← All stories
● Covered by 12 sources · 92 reportsMedium impact1 negative53 neutral27 positive

Google Expands Gemini Enterprise Agent Platform with Remote MCP Server

🔄 Updated 1d ago — new reporting from Google Cloud Blog
New to BrevFeed? We gather this story from every outlet covering it into one summary — ranked by real-world impact, not just the latest headline — so you never miss what matters. What is BrevFeed? →

Key points

  • New remote MCP server introduced for Gemini Enterprise Agent Platform.
  • Enables secure connection between external AI agents and Google Cloud.
  • Improves integration with various development environments and IDEs.
  • Supports custom function calling and credential refreshing.
  • Aims to streamline agent development and deployment.
  • Google Cloud AI published a guide for startups on transitioning AI prototypes to production.
  • The guide addresses API key management, platform migration, and quota limitations.
  • The guide provides 10 questions across onboarding, scaling, and governance phases.
  • Google Antigravity is now available for enterprise customers.
  • Antigravity is part of eligible Gemini Enterprise app subscriptions.
  • Antigravity includes administrative and spend controls.
  • New IDE extensions let developers use Antigravity in VS Code.
  • AI developer tools are included in Gemini Enterprise subscriptions.
  • Administrators can enable Antigravity and Android Studio for eligible Gemini Enterprise Standard, Plus, and Standard Emerging Market licenses.
  • Gemini Enterprise for Financial Services is introduced for capital markets and corporate banking workflows.
  • Gemini Enterprise for Legal is introduced for contract review and legal research.
  • Google Developer Relations launched "Agent Valley," a free, five-week live learning series.
  • Agent Valley teaches building AI agent systems from basic concepts.
  • Agent Valley uses hands-on exercises in a virtual world environment.
  • Annie Wang, a Google DevRel Engineer, designed the Agent Valley course.
  • Expanded billing flexibility and cost management tools are launched for AI agent workloads.
  • New pay-as-you-go option is available in Gemini Enterprise app for agent workloads.
  • Gemini Enterprise for Legal automates regulatory monitoring, document drafting, and data discovery.
  • Gemini Enterprise for Financial Services is geared towards financial professionals.
  • Gemini Enterprise for Legal and Financial Services are the first in a series of specialized industry solutions.
  • Gemini Enterprise for Legal debuted 24 hours after Thomson Reuters' Thomson AI model.
  • Google Cloud launched AI-powered Database Operations Agents.
  • Database Operations Agents include an Onboarding Agent and an Observability Agent.
  • The Onboarding Agent streamlines database setup.
  • The Observability Agent automates troubleshooting, performance optimization, and tuning.
  • Database Operations Agents integrate with Gemini Cloud Assist.
  • Database Operations Agents support AlloyDB, Bigtable, and Spanner.
  • The Onboarding Agent understands technical metrics like IOPS, latency limits, and replication lag.
  • The Onboarding Agent recommends database solutions based on workload characteristics, performance, scale, data type, and reliability.
  • The Onboarding Agent generates commands to provision, configure, and deploy database instances.
  • Pythian developed an AI Operating Model after deploying Google Cloud's Gemini Enterprise internally.
  • Pythian's AI Operating Model led to a 3x increase in user engagement.
  • Pythian's AI Operating Model led to an 80% reduction in database incident resolution times.
  • Agentic video understanding launched for Gemini 3.7 Flash, 3.6 Flash, and 3.5 Flash-Lite models.
  • Agentic video understanding improves video analysis accuracy.
  • Agentic video understanding reduces token usage and costs for long-form video content.
  • Agentic video understanding uses Gemini's native video tools.
  • Agentic video understanding enables sub-second moment retrieval, anomaly detection, and precise counting.
  • Agentic video understanding is available via the Gemini API in Google AI Studio.
  • Agentic video understanding dynamically searches, scans, and inspects video segments across visual frames, audio, and transcripts.
  • Google launched the Fairwind Program for select Google Cloud customers, government agencies, and cybersecurity partners.
  • The Fairwind Program offers early access to advanced AI cyber defense capabilities.
