How to Pass the GCP Professional Cloud Architect (PCA) Exam in 2026

The Google Cloud Professional Cloud Architect (PCA) exam is 2 hours, 50-60 multiple choice and multiple select questions, and USD 200 (plus tax). Each exam includes two of four published case studies (Altostrat Media, Cymbal Retail, EHR Healthcare, KnightMotives Automotive), and case study questions make up 20-30% of the exam. Google does not publish a passing score.

Practice this exam in timed mode with the StudyKits Professional Cloud Architect app. The Google Cloud certification hub lists every Google Cloud exam guide and practice app.

Google recommends 3+ years of industry experience, including 1+ year on Google Cloud. With the Associate Cloud Engineer behind you, plan about 6-8 weeks. Facts are from the Google pages under Sources, checked October 15, 2026.

What is the PCA exam format?

The Google Cloud Professional Cloud Architect exam tests your ability to design and plan cloud solution architecture, manage and provision cloud infrastructure, design for security and compliance, analyze and optimize technical and business processes, manage implementations, and ensure solution and operations excellence. The exam guide gives approximate weights of ~25%, ~17.5%, ~17.5%, ~15%, ~12.5% and ~12.5% for these six sections.

The exam has 50-60 questions and you get 2 hours. Google does not publish a passing score. The exam costs USD 200 (plus tax) and can be taken at a testing center or remotely (online proctored). The certification is valid for 2 years, and a shorter renewal exam is available.

Google recommends 3+ years of industry experience including 1+ years designing and managing solutions on Google Cloud. However, candidates with strong fundamentals and dedicated preparation can pass with less experience, especially if they first earn the Associate Cloud Engineer certification.

Should you take ACE before PCA?

The most effective path to the Professional Cloud Architect is through the Associate Cloud Engineer (ACE) first. The ACE teaches you the foundational services and hands-on skills. The PCA builds on that foundation with architectural design decisions, case studies, and enterprise-level considerations.

If you already hold the ACE, you have a head start. In our experience that cuts PCA study from about 10-12 weeks to 6-8.

If you are coming from AWS, check our AWS vs GCP vs Azure Certifications comparison to understand how the PCA maps to AWS certifications. The PCA is roughly equivalent to the AWS Solutions Architect Professional (SAP-C02) in terms of difficulty and scope.

How do the PCA case studies work?

The PCA exam includes case studies, and this is what sets it apart from most other cloud certifications. Google publishes four case studies before the exam; two of them appear in each exam and account for roughly 20-30% of the questions.

Each case study describes a fictional company with:

• Company overview — what the company does, its size, and its industry
• Solution concept — what they want to achieve with Google Cloud
• Existing technical environment — current infrastructure, databases, applications
• Business requirements — compliance, availability, budget constraints
• Technical requirements — performance, scalability, integration requirements

The current case studies (checked October 2026 on the Google Cloud exam page) are:

• Altostrat Media
• Cymbal Retail
• EHR Healthcare
• KnightMotives Automotive

Older study material often uses retired case studies (Helicopter Racing League, Mountkirk Games, TerramEarth, Dress4Win). Download the current case study files from the exam page and read those.

Study these case studies thoroughly before the exam. Read them multiple times. For each case study, map the requirements to specific Google Cloud services. When you see a case study question on the exam, you should already have a mental architecture diagram ready.

Which Google Cloud services must you know?

Compute

Google Kubernetes Engine (GKE) is the most heavily tested compute service on the PCA exam. You must understand:

• GKE Standard vs GKE Autopilot — Autopilot manages node infrastructure for you, Standard gives you full control
• Node pools, node auto-provisioning, and cluster autoscaling
• Workload Identity for pod-level IAM
• Multi-cluster management with GKE Enterprise (formerly Anthos)
• Network policies, Pod Security Standards, and Binary Authorization
• GKE release channels: Rapid, Regular, Stable

Compute Engine:

• Machine types, custom machine types, and preemptible/spot VMs
• Managed instance groups (MIGs) with autoscaling and autohealing
• Sole-tenant nodes for licensing and compliance
• Live migration and maintenance policies

Cloud Run and Cloud Functions:

