Google has launched Gemini 3.0, its latest and most powerful artificial intelligence model, in a sweeping integration across all its services, including a important overhaul of Google Search [[1]].The new model promises more nuanced and efficient responses to complex queries, but comes alongside cautions from CEO Sundar Pichai regarding the evolving risks within the rapidly advancing AI landscape. This rollout marks a departure for Google, delivering the update to search results promptly, impacting an estimated 13 billion queries daily [[2]].
Google has launched its next-generation artificial intelligence model, Gemini 3.0, globally, integrating it across all its services, including search. The new model is designed to improve understanding of complex queries and deliver more efficient responses. However, Google CEO Sundar Pichai also cautioned about potential risks in the rapidly evolving AI landscape.
Google began rolling out Gemini 3.0, its latest artificial intelligence model, on Tuesday, making it available across all Google services simultaneously. The most visible implementation for consumers is within Google Search, where the large language model is powering a new AI-driven overview feature.
Currently, Gemini AI is providing what Google calls an “AI Overview” for approximately 13 billion search queries each day. The model has reached 2 billion monthly users, making it the most widely used system of its kind globally, according to the company.
In a departure from previous releases, Google is deploying the newest version of Gemini directly into search results without a phased rollout. “This is the first time we are delivering Gemini in Search on day one,” CEO Sundar Pichai wrote in a blog post announcing the launch. He emphasized the model’s multimodal capabilities, stating, “It’s designed to understand depth and nuance – whether it’s recognizing subtle cues in a creative idea or deciphering the complex relationships in a challenging problem.” Pichai added that Gemini 3 is intended to better understand the context behind search queries, allowing users to achieve their desired results with fewer interactions.
Gemini 3 Aims to Outperform OpenAI, Meta, and X.ai
Google claims Gemini 3 significantly outperforms its predecessors and rivals from OpenAI, Meta, and X.ai in complex reasoning tasks. The company cited benchmark results demonstrating its superior performance. Gemini 3 can process text, images, videos, audio, and code, and is designed to assist users with learning, planning, and programming. This broad functionality positions Google to compete more directly with other AI platforms offering similar capabilities.
According to Demis Hassabis, head of Google DeepMind, the model represents “another significant step towards general AI.” For end-users, this translates into new, interactive search and learning features that not only deliver information but also present it visually and explain complex concepts. Google also highlighted the extensive safety testing conducted on Gemini 3, calling it the company’s most thoroughly vetted AI system to date.
The simultaneous rollout across all platforms suggests either exceptional efficiency in how Gemini 3.0 utilizes computing resources within Google’s data centers—or that the company currently has available capacity not being utilized by cloud customers. This comes after Google’s cloud division secured a deal with OpenAI in June to provide computing power.
Google’s AI models are optimized to run efficiently on its proprietary Tensor chips, potentially giving the company a cost advantage over competitors who rely on more expensive Nvidia chips for AI inference. The ability to efficiently execute Gemini AI internally could be a key differentiator in a competitive market.
Whether large language models can be operated profitably, given their substantial computing and energy demands, remains an open question within the current AI industry.
In a recent interview with the BBC, Google CEO Sundar Pichai warned of a potential overheating in the AI market. While acknowledging an “extraordinary moment” for AI investment, he also observed “a certain irrationality” in the current enthusiasm.
When asked if Google would be shielded from a potential AI bubble burst, Pichai responded candidly: “I think no company will be immune – including us.” The CEO’s assessment comes as valuations for technology companies reach record highs and doubts grow about the sustainability of the current boom.
This article was created for the economic competence center of WELT and “Business Insider Germany.”
Benedikt Fuest is a business editor and covers technology companies.