  • The program uses Gemini 3.8 Flash Cyber and CodeMender to find, verify, and fix software vulnerabilities.
  • The Fairwind Program aims to provide rapid patch generation.
  • Google Cloud is named a Leader in the 2026 Gartner Magic Quadrant for Strategic Cloud Platform Services.
  • Antigravity SDK provides the runtime engine from Antigravity 2.0 and Antigravity CLI.
  • Antigravity SDK includes declarative safety policies, real-time telemetry, and stateful multi-turn persistence.
  • Antigravity SDK allows developers to create centralized systems for monitoring and managing LLM workloads.
  • Google is a Leader in the 2026 Gartner Magic Quadrant for Enterprise AI Assistants.
  • Gartner evaluated Google based on its Completeness of Vision and Ability to Execute.
  • Peder Ulander is Vice President, Product Marketing, Google Cloud.
  • Google is promoting its "Intelligent Endpoints" strategy.
  • The strategy integrates AI into Chrome and other enterprise products.
  • The strategy aims to automate tasks and improve productivity.
  • The strategy enables agentic workflows directly within the browser.
  • The strategy includes a new enterprise Skills library for Gemini in Chrome.
  • David Still is Vice President, Product & Engineering, Android Enterprise.
  • Mark Berschadski is Director, Product Management, Chrome Enterprise & Security.
  • Darren Mowry is VP, Global Startups and Investor Ecosystem, Google.
  • Startups are choosing Google Cloud at a higher rate than three years ago.
  • Gemini CLI 0.61.0 was released on Wednesday.
  • Gemini CLI 0.61.0 requires explicit confirmation before editing build configuration files.
  • Gemini CLI 0.61.0 requires explicit confirmation before running build or test commands after edits.
  • Gemini CLI 0.61.0 requires explicit confirmation before executing shell commands from untrusted external content.
  • Gemini CLI 0.61.0 hardens Gemini CLI's optional sandbox.
  • Google announced at I/O in May that Gemini CLI Pro, Ultra, and free-tier users would move to Antigravity CLI.
  • Since June 18, the open-source Gemini CLI has served mainly enterprise customers and developers.
  • Andre Alves is Director, SMB Sales LATAM.
  • Latin American SMBs show 8x year-over-year growth in Google Cloud AI adoption.
  • Brazilian SMBs show 9x growth in Google Cloud AI adoption.
  • Google Earth Engine launched "Ask" to accelerate geospatial coding.
  • "Ask" integrates Gemini AI into the Google Earth Engine Code Editor.
  • "Ask" helps users write, debug, understand, and optimize geospatial queries.
  • Users can use their Gemini API key with "Ask" in the Code Editor.
  • Joel Conkling is Product Lead for Google Earth Engine.
  • Katie Friis is a Software Engineer for Google Earth Engine.
  • Google Cloud expanded its Gemini Enterprise ecosystem with new partner-built security agents and AI defenses.
  • New security agents and integrations from leading cybersecurity vendors are available.
  • New offerings include partner security agents invoked directly in Gemini Enterprise and protections for AI/agentic workloads.
  • Vineet Bhan is Director Security and Identity Partnerships.
  • Ashish Verma is Head of Partner Engineering, Security.
  • Google Cloud CLI remote MCP server is in preview.
  • Google Cloud CLI remote MCP server enables AI agents to access gcloud and bq command-line operations.
  • Prosper Nwankpa is Senior Engineering Manager, Google Cloud.
  • Adam Hwang is Software Engineering Manager, Google Cloud.
  • CloudBees is shifting to an AI-first organization.
  • Mo Plassnig is the new CEO of CloudBees.
  • Mo Plassnig announced CloudBees' AI-first pivot at the WeAreDevelopers Conference in San Jose, California.
  • AI21 Labs reduced high-priority AI job wait times from 72 hours to 12 hours.
  • AI21 Labs eliminated manual scheduling interventions for AI workloads.
  • AI21 Labs uses Google Cloud A3 and A3 Ultra instances for model training.
