GCP Professional Cloud Architect: Exam Decoded 2026

Built against the current exam guide. Updated for 2026. Covers what the exam actually tests — including the parts most courses haven't caught up to yet.

If your GCP study material doesn't cover Gemini, Model Armor, or SLSA provenance — it's teaching you last year's exam.

Google updated the Professional Cloud Architect exam guide to reflect how cloud architecture actually works in 2026 — AI-integrated, supply-chain-aware, and compliance-first. This course is built against the current exam guide, not a snapshot from two years ago.

What older courses are missing — covered here in full

Gemini Agent Platform (Section 2.4) — Grounding, function calling, agent orchestration, multi-agent systems, Vertex AI Agent Builder. Entirely absent from pre-2024 courses. Now a dedicated exam domain.

Securing AI (Section 3.1) — Model Armor for prompt injection and jailbreak protection, Sensitive Data Protection for training data scanning, and secure Vertex AI model deployment patterns. A new exam guide bullet not covered in older material.

Software supply chain security (Section 3.1) — SLSA provenance levels, Cloud Build attestation, Binary Authorization policy enforcement, Artifact Analysis continuous CVE scanning, Assured Open Source Software. Significantly expanded in the current exam guide.

Gemini Cloud Assist (Section 1.2) — AI-assisted infrastructure design, natural language to IaC generation, inline IDE suggestions, and natural language log queries. Now an explicit exam topic.

Prebuilt AI APIs (Section 2.5) — Vision AI, Natural Language AI, Speech-to-Text, Translation API, Document AI, Healthcare Data Engine. A dedicated full section where previously this was scattered or omitted entirely.

Data sovereignty and Assured Workloads (Section 3.2) — FedRAMP, ITAR, CJIS, national jurisdiction requirements, and Cloud EKM for external key sovereignty. Substantially deeper than previous exam versions.

Updated and deepened from prior exam versions

Cloud Run and serverless architecture — Min instances, concurrency tuning, cold start mitigation, Cloud Run jobs, second-generation execution environment. Deeper than the surface treatment in older courses.

GKE Autopilot and multi-dimensional autoscaling — Autopilot vs Standard trade-offs, HPA + VPA + Cluster Autoscaler + KEDA working together, Workload Identity for GKE, GKE Backup. Exam questions now test Autopilot-specific constraints.

Cloud Deploy and progressive delivery — Release pipelines, approval gates, canary and blue/green deployments, rollback strategy. Cloud Deploy is now the exam-expected delivery pattern, not ad-hoc CI/CD scripts.

Workload Identity Federation — OIDC token exchange for GitHub Actions, AWS, and Azure workloads authenticating to GCP without SA key files. Heavily tested in current exams; absent from courses written before 2022.

Chaos engineering and reliability testing — Fault injection testing, load vs stress vs soak testing, DiRT (Disaster Recovery Testing), penetration testing governance. Full dedicated coverage in Phase 6.

Apigee and API management — Apigee vs Cloud Endpoints vs API Gateway — when each is correct, monetisation, developer portal, hybrid deployment. The exam now tests which tier fits the scenario.

For renewal candidates specifically: the second practice test (250 questions) is calibrated to the current exam guide's scenario density and question style — including all new AI and supply chain security topics. If you passed the PCA two years ago, this is what has changed since then.

What hasn't changed — and still matters

The fundamentals are still heavily tested. IAM, VPCs, Cloud SQL vs Spanner vs Bigtable, Cloud Storage tiers, multi-zone vs multi-region HA, RTO and RPO, CMEK and key management, SLOs and error budgets — this course covers all of it at the same depth as the new topics. Nothing is assumed from a previous course.