Google pushed Gemini toward proactive behavior in July 2026, tying it more tightly to Google’s suite of apps. The company highlighted newer model families like Gemini 3.5 Flash and Gemini Omni, betting that tight integration with Gmail, Docs, and other services would make Gemini essential to daily work.

New features include screen recording reactions, AI-powered video and music creation, and floating app bubbles for easier multitasking. Google Docs expanded Gemini-powered document creation to 11 more languages, making the tool accessible to more international users.

Where Progress Is Slowing

Behind the feature announcements sits a tougher reality: Google is months behind on delivering Gemini 3.5 Pro, its most powerful flagship AI model. The company has taken extra time to improve capabilities—especially in coding—which means Gemini’s top tier isn’t available on the original timeline.

Coding is where frontier AI models prove their worth. If Google can’t ship a competitive coding engine, it loses credibility with developers, who often try multiple AI systems before committing to one.

New Hardware on the Horizon

Google is designing a new server chip internally called Frozen v2, aimed at release in 2028. The company claims Frozen v2 could be 6 to 10 times more efficient than existing AI chips, measured by tokens generated per unit of power.

Efficiency matters at scale. Every watt saved translates to lower operating costs and faster response times for users. If Frozen v2 delivers, it could give Google an infrastructure advantage over competitors.

Education Push

Google launched ATL Saathi, a Gemini-powered web app piloting in 100 Indian schools. The tool gives Tinkering Lab educators a 24/7 planning and training assistant, supporting curriculum-aligned guidance and micro-learning for teachers in resource-constrained settings.

Google’s Gemini strategy hinges on integration and reliability. Features matter less than making AI work within workflows people already use.

FYI (keeping you in the loop)

Why are tokens per second important?

Faster token generation means quicker AI responses. Users see answers instantly instead of waiting. At scale across millions of requests, speed is a competitive edge.