Google Rolls Out May Core Update Amid AI Search Overhaul

Google began rolling out its May 2026 core update on May 21, marking the second such update this year and the fourth confirmed ranking update for 2026 [1, 2]. This update coincides with significant announcements from Google I/O, including a redesigned Search box, an upgrade to AI Mode with Gemini 3.5 Flash, and a preview of new AI-powered information agents [1]. The rollout is expected to take up to two weeks, potentially overlapping with other changes and making it challenging to isolate the causes of any ranking shifts [1, 2].

Key takeaways

• Google's May 2026 core update began rolling out on May 21 and is expected to last up to two weeks [1, 2].
• This update is the second core update and fourth confirmed ranking update of 2026 [1, 2].
• The update coincides with Google I/O announcements, including a redesigned Search box and AI Mode powered by Gemini 3.5 Flash [1].
• Google released its first AI Mode usage data, reporting over 1 billion monthly users and a tripling of query length compared to traditional searches [1].
• Conflicting guidance exists between Google's Search and Lighthouse teams regarding the llms.txt file for AI agents [1].

Core Update and AI Search Evolution

The May 2026 core update commenced without a companion blog post or specific stated goals from Google, similar to the March core update [1, 2]. Google's Search Central account announced the rollout on X, confirming its progress [2]. SEO professionals, such as Marie Haynes, have connected the timing of the update to Google I/O, suggesting it aligns with Gemini 3.5 Flash powering AI features in Search [1]. Some experts, like Harpreet Singh Chatha, speculate the update could target websites over-optimizing for AI citations [1].

Alongside the core update, Google announced what it described as the biggest upgrade to the Search box in over 25 years [1]. This redesigned Search box expands dynamically, supports multimodal inputs like images and files, and offers AI-powered suggestions [1]. Gemini 3.5 Flash is now the default model in AI Mode, which Google says has over 1 billion monthly users, with queries doubling each quarter [1]. These AI Mode searches are reportedly three times longer than traditional queries, and follow-up queries have increased by 40% monthly in the U.S. [1]. Jeffrey Cohen of Skai noted that shoppers are using AI for detailed research, shifting from short keywords to conversational briefs [1].

Conflicting Guidance and Future Implications

Google's guidance on the llms.txt file, intended for AI agents, has presented conflicting signals from its Search and Lighthouse product teams [1]. While Google's AI guide states llms.txt isn't necessary for AI Search, Lighthouse 13.3 checks for the file by default and flags sites with errors [1]. John Mueller clarified that markdown pages are useful for documentation but not for most websites, advising sites to prioritize current needs over future agentic browsing readiness [1]. This discrepancy highlights an unresolved conflict within Google's own teams regarding AI agent interaction [1].