before asking the authorized owner.

Suggested Reading

Power Infrastructure Awareness

Recognize the major parts and basic purpose of the electrical path serving AI equipment. The assessment is about reading supplied information and escalating safely—not calculating or selecting a design.

Power Infrastructure Awareness Potential Exam Topics:

  • Identify major power-path components in a simple diagram or scenario, such as source, switchgear, backup generation, UPS/BESS, distribution, PDU or busway, and rack connection.
  • Explain the basic purpose of backup power, redundancy, separation, and monitoring without calculating or selecting a design.
  • Recognize how dense equipment can affect available capacity, routing, space, access, support, and coordination with cooling or mechanical work.
  • Read basic supplied electrical information such as a label, simplified one-line, status indication, capacity statement, or stated operating limit.
  • Recognize a basic electrical warning, alarm, abnormal status, damaged condition, or scope boundary and choose the correct safe communication or escalation step.

18%
Exam Weight

What You Need to Know

NVIDIA’s Dynamic Power Software documentation describes a one-line diagram as a simplified representation of a three-phase power system. It identifies sources such as utility, generators, and UPS systems; distribution equipment such as transformers, switchgear, panels, and PDUs; protection such as breakers, fuses, and disconnects; and loads such as servers, networking equipment, and storage.

Use this conceptual sequence:
Utility or onsite source → transformer/substation → switchgear or bus → generator/UPS/BESS functions → floor or room distribution/busway → rack PDU or remote panel → rack power shelf/server PSU → compute load.

Trace from source to load, then from load back to source. At each step, identify the equipment name, its upstream parent, downstream children, protection, status, stated limit, and path label.

Backup, Redundancy, and Monitoring

  • Generator: standby source for longer-duration facility support after startup and transfer.
  • UPS: short-duration ride-through and conditioned power for designated loads while upstream supply changes or recovers.
  • BESS: stored-energy system whose function can include rapid buffering, grid/load support, and disturbance ride-through; exact design and operating role are site-specific.
  • Redundancy: preserves the required equipment or power paths during a fault or maintenance event. “Two cords” is not proof of end-to-end independence.
  • Monitoring: provides observed load, provisioned capacity, active target, utilization, health, alarms, trends, and communication paths. Sample alert rules must be adapted to the site.

NVIDIA’s DGX H100 design guidance shows that the number and arrangement of energized power supplies matter to workload continuity; it also recommends monitored rPDUs and validation of rack-level redundancy. NVIDIA rack-scale systems also illustrate the final path from remote power panel/whips to power shelves and a rack bus bar.

Preparation Activities

  • Take a simple one-line and identify sources, distribution, protection, loads, and redundant paths.
  • Explain why high rack density affects electrical capacity, cable/busway routing, floor space, service clearances, structural support, and cooling coordination.
  • Compare a stated limit with a supplied reading and practice saying only whether it is normal, near a limit, or requires escalation.
  • Read supplied electrical information
    • Start with the title block, revision, legend, and stated operating condition.
    • Follow lines and arrows rather than guessing from physical proximity.
    • Match equipment, PDU, breaker, feed, and path identifiers exactly.

Suggested Reading

Cooling & Mechanical Infrastructure Awareness

Recognize how heat is collected, moved, and rejected using air- and liquid-cooling systems. Study relationships and supplied limits; do not memorize one product’s values as universal requirements.

Cooling & Mechanical Infrastructure Awareness Potential Exam Topics:

  • Follow the basic path of heat from AI equipment to the room, liquid loop, facility system, and heat-rejection equipment using a simple diagram.
  • Distinguish air cooling, direct liquid cooling, and facility heat rejection at a conceptual level.
  • Identify the basic purpose of common mechanical and cooling components such as fans, pumps, piping, valves, manifolds, cooling distribution units, heat exchangers, chillers, and cooling towers when shown in context.
  • Recognize why temperature, flow, pressure, cleanliness or water quality, leak detection, condensation control, and drainage can matter without calculating system values.
  • Read basic supplied cooling information such as a label, simplified flow diagram, status indication, temperature or pressure limit, leak alarm, or maintenance notice.
  • Recognize a basic cooling or mechanical warning, leak, abnormal status, damaged condition, or scope boundary and choose the correct safe communication or escalation step.

17%
Exam Weight

What You Need to Know

NVIDIA rack-scale systems combine direct liquid cooling for CPU/GPU cold plates with fan-driven air cooling for remaining components.[9] NVIDIA’s DSX facilities reference traces heat from cold plates through a technical cooling system and liquid-to-liquid CDUs to facility water and outdoor heat rejection; it also shows separate air-cooling service for remaining loads.