  • Google Cloud A3 instances are powered by NVIDIA H100 Tensor Core GPUs.
  • Google Cloud A3 Ultra instances are powered by NVIDIA H200 Tensor Core GPUs.
  • AI21 Labs' model training runs on Google Kubernetes Engine (GKE) clusters.
  • Barak Peleg is VP, Technology and Architecture at AI21.
  • Asaf Ben-Tovim is DevOps Engineer at AI21.
  • Google Cloud launched Google Cloud Modernize, a new portfolio.
  • Google Cloud Modernize integrates existing migration and modernization tools with AI capabilities.
  • Google Cloud Modernize includes Migration Center, VMware Engine, Mainframe Modernization, and EKS-to-GKE Migration Agent.
  • Modernization Hub is a new in-console experience for developers and architects.
  • Modernization Hub analyzes source code, maps dependencies, and accelerates modernization for Java, .NET, and mainframe applications.
  • Agentic Quick Estimator converts VMware inventory exports and infrastructure inputs into total cost.
  • The Gemini agent is a new universal AI agent for enterprise customers.
  • The Gemini agent is available in Gemini Enterprise, Workspace, and third-party services.
  • The Gemini agent can handle knowledge work, create media, and write/run code.
  • The Gemini agent can utilize various AI models including Gemini and Claude.
  • The Gemini agent can take objectives rather than specific instructions.
  • The Gemini agent can function as a personal assistant and team member.
  • The Gemini agent integrates with existing tools and operates across various devices and applications.
  • The Gemini agent works directly inside Gmail, Drive, Docs, Slides, Sheets, Chat, and Calendar.
  • The Gemini agent has new data and analytics skills.
  • The Gemini agent offers leading cost controls through multi-model orchestration, Smart Routing, and real-time spend caps.
  • The SUNY AI Platform is a secure AI and high-performance computing environment built on Google Cloud with Gemini Enterprise.
  • The SUNY AI Platform is initially available across nine of SUNY’s research-intensive campuses.
  • Google Cloud estimates its operations will add 10 billion euros to Irish GDP in 2025.
  • Google will initially focus on bringing the agent to businesses before rolling it out to consumers.
  • Sundar Pichai stated Gemini has over 1 billion monthly active users.
  • Nearly 90% of Fortune 100 businesses now use Gemini Enterprise at work.
  • Antigravity SDK supports local models for AI agent workflows.
  • Antigravity SDK allows hybrid agent orchestration.
  • Antigravity SDK implements the Architect-Builder pattern.
  • Gemini 4 Argon frontier model is now widely available.
  • Gemini 4 Argon is trained to maintain context across long, multi-step tasks.
  • Gemini is integrated with Borderless Lakehouse.
  • The Gemini-Borderless Lakehouse integration allows AI agents to access and reason over data across multiple clouds and on-premises systems.
  • The Gemini-Borderless Lakehouse integration enables natural language querying of cross-cloud data.
  • The Gemini-Borderless Lakehouse integration is currently in preview.
  • Google is a Leader in the 2026 IDC MarketScape for Agentic AI Platforms.
  • IDC MarketScape evaluated Google's Gemini platform for deploying AI agents.
  • MCP Apps are introduced for Gemini Enterprise.
  • MCP Apps enable interactive user interfaces within the chat flow.
  • MCP Apps allow visual, multi-step work without constant model interaction.
  • MCP Apps render UIs directly in the chat flow inside a sandboxed iframe.
  • MCP Apps can expand to fullscreen or a side panel.
  • MCP Apps edits run locally while the workspace and assistant stay synchronized.
  • MCP Apps use a private Cloud Run service for hosting the MCP server.
  • Enterprise search with Gemini on Google Distributed Cloud (GDC) air-gapped is in preview.
  • Managed RAG search operates within a customer's physical security perimeter without public internet.
  • Gemini for government services is in preview.
  • Gemini for government services is a new suite of AI tools for public sector work.
  • Gemini for government services adds skills built around government work.
  • Gemini for government services connects to existing agency systems.
  • Gemini for government services includes specialized partner agents.