• Cloud Run for containerized serverless workloads
• Cloud Functions for event-driven functions
• When to choose Cloud Run vs Cloud Functions vs GKE vs Compute Engine

Anthos (GKE Enterprise):

• Running workloads across GCP, on-premises, and other clouds
• Anthos Config Management for GitOps
• Anthos Service Mesh for microservices observability
• Migration strategies using Migrate to Containers

Data and Analytics

BigQuery is to GCP what S3 is to AWS — it appears everywhere:

• Serverless data warehouse architecture
• Partitioned tables (time-based, range, integer) and clustered tables
• BigQuery ML for machine learning without exporting data
• BigQuery BI Engine for sub-second query performance
• Federated queries to Cloud SQL, Cloud Storage, and Bigtable
• Data transfer service and scheduled queries
• Slot reservations vs on-demand pricing

Cloud Spanner:

• Globally distributed relational database with strong consistency
• When to choose Spanner vs Cloud SQL vs AlloyDB
• Schema design: interleaved tables, primary key selection for even distribution
• Multi-region configurations for 99.999% availability

Cloud SQL and AlloyDB:

• Cloud SQL for MySQL, PostgreSQL, and SQL Server
• AlloyDB for PostgreSQL-compatible workloads needing high performance
• High availability, read replicas, and automated backups

Bigtable:

• Wide-column NoSQL database for low-latency, high-throughput workloads
• Row key design for even data distribution
• When to use Bigtable vs Firestore vs Memorystore

Dataflow and Dataproc:

• Dataflow for Apache Beam pipelines (both batch and streaming)
• Dataproc for Apache Spark and Hadoop workloads
• When to use Dataflow vs Dataproc

Networking

• VPC design — shared VPC, VPC peering, Private Google Access, Private Service Connect
• Cloud Load Balancing — global vs regional, HTTP(S) vs TCP/UDP, internal vs external
• Cloud CDN — integration with load balancers, cache modes, signed URLs
• Cloud Interconnect and Cloud VPN — Dedicated vs Partner Interconnect, HA VPN
• Cloud DNS — public vs private zones, DNS policies, DNS peering
• Cloud NAT — outbound internet access for private instances
• Cloud Armor — DDoS protection, WAF rules, adaptive protection

Security and Identity

• IAM — roles (basic, predefined, custom), service accounts, Workload Identity Federation
• Organization policies — constraints at the org, folder, or project level
• VPC Service Controls — security perimeters around GCP resources to prevent data exfiltration
• Cloud KMS — encryption keys, key rotation, CMEK vs CSEK
• Secret Manager — storing and accessing secrets
• Identity-Aware Proxy (IAP) — zero-trust access to applications
• Security Command Center — threat detection, vulnerability scanning, and compliance monitoring

Operations

• Cloud Monitoring — metrics, dashboards, alerting policies, uptime checks, SLIs/SLOs
• Cloud Logging — log routing, log sinks to BigQuery/Cloud Storage/Pub/Sub, log-based metrics
• Cloud Trace and Cloud Profiler — distributed tracing and application performance profiling
• Error Reporting — automatic error grouping and notification

Which design principles does the exam test?

The PCA exam tests your ability to apply Google’s recommended design principles:

Design for High Availability

• Deploy across multiple zones by default, multiple regions for critical workloads
• Use managed services (GKE, Cloud SQL HA, Spanner) for built-in redundancy
• Implement health checks and autoscaling
• Design for graceful degradation

Design for Scalability

• Use autoscaling at every layer: GKE cluster autoscaler, MIG autoscaling, BigQuery on-demand
• Choose serverless options (Cloud Run, Cloud Functions, BigQuery) when possible
• Design data stores for expected growth: Bigtable for high-throughput, Spanner for global scale

Design for Security

• Apply least privilege IAM at every level
• Use VPC Service Controls for sensitive data
• Encrypt data at rest and in transit (Google encrypts by default, but know when to use CMEK)
• Use Private Google Access and Private Service Connect to avoid public internet exposure

Design for Cost Optimization

• Use committed use discounts for predictable workloads
• Use preemptible/spot VMs for fault-tolerant workloads
• Choose the right storage class: Standard, Nearline, Coldline, Archive
• Right-size instances using recommendations from Recommender

Design for Operations

• Implement Infrastructure as Code with Terraform or Deployment Manager
• Use Cloud Build for CI/CD pipelines
• Set up proper monitoring, logging, and alerting from day one
• Define SLIs, SLOs, and error budgets

What does an 8-week PCA study plan look like?