  • Air path: component → server fans → room supply/exhaust air → CRAH/CRAC or other air-side cooling → facility plant → outdoor heat rejection.
  • Direct-liquid path: GPU/CPU → cold plate → rack manifold → technical cooling loop → CDU/heat exchanger → facility-water loop → chiller, dry cooler, or cooling tower → outdoors.
  • Hybrid system: liquid removes heat from the highest-density components while air continues cooling networking, storage, power, and other components.

Conditions That Matter

  • Temperature: relates to equipment limits and heat-transfer performance; excessively cold surfaces can create condensation risk when below dew point.
  • Flow and pressure: indicate whether coolant can reach loads and overcome system resistance. A low reading may have many causes; do not diagnose beyond supplied evidence.
  • Water/coolant quality and cleanliness: affect corrosion, scale, fouling, biological growth, filters, and small cold-plate passages. Facility and technical loops may have different requirements.[14]
  • Leak detection: can exist at tray, rack, and facility/CDU levels. NVIDIA documents tray sensors, rack ropes/spot sensors, and BMS-integrated facility detection.[12]
  • Drainage and containment: are site-specific design and environmental controls. Follow the current plan; do not assume a visible drain is approved for coolant.

Reading Cooling Information

Confirm diagram revision, flow direction, loop name, supply/return, asset or valve tag, units, normal range, alarm state, and whether the indicated condition is measured, commanded, or inferred. Compare only values that use the same units and refer to the same location and operating state.

Leak or abnormal condition: Do not silence, reset, isolate, tighten, close, open, drain, or clean unless your role and procedure authorize it. Keep personnel and electrical hazards in mind, preserve evidence, identify the exact location and alarm, and escalate through the approved path.

Preparation Activities

  • Draw separate facility-water and technical-cooling loops with a heat exchanger between them.
  • Explain the difference among air cooling, direct liquid cooling, and outdoor heat rejection.
  • Practice reading a supplied temperature, pressure, flow, or leak status without making an unsupported diagnosis.

Recommended Training

  • Electrical and Liquid Cooling Installation self-paced course (coming soon)
  • MGX Physical Installation self-paced course (coming soon)

Suggested Reading

Installation, Integration & Commissioning Awareness

Understand why trades coordinate routes, interfaces, readiness, testing, evidence, deficiencies, and handover before systems enter service.

Installation, Integration & Commissioning Awareness Potential Exam Topics:

  • Identify common cross-trade coordination points such as routes, supports, clearances, access, containment, drainage, labeling, cleanliness, sensors, and sequence of work.
  • Recognize basic readiness prerequisites before equipment is energized, filled, started, tested, or handed over.
  • Explain the beginner-level purpose of visual inspection, prefunctional checks, functional tests, and integrated testing.
  • Identify what objective evidence may show that work is complete or incomplete, including checklists, readings, labels, photos, test records, deficiencies, and approvals.
  • Recognize the purpose of point naming, sensors, alarms, controls visibility, and communication paths before operational handover.
  • Choose the appropriate response when a prerequisite, document, test result, label, responsible owner, or approval is missing.

21%
Exam Weight

What You Need to Know

Coordinate Before You Test

High-density rack placement depends jointly on power and cooling capacity, cable routing, adjacent equipment, service clearance, airflow, and expansion space. Mechanical distribution may also share space with power bus, cable cartridges, manifolds, supports, sensors, and access zones.

Coordination means proving that each trade’s work is compatible with the full system and the planned sequence. A route that is individually valid can still block maintenance access, violate clearance, interfere with containment, hide a label, obstruct airflow, complicate drainage, overload a support, or prevent a later test.

Readiness and Test Levels

ASHRAE’s AI Data Center Energy Performance Framework describes a progression from factory acceptance and delivery/installation verification through pre-functional checks, functional performance testing, and integrated-systems testing under operational and failure scenarios.

Objective Evidence and Handover

A claim that work is “done” is weaker than an approved record showing what was checked, by whom, against which requirement, when, with what result.

  • Missing or inconsistent labels, unresolved deficiencies, incomplete checklists, unapproved scripts, out-of-range readings, failed tests, or absent responsible owners are evidence that readiness is not established.
  • Commissioning monitoring should be available early enough to collect test evidence; retained trends can establish an operating baseline for later comparison.
  • Point names, sensor identity, alarm routing, controls visibility, and communication paths must be understandable before handover so operators can distinguish asset, location, condition, and status.

