Google Cloud Subsequent 25 is happening right this moment, and Google has introduced quite a few new bulletins, primarily associated to enhancing Gemini choices and making it simpler to construct and undertake brokers.
Listed here are a few highlights from the occasion:
New instruments for constructing AI brokers
The corporate introduced the Agent Improvement Equipment (ADK), an open-source framework that covers the end-to-end technique of constructing and deploying brokers and multi-agent programs.
Google additionally introduced the Agent2Agent (A2A) protocol, an open protocol that permits AI brokers to speak with one another, trade data in a safe manner, and coordinate actions on prime of enterprise purposes. It was developed with help and contributions from over 50 different corporations, together with Atlassian, Field, MongoDB, Salesforce, ServiceNow.
Agentspace up to date with new options for simpler adoption of brokers
Google Agentspace is a platform the place builders can construct brokers combining Gemini fashions, Google-quality search performance, and enterprise knowledge.
The corporate introduced that organizations will now be capable of give their workers entry to the platform’s unified enterprise search, evaluation, and synthesis capabilities from inside Chrome’s search field.
Different new capabilities embody a no-code Agent Designer for constructing customized brokers, and entry to 2 new brokers constructed by Google: Deep Analysis and Thought Era.
seventh era TPU, Ironwood, is introduced for late 2025 launch
Ironwood is a Tensor Processing Unit (TPU) that’s designed particularly for inference functions. It is going to be accessible in two sizes: 256 or 9,216 chip configurations.
In comparison with its sixth gen TPU, Trillium, it achieves a 2x enchancment in efficiency/watt. It additionally options enhancements in Excessive Bandwidth Reminiscence capability and bandwidth, and Inter-Chip Interconnect (ICI) bandwidth.
“Ironwood represents a big shift within the improvement of AI and the infrastructure that powers its progress. It’s a transfer from responsive AI fashions that present real-time data for individuals to interpret, to fashions that present the proactive era of insights and interpretation. That is what we name the “age of inference” the place AI brokers will proactively retrieve and generate knowledge to collaboratively ship insights and solutions, not simply knowledge,” Amin Vahdat, VP and GM of ML, Programs, and Cloud AI at Google, wrote in a weblog submit.
New capabilities in Vertex AI
Lyria—Google’s text-to-music mannequin—has been added in Vertex AI as a non-public preview.
Different updates embody:
- New modifying and digicam management options in Veo 2 (preview)
- Immediate Customized Voice in Chirp 3 to allow customers to create customized voices from 10 seconds of audio
- Improved picture era and inpainting capabilities in Imagen 3 for fixing lacking parts of a picture or eradicating objects
New Gemini capabilities in Google Workspace
The corporate is introducing Google Workspace Flows, enabling customers to create agentic workflows to streamline processes and automate repetitive duties throughout Workspace.
Audio capabilities are additionally being built-in into Google Docs to allow customers to create audio variations of their paperwork, together with producing podcast-style overviews.
Different new Workspace AI capabilities embody an AI writing device in Google Docs, Veo 2 picture era in Google Vids, new Gemini capabilities in Google Chat, and AI-powered knowledge evaluation and surfacing of key insights in Google Sheets.
