Google Bets on 'AI Assistant' Strategy... Unveils Gemini Spark and Omni

[Financial News, New York = Reporter Lee Byung-chul] Google unveiled a personal assistant-style AI agent and a video AI model that simulates the real world as it moves to strengthen its Artificial Intelligence (AI) competitiveness. As OpenAI and Anthropic push ahead with Initial Public Offering (IPO) plans and competition for leadership in the AI market intensifies, Google is also accelerating efforts to expand its AI ecosystem by combining search, cloud, and semiconductors.
At its annual developer conference, Google I/O 2026, held on the 19th in Mountain View, California, Google introduced a new AI agent, Gemini Spark, and a video generation model, Gemini Omni.
The main focus of the event was Agent AI. Google put AI at the forefront as a system that goes beyond simple question-and-answer functions to understand users' digital lives and carry out real actions on their behalf.
Alphabet CEO Sundar Pichai said, "We are focused on bringing cutting-edge agent capabilities to consumers in a safe and secure way."
The newly unveiled Gemini Spark connects with a range of Google services, including search, email, calendars, and documents. Depending on user instructions, it can automatically perform tasks such as web searches, information organization, reservations, and status updates. Google said it is currently testing the service with a limited group of users and plans to roll it out first next week to subscribers of Google AI Ultra, which costs about $100 per month.
Google also unveiled Gemini Omni, an AI based on a world model that simulates the real world. The model can generate new videos or edit existing scenes using not only text but also images, audio, and video.
Google explained, "Users can modify actions in videos they have shot or add new characters and objects."
Video generation AI is widely seen as a field that requires enormous computing resources. Earlier this year, OpenAI also scaled back development of its video generation model Sora and redirected computing resources to other projects. Against this backdrop, Google's decision to put video AI at the center is being viewed as a sign of confidence in its AI infrastructure capabilities.
Google also unveiled Gemini 3.5 Flash, which is optimized for coding and AI agent tasks. The company said it delivers speeds up to four times faster than existing state-of-the-art AI models. A more powerful Gemini 3.1 Pro will be released within weeks.
Google is also expanding investment in AI infrastructure. The company recently unveiled its eighth-generation TPU and is also pushing to sell chips directly to some customers. It also plans to establish an AI infrastructure company with private equity firm Blackstone to respond to the expanding market.
Market watchers say Google is rapidly transforming from a simple search company into an AI platform company. Alphabet's stock has risen about 25% this year, and its market capitalization is nearing $5 trillion. Revenue at Google Cloud in the first quarter of this year reached $20 billion, up 63% from a year earlier.
\r\n
\r\n\r\n
\r\n
pride@fnnews.com Reporter Lee Byung-chul Reporter