Google says AI Mode is expanding globally faster than search features

Google’s AI Mode is moving across countries and languages on a timetable that would have looked impossible in older Search. Liz Reid said the product’s multilingual model architecture let Google expand far faster than previous search features, which often took months or even years to reach every market.

That speed changes the agency playbook. If AI Mode can reach multiple languages in a matter of months, multilingual SEO stops looking like a slow, market-by-market rollout and starts looking like a competitive land grab. For international clients, the pressure shifts to localization workflows, entity consistency, and content architecture that can hold up across regions at once.

Google has tied that expansion to its existing Search systems, not to a separate AI-only layer. The company said it uses its current ranking systems to ground AI Mode responses based on location, which matters because local signals still appear to shape what users see even as the interface gets more generative. For agencies, that means the old work still counts: structured entity naming, local business data, regional relevance, and trust signals are now part of AI visibility, not just classic blue-link SEO.

The rollout has accelerated quickly. On Oct. 7, 2025, Google said AI Mode was launching in more than 35 new languages and over 40 new countries and territories, pushing availability to more than 200 countries and territories total. In May 2025, Google positioned AI Mode as its most powerful AI search, combining Gemini 2.0 with Google’s information systems, including the Knowledge Graph and real-time sources. By May 2026, Google said AI Mode was available globally and that Gemini 3.5 Flash was the default model in AI Mode for everyone globally.

Scale has followed the rollout. Google said AI Mode had surpassed a billion monthly active users globally, and AI Mode queries had more than doubled every quarter since launch. Even so, Google’s Search Help pages still describe AI Mode as an opt-in experiment in Search Labs, and they warn that AI responses may include mistakes. That combination is exactly why agencies cannot treat multilingual AI search as a finished product.

The practical move now is to build for fast international repetition. Local pages need the same entity signals, the same factual backbone, and the same naming discipline across languages before AI search visibility shifts again. If Google keeps compressing the rollout cycle, the agencies that get multilingual architecture right first will be the ones clients notice in every market.