  • Google Distributed Cloud (GDC) air-gapped environments now support open-weight models.
  • Agent Compute Engine is a new compute service for AI agents.
  • Agent Compute Engine will be available in preview in the coming months.
  • Agent Compute Engine provides hardware-isolated runtime sandbox security.
  • Agent Compute Engine powers Gemini for work tasks.
  • Agent Compute Engine provides hardware-isolated microVMs for AI agents.
  • Agent Compute Engine combines container speed with sandbox security.
  • Agent Compute Engine originated from Google's internal agentic workloads.
  • Agent Compute Engine was announced at AGNTCon + MCPCon.
  • E2B and Daytona offer agent sandboxes.
  • AWS Bedrock AgentCore offers agent sandboxes.
  • E2B runs sandboxes in Firecracker microVMs.
  • E2B can pause and resume a sandbox with memory and filesystem intact in about one second.
  • Daytona added microVM sandboxes in July.
  • Daytona microVMs can pause a full machine or fork a running one.
  • AWS Bedrock AgentCore gives each agent session its own microVM.
  • AWS does not charge for CPU while an agent waits on a model.
  • Gemini 4 Argon optimized the Agent Development Kit's 200,000-line Python codebase.
  • The optimization reduced in-memory session-preparation overhead by 40%.
  • The optimization reduced cold package import time by 21%.
  • The optimization was achieved in an 11-hour unattended run.
  • The Agent Development Kit codebase spans 800 files and 17,000 unit tests.
  • Gemini Robotics Embodied Reasoning 2 (ER 2) is now in public preview on the Gemini platform within Google Cloud.
  • Gemini Robotics ER 2 enables robots to converse with humans, orchestrate action models, and execute multi-minute tasks.
  • Google Cloud published a guide on connecting Gemini Enterprise to Snowflake data warehouses using MCP and ADK.
  • The guide demonstrates how AI agents can generate and execute SQL queries with user-level permissions in Snowflake.
  • Google Cloud and Oracle expanded their partnership to integrate Oracle AI Database@Google Cloud with Google's Agentic Data Cloud and Gemini Enterprise.
  • The integration aims to unify Oracle's operational data with Google Cloud's data and AI ecosystem.
  • Google Skills launched free video learning bundles for Gemini.
  • Gary Eimerman is Managing Director, Google Cloud Learning.
  • Google Cloud introduced cross-cloud interconnect and caching for AWS and Azure.
  • Sirish Chandrasekaran is VP, Product Strategy & Growth, Data Cloud.
  • Raj Pai is VP, Product Management, Databases.
  • PayPal, Siemens, Telus, Vodafone, and Walmart use Google's Agentic Data Cloud.
  • Google introduced a method to optimize AI agent skills using AlphaEvolve and Harbor.
  • The system allows AI agents to evaluate and improve their operational instructions.
  • The system iteratively proposes, tests, and scores skill variations on real-world coding tasks.
  • Ivan Nardini is an AI Developer Relations Engineer.
  • Anant Nawalgaria is a Group Product Manager & Sr. Staff Engineer.
  • Coding agents like Google Antigravity and OpenCode are used for software engineering tasks.
  • OpenCode with Gemini 3.8 Flash on Terminal-Bench 2.0 passed about two-thirds of its attempts.
  • Data Cloud connectors in Gemini for business are now in preview.
  • Data Cloud connectors allow querying live enterprise data in natural language.
  • Data Cloud connectors enforce row- and column-level security and end-user credentials.
  • Antigravity version 2.29 is rolling out.
  • Antigravity 2.29 features a redesigned conversation page with the prompt box at the bottom.
  • Antigravity 2.29 has a default Google Sans font and Onyx theme.
  • Antigravity 2.29's single-repo projects combine branch and Git worktree selection into one menu.
  • Antigravity 2.29's agent picker now lives inside the model menu.
  • Antigravity 2.29 includes a new Voice Mode powered by Gemini Live.
  • Antigravity Voice Mode allows talking to Antigravity and handing requests by voice.
