AI Product July 14, 2026

Google brings image generation to AI Overviews: text-to-image directly in Search, powered by Nano Banana

On July 14, 2026, Google (announced by Brad Kellet, Senior Engineering Director for Search) unveiled image generation from a text prompt inside AI Overviews, powered by Google's Nano Banana model. It rolls out over the coming weeks, starting in English and in regions that already support image creation in AI Mode. The announcement coincided with a redesigned Google Images homepage marking the service's 25th anniversary.

Navy threads emerging from a command bar and weaving into an empty frame, with an emerald generation core glowing at the center

Google has embedded image generation inside AI Overviews: a user types a text prompt and an original image, rendered from scratch by the Nano Banana model, appears within the search result. In the July 14, 2026 announcement, the company said the feature would open first in English and in regions that already support image creation in AI Mode, rolling out over the coming weeks. The same day, Google Images got a redesigned homepage for its 25th anniversary.

The threshold here is production moving from an app into search itself. You no longer visit a separate tool to make an image; the distance between intent and picture collapses into the search box. It points the same way as the board’s Adobe Firefly and Veo cost cards: creative production is becoming abundant and cheap. For advertisers the question is no longer “how do I produce the visual” but, amid abundance, what gets chosen and whose brand is remembered.