Missing prerequisite = hold the affected step, document the gap, protect the current safe state, notify the responsible owner, and proceed only after the required evidence or approval is supplied. Do not invent a substitute or accept an undocumented verbal assumption.

Preparation Activities

  • Classify sample evidence as inspection, prefunctional, functional, integrated, deficiency, correction, retest, or approval.
  • Given a missing label or test result, state what must be held and what information the owner needs.
  • Trace one control point from sensor through naming, controller/BMS visibility, alarm, historian, notification, and operator response.

Suggested Reading

Operations & Maintenance Awareness

Use supplied status, ranges, trends, procedures, and records to support routine awareness, controlled response, resource efficiency, and maintenance—not complex diagnosis.

Operations & Maintenance Awareness Potential Exam Topics:

  • Identify the basic purpose of dashboards, alarms, trends, logs, rounds, and equipment status indicators.
  • Compare a supplied reading or status with a stated normal range, limit, baseline, or procedure and recognize whether escalation is needed.
  • Explain why maintenance, isolation, shutdown, restart, and configuration change activities require planning, authorization, stakeholder communication, and rollback preparedness.
  • Describe the role of SOPs and MOPs in ensuring safe, reliable, and controlled execution.
  • Describe the basic interdependencies between power, cooling, mechanical, controls, and compute infrastructure, and recognize common cross-system symptoms during operational events without performing detailed troubleshooting.
  • Identify the information needed for a clear incident report, work record, shift handoff, or escalation.
  • Describe how day-to-day operational decisions impact energy and water efficiency, and recognize opportunities to improve resource utilization while maintaining required operating conditions.
  • Describe the role of preventive maintenance in sustaining data center operations by recognizing equipment conditions, trends, and maintenance requirements that may affect reliability or performance.

15%
Exam Weight

What You Need to Know

Evidence Before Interpretation

  • Dashboard: summarizes selected current and historical information.
  • Alarm: calls attention to a defined event or threshold; it does not prove root cause.
  • Trend/historian: shows change over time and supports comparison with baselines and events.
  • Log: records events, actions, states, messages, or access; timestamps and asset identity matter.
  • Round: scheduled observation that adds local condition evidence.
  • Status indicator: communicates a current state as defined by the equipment or control system.

NVIDIA Mission Control uses metrics, triggers, alerts, health checks, dashboards, and centralized logs for observability. NVIDIA DPS dashboards distinguish capacity, provisioned load, observed power, targets, utilization, and health signals.

Compare, Then Escalate

Use the range, limit, baseline, or procedure supplied in the scenario. Confirm units, asset, measurement location, timestamp, and operating state. Describe the difference factually. Escalate when the source requires it, the value is outside limits, the trend is materially abnormal, status is inconsistent, evidence is missing, or the response exceeds your authority.

Controlled Work and Change

ASHRAE recommends formal SOPs and MOPs aligned with control logic and alerts while retaining human accountability for interpretation, authorization, safety, compliance, and maintenance execution. In common usage, an SOP describes an approved recurring operating method, while a MOP provides a controlled plan for a specific change or maintenance activity. Site definitions and approval requirements control.

Before maintenance, isolation, shutdown, restart, or configuration change, verify scope, current state, hazards, dependencies, approvals, communication plan, execution steps, hold points, verification, rollback/recovery, and recordkeeping. NIST’s configuration controls reinforce documented approval, impact analysis, controlled implementation, validation, record retention, and baseline updates.

Cross-System Events

A compute alarm may reflect power or thermal conditions; a cooling event can change electrical load; a power transfer may affect pumps, controls, communications, and IT equipment; a sensor or network issue can make a healthy system appear unavailable. Correlate timestamps and affected assets, but do not label correlation as confirmed cause.

Incident and Handoff Information

Use a concise structure: affected system and exact asset/location; observed condition; time and current status; stated limit/baseline; immediate safety or service concern; actions already taken; approvals/notifications; evidence or logs attached; owner; next action and next checkpoint. Preserve uncertainty honestly.

Efficiency and Maintenance

Energy and water efficiency must be managed together with equipment limits, reliability, climate, heat-rejection method, and operating requirements. ASHRAE identifies airflow management, liquid-cooling architectures, intelligent controls, real-time monitoring, continuous commissioning, heat reuse, and low/no-water cooling as relevant strategies.[17] Preventive maintenance uses schedules and recurring tasks; predictive/condition-based approaches use trends and condition evidence. Current OEM and site requirements govern equipment-specific intervals.

Governing rule: Observe → verify evidence → compare with the approved reference → communicate → act only within authorization → record the result. Never tune, reset, bypass, switch, or isolate merely because a dashboard suggests an opportunity.

