As the AI competition enters a feverish phase, Google CEO Sundar Pichai has officially confirmed that the company's next-generation large model, Gemini 3, will be released in 2025. Although Pichai cautiously stated that this upgrade "will not bring disruptive breakthroughs," this move clearly sends a signal — Google is systematically positioning itself to narrow the technical gap with OpenAI's GPT-5 and accelerate the construction of a complete AI ecosystem covering endpoints, cloud services, and infrastructure.
Gemini 3 focuses on "intelligent agents," 650 million monthly active users as a moat
As the core carrier of Google's AI strategy, the Gemini series has accumulated over 650 million monthly active users. Although it still lags behind ChatGPT in terms of user scale, it shows rapid growth. The upcoming Gemini 3 will focus on enhancing AI agent capabilities, enabling the model not only to answer questions but also to proactively plan, call tools, and perform multi-step tasks, providing users with a more coherent and autonomous interaction experience. This evolution from a "Q&A assistant" to an "action agent" is seen as a key watershed in the next generation of AI competition.
AI-driven search and cloud business grow together, 70% of enterprise customers are on board
Google's AI offensive is not limited to the model itself. Its traditional competitive advantage — search — is achieving new growth through deep integration of generative AI features, although some changes have sparked controversy by altering user habits. At the same time, Google Cloud has become a key engine for Alphabet's AI commercialization: over 70% of existing cloud customers have adopted its AI toolkits, including Vertex AI, Gemini API, and industry-specific solutions. The company is accelerating the construction of a full-stack AI infrastructure from chips (TPU v5), models, to applications to meet the growing demand for computing power and development needs of enterprises.
Multipronged approach: generative models + autonomous driving, global expansion in parallel
In addition to the Gemini main line, multiple generative AI projects are also being advanced internally at Google, covering areas such as multimodal understanding, code generation, and scientific computing, aiming to maintain a leading position at key technology nodes. Meanwhile, Waymo, the autonomous driving star under Alphabet, is leveraging the AI wave to accelerate its international expansion — planning to enter cities like London and Tokyo, expanding AI-driven mobility services to high-density urban areas around the world, further expanding its real-world data loop and user base.
Conclusion: Steady progress, Google's AI counterattack logic
Differing from OpenAI's emphasis on "paradigm breakthroughs," Google has chosen a more stable path: anchoring on its existing product matrix and achieving deep penetration of AI capabilities through incremental innovation. Gemini 3 may not be a "groundbreaking work," but it will be a crucial link for Google to connect the consumer side, enterprise side, and infrastructure side. In this AI competition that determines the dominant position for the next decade, 650 million active users, a global cloud network, and deep technological reserves are becoming Google's most solid counterattack capital.
