Google Cloud Platform (GCP) is the cloud computing division of Alphabet Inc., launched in 2008 from Mountain View, California. It holds approximately 12% of the global cloud infrastructure market and is particularly strong in data analytics (BigQuery), machine learning (Vertex AI), Kubernetes (GKE — which Google originally developed), and global networking. GCP is the cloud of choice for many data-intensive organizations and AI-first startups.
Pricing is fully pay-as-you-go with two notable structural advantages over competitors: automatic sustained use discounts and per-second billing. Sustained use discounts apply automatically when a VM runs for more than 25% of a billing month, adding no administrative overhead — discounts scale up to 30% for VMs running the full month with no commitment required. Committed use discounts (CUDs) for predictable workloads offer up to 57% off on 1-year or 70% off on 3-year commitments.
New accounts receive $300 in free credits valid for 90 days to evaluate any GCP service. An always-free tier includes one e2-micro VM instance per month in select US regions (0.25 vCPU, 1GB RAM), 5GB Cloud Storage standard, 1TB BigQuery queries per month, 2 million Cloud Functions invocations per month, and Cloud Run with 2 million requests per month. The e2-micro free VM is a genuine always-on resource for lightweight backends; on-demand pricing for an e2-standard-2 starts at approximately $48.92/month.
GCP’s AI and data stack is a significant competitive differentiator. BigQuery’s serverless SQL analytics model bills by query data processed, making it extremely cost-effective for variable analytics workloads. Vertex AI consolidates model training, deployment, and MLOps in one managed platform. Google Gemini models power AI features across the stack with Gemini API pricing on a per-token basis.
The main weaknesses are a narrower overall service catalog than AWS, a smaller third-party integration ecosystem, and billing complexity around AI and data services that surprises new users. The Gemini Developer API costs are excluded from the $300 free credit and GPU usage on the trial account is not covered, which frustrates developers testing AI workloads.