From the main stage to the terminal
summary
Gemini Enterprise Agent Platform now ships an end‑to‑end stack for building, scaling, governing and optimizing agents: Agent Development Kit (Python/Go/TypeScript/Java), Agent CLI that scaffolds agents into production, and governance features (gateway, cryptographic agent identity, registry, anomaly detection). New runtime pieces include memory bank and long running agents that persist state up to seven days, agent eval + simulation and tracing for observability. Gen models: Gemma 4 family (2–31B, 40M+ downloads, Apache 2.0), Nano Banana/Veo/Lyria multimodal media, and TPU V8 / Blackwell GPUs (96GB VRAM, ~5x inference gains) powering production demos that do autoscaling and low‑latency inference.
AI-written summary · about
chapters (12)
- 0:00Holding slide
- 9:02Intro
- 15:24Jason Davenport x Addy Osmani x Dave Elliott
- 44:22Jason Davenport x Ben Gilbert x David Rosenthal
- 1:13:08Jason Davenport x Shubham Saboo
- 1:47:27Jason Davenport x Philip Kiely x Jay Rodge
- 2:29:47Stephanie Wong x Yasmeen Ahmad
- 3:12:14Stephanie Wong x Khulan Davaajav
- 3:52:23Stephanie Wong x Katie Nguyen
- 4:24:55Aja Hammerly x Tomek Porozynski
- 4:51:00Muhammad Farooq x Omar Sanseviero
- 5:12:20Logan Kilpatrick x Sam Witteveen
related talks
39mBuild the new generation of AI Agents
10mBuilding intelligent agents with Gemma 3
45mCreate and deploy your first agent with Google Cloud Gemini Enterprise
49mLet's build an AI agent
23mScaling GenAI Applications and Agentic AI for Business Competitive Differentiation
37mBuild agents with Gemini API
24mScalable GenAI in Production
37mAGENTS OF S.E.A.L.E.D: AI Agentic Cybersecurity Framework
