Google Earth adds Nano Banana 2 AI image generation on web

Google Earth rolled out Nano Banana 2, its newest AI image‑generation model, directly in the web version, letting users create location‑based visuals with a single click—but the feature is limited to a few daily generations and is unavailable on the mobile app, raising questions about scalability and potential abuse [1].

How the upgrade works

The Nano Banana 2 model, previously embedded in Google’s suite of products such as Chrome and Gemini Notebook, now lives inside Google Earth’s web interface. Users zoom to any location, tap a button, and the AI produces a new image—ranging from historical reconstructions (e.g., a bustling Pompeii before 79 AD) to custom infographics like annotated views of the Statue of Liberty. The tool also offers a “Refine Image” step before saving the result to a project. Google markets the feature as a way for educators, architects, and planners to visualize past scenes or future developments, though the model’s output can contain factual inaccuracies because it “draws from multiple sources” without citing them [1][2].

Limitations and rollout quirks

At launch, the feature imposes daily generation caps and each image can take up to two minutes to render. Users cannot adjust the camera angle after submitting a prompt; they must restart the process if they wish to change the view. Moreover, the AI capability is tied to Google Earth’s own subscription tiers, separate from Google One, meaning even paid Google AI subscribers receive only a handful of generations per day [1]. The most notable restriction is platform exclusivity: the Nano Banana 2 tool is only available on the web version of Google Earth, leaving the Android app—downloaded over 500 million times—without access, a gap the company has not yet addressed [1].

Market and competitive context

Google’s integration of Nano Banana 2 follows earlier deployments of the model in Chrome, Photos, Search, and Gemini Notebook, signaling a broader strategy to embed generative AI across its product ecosystem. Competitors such as Microsoft’s Copilot and Adobe’s Firefly are also embedding image‑generation tools in mapping or design platforms, but Google’s advantage lies in its massive geospatial data assets and the ability to combine AI with real‑time satellite imagery. However, the limited daily quota and lack of mobile support may cede early‑adopter advantage to rivals that offer more seamless, cross‑device experiences.

What to watch

• Full‑device rollout: Whether Google extends Nano Banana 2 to the Android and iOS apps in the coming months.
• Subscription tier changes: Adjustments to daily generation limits or integration with Google One plans.
• Regulatory scrutiny: Potential policy responses if the feature is misused for misinformation or deep‑fake creation, especially after the feature was rolled back following misuse concerns [1].

The launch underscores Google’s push to make AI a native layer in its mapping services, but the constrained rollout and platform gaps suggest the company is still testing the balance between innovation, user demand, and responsible use.