Google Cloud Next 2025: Gemini 2.5 Flash, Workspace Upgrades, and Agentic AI Take Center Stage

Google's annual Google Cloud Next event has once again become a hub for AI announcements, with a strong focus on its Gemini models and advancements in agentic AI. Here’s a breakdown of the key updates:

Gemini 2.5 Flash: Faster, Cheaper, and Smarter

Google unveiled Gemini 2.5 Flash, a streamlined version of its Gemini 2.5 Pro model. Optimized for speed and cost-efficiency, Flash uses "test-time compute"—a technique that dynamically adjusts processing power based on task complexity. This approach, also seen in DeepSeek's R1 model, reduces operational costs while maintaining performance.

• Availability: Coming soon to Vertex AI, AI Studio, and the Gemini app.
• Gemini 2.5 Pro: Now in public preview on Vertex AI and Gemini, this model recently topped the Chatbot Arena leaderboard.

Workspace Gets an AI Boost

Google is integrating Gemini models into Google Workspace, introducing features like:

• Audio Docs: Convert text documents into audio.
• Automated Sheets Analysis: AI-driven data insights in Google Sheets.
• Workspace Flows: Automate workflows (e.g., customer service requests) across Workspace apps.

These enhancements are powered by agentic AI, which performs multi-step reasoning to execute complex tasks.

Adopting Anthropic’s Model Context Protocol (MCP)

To improve data access for AI models, Google is adopting MCP, an open-source standard developed by Anthropic. MCP enables secure connections between data sources and AI tools, ensuring more reliable responses.

• Why it matters: OpenAI has also adopted MCP, signaling industry-wide alignment on this framework.

What’s Next?

With more announcements expected on Day 2, Google continues to push the boundaries of AI innovation. Stay tuned for updates on Gemini, Workspace, and beyond.

Coverage by Cecily Mauran, Tech Reporter at Mashable. Follow her on X (@cecily_mauran).