This analysis is part of Google: The Gravitational Center of the AI Market, a deep dive by The Business Engineer.

Competitors offer APIs to developers — but the developer must bring their own distribution. Google offers APIs embedded in surfaces that already reach billions — and then expands developers into the full stack.
The Developer Playing Field
OpenAI/ChatGPT: GPT-4.5, o3 API, ~300M MAUs — API only, no platform surface. GitHub Copilot: Azure AI + VS Code — strong IDE integration, depends on OpenAI models. Anthropic/Claude: Safety-focused dev tools — growing adoption, API only. Cursor, Replit, Hugging Face: Niche/open-source tools.
Google: API Is the Entry Point, Surfaces Are the Moat
Gemini API (10B+ tokens/min) is the entry point, surrounded by: Gemini App (750M MAUs), Google Workspace (3B+ users), Android (580M+ Circle to Search), Chrome (3B+ users), YouTube ($60B+ content), Search (2B+ users), Google Cloud + TPU (cost backbone), and Infra + Data + Platform (1.8x cross-product expansion).
The Land-and-Expand Flywheel
1. LAND via API: Developer enters through Gemini API. 2. EXPAND into Surfaces: Workspace, Android, Search, Chrome (3B+ users). 3. 1.8x Cross-Product: Infra + Data + Platform upsell — AI customer becomes full cloud customer. 4. RETAIN: Vertical integration (TPU + Gemini + Vertex = sticky).