Weeks 1-2: Core Services Deep Dive

• Master GKE: cluster types, networking, security, autoscaling
• Learn Compute Engine MIGs, load balancing, and network design
• Study Cloud Run and Cloud Functions use cases
• Hands-on: deploy a multi-tier application on GKE with an HTTP(S) load balancer
• 20 practice questions per day

Weeks 3-4: Data Services and Analytics

• Deep dive into BigQuery: architecture, optimization, security
• Study Cloud Spanner, Cloud SQL, AlloyDB, Bigtable, and Firestore
• Learn Dataflow and Dataproc for data processing
• Hands-on: build a data pipeline from Cloud Storage to BigQuery with Dataflow
• 25 practice questions per day

Weeks 5-6: Security, Networking, and Operations

• Master IAM, organization policies, and VPC Service Controls
• Study networking: VPC design, load balancing, interconnect options
• Learn Cloud Monitoring, Logging, and SRE principles
• Study Anthos/GKE Enterprise for hybrid and multi-cloud scenarios
• Hands-on: set up a shared VPC with VPC Service Controls
• 30 practice questions per day

Weeks 7-8: Case Studies and Practice Exams

• Study all published case studies. Map each requirement to GCP services.
• Take full-length practice exams under timed conditions
• Review weak areas and re-study relevant documentation
• Focus on architectural decision-making, not just service knowledge
• 40 practice questions per day
• Schedule your exam

How should you answer case study questions?

For each case study question:

1. Recall the company’s key requirements (you should have these memorized)
2. Identify which requirement the question is testing
3. Eliminate answers that violate stated requirements (for example, if the company requires no public IP addresses, any answer exposing resources publicly is wrong)
4. Choose the answer that best balances technical requirements and business constraints

Do not try to memorize specific answers to case study questions. The questions change between exam versions. Instead, understand each company deeply enough to make architectural decisions for any scenario they present.

What should you do on exam day?

• Read case study questions carefully. The question will reference a specific company, and the correct answer depends on that company’s requirements.
• For non-case-study questions, look for the “Google Cloud way” of solving problems. Google favors managed services, serverless options, and BigQuery for analytics.
• When two answers seem correct, consider operational overhead. Google almost always prefers the solution with less management burden.
• Use the 2-hour time limit wisely. Case study questions take longer — allocate extra time for them.
• Flag questions you are unsure about and revisit them at the end.

What comes after PCA?

The Professional Cloud Architect opens doors to other professional-level certifications. Consider the Professional Data Engineer if you work with data, the Professional Cloud Security Engineer if you focus on security, or the Professional Cloud Developer if you build applications.

How can you practise for PCA?

Read the four case studies in the exam guide before anything else, then follow the study plan. The StudyKits PCA app has original practice questions with explanations. Its questions follow the previous exam outline; update in progress.

Frequently asked questions

What are the current PCA case studies?

Altostrat Media, Cymbal Retail, EHR Healthcare and KnightMotives Automotive. Two appear in each exam, and you can view them on a split screen during the exam.

What is the PCA passing score?

Google does not publish one.

How do I renew the Professional Cloud Architect?

It is valid for 2 years. Renew with the standard exam; a 1-hour, 25-question renewal exam (USD 100) built around one generative AI case study; or designated Google Skills courses, which add 1 year.

Is PCA harder than ACE?

Yes. PCA is a professional exam with case studies and a 3+ year experience recommendation. ACE is associate level.

Can I take PCA in Japanese?

Yes. The exam is offered in English and Japanese.

Sources

Checked October 15, 2026:

• Google Cloud Professional Cloud Architect
• PCA exam guide (PDF)
• PCA renewal exam guide (PDF)
• Google Cloud certification renewal