On May 19, 2026, Google's annual conference opens at the Shoreline Amphitheatre in Mountain View, California. The main keynote is broadcast from 7 p.m. (French time) on YouTube, and the event is shaping up to be the most artificial-intelligence-focused in the company's entire history. At the heart of the program: Gemini 4, Google's new flagship model, and a series of announcements that could redefine our everyday relationship with digital tools.

Gemini 4: Google's most ambitious AI

Gemini 4 is unquestionably the star of this edition. The new model promises capabilities far beyond its predecessor on every front: more precise reasoning, better understanding of complex instructions, handling of long conversations, and native multimodality that unifies text, image, audio, video and code in a single request.

But what is really getting people talking is the feature called Personal Intelligence: a connection between Gemini and all of the user's personal data — Gmail, Google Photos, Drive, Calendar, YouTube — to create an assistant that truly knows you. The ambition is to allow Gemini to understand the context of your digital life so it can give you relevant answers without you having to explain everything again each time.

Google plans to roll out this feature to 2 billion users in more than 200 countries — an ambition that raises as much hope as it does questions about privacy.

Remy, the personal AI agent that acts on your behalf

One of the most anticipated announcements is that of an autonomous AI agent, internally nicknamed Remy (a reference to Pixar's famous cooking rat). The idea is simple but radical: an assistant capable of really acting on your behalf, without you having to monitor every step.

In practical terms, Remy could:

• Reply to your emails according to your style and habits
• Schedule appointments while taking your constraints into account
• Carry out research and provide you with a concise summary
• Interact with third-party applications autonomously

This agent marks an important step in the transition toward agentic AI: we are no longer talking about a tool that answers questions, but about an assistant that takes initiatives and executes tasks from end to end.

Android 17 and XR glasses: hardware joins the dance

Google I/O 2026 is not limited to software. Android 17 will be presented in detail during the event, with a stable version expected in June. The operating system has been deeply redesigned to integrate AI at every level: contextual suggestions, automatic notification summaries, real-time transcription and instant translation even without a connection.

Another major hardware announcement: the Android XR glasses. Google has confirmed that this product will be released during 2026. Developed in partnership with eyewear manufacturers, they integrate Gemini directly into the lens, making it possible to interact with the physical environment through voice or gaze. It is Google's direct response to Meta's Ray-Ban glasses and the Apple Vision Pro.

Creative tools reinvent themselves

Google is also highlighting its AI creation tools, which have made considerable progress since their launch:

• Veo: AI video generation reaches a new level of realism, with high-definition clips lasting up to several minutes
• Lyria: music generation gains new voices and instruments, with the ability to create complete compositions from a simple text prompt
• ImageFX: the AI image editor gains new inpainting and frame extension features

These tools are gradually being integrated into Google Workspace products, which means they will be accessible directly from Docs, Slides or Gmail for business subscribers.

Firebase becomes a platform for autonomous agents

Less visible to the general public but very significant for developers: Firebase is repositioning itself as a native platform for autonomous agents. Google is integrating specific features to orchestrate AI agents, manage their memory, their access rights and their lifecycle. It is a strong signal that agentic AI is no longer a distant horizon, but a reality that technical teams must integrate into their architectures right now.

A context of accelerated transformation in the tech sector

Google I/O 2026 opens in a tech sector undergoing major change. On the same day, Meta triggers the first wave of its 8,000 layoffs, representing 10% of its global workforce. The group justifies this decision by the need to fund a $115 billion to $135 billion investment in AI in 2026. Amazon, Microsoft and Oracle have launched similar moves.

This context raises a fundamental question: will AI create as many jobs as it destroys? Big tech companies are betting that it will, but the current signals — 78,000 job cuts in the sector since January 2026 — call for caution.

What to remember

Google I/O 2026 marks a turning point. The company is no longer presenting only isolated AI features: it is unveiling a complete vertical integration strategy, from in-house TPU chips to connected glasses, via language models and operating systems. Google's challenge is to demonstrate that this full stack delivers a superior experience to that of its competitors.

For users, the announcements of May 19 could have a very concrete impact in the coming months: an assistant that truly manages your schedule, a phone that understands your context even better, and perhaps soon glasses that overlay AI onto the real world. The conference can be followed live tonight on YouTube from 7 p.m.