Sundar Pichai outlined a transition in artificial intelligence from prioritizing model size to focusing on faster, cheaper, and more accessible AI technologies that can scale to billions of users globally.
- Google’s Gemini 3.5 nearly matches leading AI models with greater speed and lower cost
- AI usage on Google’s platforms surged to over 3 quadrillion tokens processed monthly
- Google expands AI agent tools in Workspace and multimodal AI capabilities
What happened
Sundar Pichai announced that the AI race is evolving beyond simply developing larger models to delivering AI solutions that are faster, more cost-effective, and easily accessible to billions of people. Google’s Gemini 3.5 Flash model now achieves close to 90% of the performance of the largest frontier AI models while being four times faster and costing significantly less to run.
Pichai also highlighted the massive growth in AI usage within Google’s ecosystem. The company now processes more than 3.2 quadrillion tokens monthly across its services, a dramatic increase from 480 trillion tokens a year prior. Over 8.5 million developers are leveraging Google’s Gemini models to build innovative applications, while popular consumer AI tools have attracted billions of active users.
Why it matters
This shift underscores a critical phase in AI development where scaling deployment efficiency and affordability are as crucial as raw model capabilities. By narrowing the performance gap while improving cost and speed, Google aims to democratize access to advanced AI technologies, making them practical for a broader user base across diverse sectors.
Pichai also noted that AI’s growing role in areas such as security highlights the ongoing value of large models for complex tasks. However, the strategic focus on practical deployment underlines a new competitive landscape where seamless integration, multimodal input handling, and agent-based assistance will differentiate market leaders.
What to watch next
Google is intensifying its efforts to embed AI agents into core products, including Gmail, Docs, and other Workspace tools, through innovations like Gemini Spark and Gemini Omni. These developments aim to automate routine tasks and generate rich multimedia content from varied inputs, expanding AI’s utility in everyday workflows.
Further, Google’s leadership in Search continues to be a major advantage as it transforms into a conversational, multimodal AI experience. Partnerships with industry players on AI content watermarking and commercial protocols hint at Google’s push toward transparent, scalable AI-driven commerce and user interaction models globally.