During the company’s annual developer conference in Mountain View, CEO Sundar Pichai remarked that artificial intelligence is “transforming every part of the company,” emphasizing that Google is experiencing a phase of “hyper progress” spurred by swift technological advancements and ongoing product releases.
Pichai noted that the company’s AI-driven transformation is already enhancing engagement, particularly with increased Search usage. He also highlighted the robust growth of the Gemini AI ecosystem, with the Gemini app reaching over 900 million monthly users, having more than doubled in usage over the past year.
At the conference, Google presented a set of developer tools within the “Antigravity” platform, designed to assist programmers in developing software and managing AI agents. This initiative leverages talent and technology gained through their acquisition of the startup Windsurf in a deal valued at $2.4 billion.
The company also introduced new AI models, including Gemini 3.5 Flash, which is touted as its most advanced coding-focused model to date. Another model, Gemini Omni, further broadens capabilities by enabling content generation across multiple formats—allowing users to create and modify videos through simple text, image, audio, or video prompts. Google indicated that future updates will enhance Omni’s functionality for image and audio generation.
Agentic commerce initiative: Universal Cart launched
Google also revealed a new AI-powered shopping system named Universal Cart at its I/O 2026 event, as part of its larger push into “agentic commerce.” This feature aims to streamline user shopping across Google services by merging product discovery, pricing insights, and checkout into a cohesive AI-enhanced experience.
Universal Cart operates across Google Search, Gemini, YouTube, and Gmail, enabling users to easily add items while browsing and automatically track them in a central cart.
The system utilizes Gemini AI models to monitor price fluctuations, stock levels, and historical pricing trends, as well as to flag compatibility issues—such as identifying mismatched components in complex product configurations and suggesting alternatives.
The cart integrates with Google Wallet and Google Pay, allowing users to finalize purchases with just a few taps or redirect to merchant websites. It also considers loyalty programs, card benefits, and promotions to highlight potential savings.
Initial retail partners include Nike, Sephora, Target, Ulta Beauty, Walmart, Wayfair, Fenty, and Steve Madden. The service will initially launch in the US via Search and the Gemini app, followed by YouTube and Gmail.
New commerce protocols: UCP and AP2 announced
Google also unveiled updates to its Universal Commerce Protocol (UCP) to standardize AI-driven shopping experiences across various platforms. Additionally, the company introduced the Agent Payments Protocol (AP2), which allows AI agents to make purchases on behalf of users while adhering to predefined spending limits and security measures.
The company stated that AP2 incorporates privacy-centric safeguards and digital verification systems to ensure transparency in AI-assisted transactions.
Gemini Omni Flash and video generation enhancements
Google introduced Gemini Omni Flash, a new AI video-generation model that is part of its Omni family. This model is regarded as a significant improvement over previous tools like Veo and Nano Banana, featuring enhanced capabilities for understanding motion, physics, and scene continuity.
During the keynote, Demis Hassabis from Google DeepMind demonstrated how the model can generate and manipulate videos using simple prompts, creating more lifelike and editable AI-generated scenes.
He stated that the model has a better grasp of complex concepts, such as kinetic and potential energy, facilitating more accurate movements and interactions in generated videos.
SynthID expansion for AI content identification
Google also expanded its SynthID watermarking system, designed to detect AI-generated content. The upgraded system will now encompass video, images, audio, and text, in collaboration with companies including NVIDIA, OpenAI, Kakao, and ElevenLabs.
SynthID embeds invisible digital watermarks into AI-generated content without compromising quality, enabling users and platforms to verify whether media has been produced or modified using AI tools.
Availability and rollout
Gemini Omni Flash is accessible starting today in the Gemini app, Google Flow, and YouTube Shorts, with API access for developers expected soon. The comprehensive Gemini Omni suite is being rolled out to Google AI Plus, Pro, and Ultra subscribers.
Also Read: Google, Blackstone to create AI cloud firm with in-house chips