Preparation Activities

  • Build a short event timeline from a dashboard, alarm, trend, round, and log; separate observation from inference.
  • Practice a shift handoff that identifies current state, outstanding risk, owner, and next checkpoint.
  • For an efficiency opportunity, name the operating condition that must remain protected before any change is considered.

Suggested Reading

Safety, Work Boundaries & Team Communication

Demonstrate professional conduct and situational awareness on a complex multidisciplinary site. Local law, the authority having jurisdiction, employer rules, and site procedures always control.

Safety, Work Boundaries & Team Communications Potential Exam Topics

  • Explain the importance of adhering to site-specific safety requirements, permits, PPE, isolation practices, and approved procedures.
  • Identify conditions that require a person to stop, protect the area, avoid making an adjustment, and contact an authorized role.
  • Use clear basic terminology to communicate the affected system, location, observed condition, time, status, and immediate concern.
  • Explain the importance of current documents, revision status, labels, change control, and accurate records.
  • Handle facility, customer, system, and assessment information according to site and program confidentiality and security requirements.

15%
Exam Weight

What You Need to Know

Site Rules and Hazardous Energy

OSHA’s U.S. general-industry hazardous-energy rule requires an energy-control program with procedures, training, and periodic inspections; covered equipment must be isolated and stored/residual energy controlled, and an authorized employee verifies isolation before work. OSHA also requires appropriate electrical PPE and safeguards such as insulated tools, shields, barriers, signs, barricades, or attendants where applicable.

Study synthesis. A permit does not replace qualification, hazard assessment, deenergizing where required, isolation verification, PPE, or authorization. Stay within the assigned task, physical boundary, permit, role, and approved procedure.

Stop, Protect, and Escalate

Pause the affected work and contact the authorized role when conditions differ from the plan; a guard, label, barrier, or isolation is missing; a permit or procedure is unclear, expired, or mismatched; required PPE is unavailable; a leak, odor, sound, damage, alarm, exposed energy, or other new hazard appears; required evidence or approval is absent; or the task exceeds your authorization.

Keep clear, warn affected people, preserve the safe state, and avoid resets or adjustments. Resume only after the responsible authority resolves the issue and re-briefs or re-authorizes the work.

Clear Communication

State the system, exact location/asset, observed condition, time, current status, and immediate concern. Distinguish direct observation from interpretation. Repeat back critical limits and instructions, identify the person in charge, record the handoff, and re-brief after a meaningful change.

Current Documents and Change Control

Before acting, verify asset ID, drawing/procedure title, revision, prerequisites, ticket/permit status, and approval. If the field condition and document disagree, stop and route the discrepancy through change control. NIST SP 800-53 describes baseline and change-control practices that include documented scope, approval decisions, controlled implementation, validation, monitoring, and retained records.

Confidentiality and Security

  • Use only approved accounts, devices, removable media, storage locations, and communications channels.
  • Access only the areas, racks, consoles, records, and data needed for the assigned task.
  • Protect labels, drawings, photos, logs, screens, customer names, network details, credentials, and assessment content.
  • Do not tailgate, lend badges or credentials, bypass access controls, or share exam content.
  • Report suspected loss, disclosure, tampering, or unauthorized access immediately.

Jurisdiction note: The OSHA sources in this guide are U.S. general-industry references. Construction work and non-U.S. work may be governed by different or additional requirements. The exam should supply any jurisdiction-specific condition needed to answer.

Preparation Activities

  • Practice converting an unsafe or unclear observation into a concise escalation message.
  • Compare two document revisions and identify why title, revision, status, and approval matter.
  • Classify information by need-to-know and identify the approved handling channel.

Suggested Reading

Practice Questions

Explore sample questions to get familiar with the topics and question styles you may encounter on the certification exam.

Learning Mode

Answer each question and get immediate feedback, including the correct answer, an explanation, and a link to learn more.

Ready to Get Certified?

Validate your expertise. Whether you are installing, managing, or supporting AI power and cooling infrastructure, earning this certification showcases your skills and helps you stand out in a rapidly growing field.

Contact Us

NVIDIA offers training and certification for professionals looking to enhance their skills and knowledge in the field of AI, accelerated computing, data science, advanced networking, graphics, simulation, and more.

Contact us to learn how we can help you achieve your goals.

Stay Up to Date

Get training news, announcements, and more from NVIDIA, including the latest information on new self-paced courses, instructor-led workshops, free training, discounts, and more. You can unsubscribe at any time.