Google Cloud computing services cover the infrastructure most apps actually run on, virtual machines, containers, serverless functions, storage, and networking, billed on a pay-as-you-go basis so you are not paying for idle capacity.
Four Ways to Run Code
- Compute Engine — full-control VMs with custom machine types and optional GPUs, for lift-and-shift migrations or performance-sensitive batch jobs.
- Kubernetes Engine (GKE) — managed scaling, upgrades, and security for containerized apps, the default choice for microservices needing portability.
- Cloud Run — fully managed containers that scale to zero between requests, now Google’s recommended starting point for most new serverless projects, since it covers what App Engine’s flexible environment used to do at a lower cost with fewer constraints.
- Cloud Functions — event-driven code for a single trigger (an HTTP request, a Pub/Sub message, a storage event); as of Cloud Functions 2nd gen, it actually runs on top of Cloud Run’s infrastructure rather than a separate stack.
App Engine still exists and standard environment remains a reasonable fit for a narrow set of legacy cases (tight Datastore coupling, very low-volume apps that fit comfortably in the free tier), but for a new project in 2026, Cloud Run is the more current default; App Engine’s flexible environment specifically is one Google itself now points teams away from in favor of Cloud Run. See Google Cloud Platform services for the full service catalog beyond compute.
Storage and Databases
Cloud Storage handles unstructured data, photos, logs, backups, with multiple storage classes to control cost. For structured data, Cloud SQL runs managed MySQL, PostgreSQL, or SQL Server, while Spanner offers a globally distributed database with ACID transactions for workloads that need to scale horizontally without giving up consistency.
Security and Migration
IAM, encryption at rest and in transit, DDoS protection, and compliance certifications (HIPAA, PCI-DSS, GDPR) come built in, details in Google Cloud security services. For moving existing workloads in, Migrate for Compute Engine and Transfer Appliance automate replication and cutover, covered further in Google Cloud migration services.
AI, Analytics, and Support
Pre-trained APIs and the Gemini Enterprise Agent Platform (Google’s current name for what was Vertex AI) cover machine learning, detailed in Google Cloud AI services; BigQuery, Dataflow, and Dataproc handle large-scale analytics. If you would rather not run day-to-day operations yourself, Google Cloud managed services covers monitoring and patching, and Google Cloud consulting services is there for architecture reviews and cost optimization.