Q1 2026 earnings call: Remarks from our CEO

Source: Google Date: 2026-04-29 URL: https://blog.google/company-news/inside-google/message-ceo/alphabet-earnings-q1-2026/

Summary

Alphabet’s Q1 2026 results showed Google Cloud exceeding $20B revenue for the first time (63% YoY growth) with a backlog nearly doubling to over $460B. Generative AI products on Cloud grew nearly 800% YoY, and Google’s first-party models are processing 16 billion tokens per minute via direct API. Gemma 4 hit 50 million downloads within weeks of release, with the broader open model portfolio surpassing 500 million total downloads. The company introduced eighth-generation TPUs optimized for both training and inference.

Implications

• The 16B tokens/minute API figure is a concrete scale benchmark; as a comparison point it calibrates just how much inference volume a hyperscaler-class deployment absorbs at this stage of adoption.
• Gemma’s download velocity (50M in weeks) signals that Google’s open-weight strategy is working as a developer acquisition funnel — putting Google-trained model weights into agentic pipelines that may later graduate to Cloud APIs.
• The backlog doubling to $460B locks in enterprise cloud-AI spend at a horizon that reduces competitive risk from model commoditization — infrastructure relationships outlast any single model generation.
• Eighth-gen TPUs covering both training and inference workloads (previously separate) reflects growing pressure to reduce cost-per-token on long-running agentic tasks, not just one-shot inference.