  • Antigravity Voice Mode offers various voice options and customizable speed from 0.75x to 2x.
  • Antigravity Remote Agents are rolling out to paid Google AI Pro and Ultra subscribers.
  • Remote Agents create isolated cloud environments that persist when a laptop is closed.
  • Remote Agents can be controlled from antigravity.google.com/remote.
  • Google Cloud introduced TPU 8tx and TPU 8i Max.
  • TPU 8tx is for secure and confidential inference.
  • TPU 8i Max focuses on massive context windows and high compute density.
  • Google Cloud Next in April introduced TPU 8t and 8i.
  • TPU 8t and 8i offer up to 2.7 times better training performance.
  • TPU 8t and 8i offer 80% better inference price performance.
  • Gemini for healthcare is an AI solution for healthcare organizations.
  • Gemini for healthcare integrates and orchestrates tasks across clinical, financial, and administrative systems.
  • Gemini for healthcare aims to reduce manual data reconciliation for clinicians.
  • Gemini for healthcare is built to support HIPAA compliance.
  • Gemini for healthcare is available in preview.
  • Thomas Kurian is CEO, Google Cloud.
  • Gemini for healthcare is initially available with launch partners HCA Healthcare, Highmark Health, and Northwell Health.
  • Google DeepMind released Gemini Robotics 2.5.
  • Gemini Robotics 2.5 shows improved dexterity and speed in tasks.
  • Gemini Robotics 2.5 was evaluated across more than 200 activities.
  • Gemini Robotics 2.5 was used in Mercedes-Benz factories.
  • Google DeepMind introduced an open-source safety benchmark called ASIMOV-jailbreak.
  • ASIMOV-jailbreak prepares for human-robot collaboration.
  • Gemini Embodied Reasoning (ER) is the vision-language model piece of the physical AI platform.
  • Gemini Robotics is the vision-language-action (VLA) component.
  • Apptronik’s Apollo 2 humanoid was shown with Gemini Robotics 2.5.
  • Google Cloud launched the Gemini factory, a co-innovation program.
  • The Gemini factory helps enterprise software partners integrate solutions with Gemini.
  • The Gemini factory aims to speed up development of AI solutions listed on Google Cloud Marketplace.
  • Noel Kenehan is Head of Gemini factory.
  • Renee Li is Technical Solutions Consultant.
  • Google Cloud introduced the Professional Agentic Architect certification.
  • The certification is part of the Gemini Agent Ready (GEAR) initiative.
  • The certification has a hybrid exam format with conceptual knowledge and hands-on labs.
  • Google launched the Gemini Decisions API.
  • The Gemini Decisions API is powered by the new Gemini Decision 1 model.
  • The Gemini Decisions API processes multimodal inputs across text, images, audio, video, and PDFs.
  • The Gemini Decisions API returns probabilities and confidence scores for up to 1,000 questions in 150-250 ms.
  • The Gemini Decisions API costs $0.10 per 1M input tokens with zero output tokens billed.
  • Gemini Decisions API evaluates predefined options in a single pass.
  • Gemini Decisions API supports a 256,000-token context window.
  • OpenAI, TypeSafe AI, and Microsoft also offer Decisions APIs.
  • Antigravity's /teamwork is a new multi-agent system built on Gemini models.
  • Antigravity's /teamwork enhances human-AI collaboration in science and engineering.
  • Antigravity's /teamwork is now available externally for researchers and developers.
  • Antigravity's /teamwork was launched on November 13, 2026.
  • Vahab Mirrokni is Google Fellow and VP, Google Research.
  • Antigravity's /teamwork uses Gemini 4 Carbon and Gemini 4 Hydrogen models.
  • Antigravity's /teamwork includes /boost and other agents.
  • Gemini AI models are available on Google Cloud Dedicated in Europe.
  • Google Cloud Dedicated is a partner-operated sovereign cloud offering.
  • Google for Startups Accelerator: Sovereign AI program is open for applications in Europe and Israel.
  • The accelerator supports startups building AI solutions compliant with EU AI Act, NIS2, and DORA.
  • Google Cloud is a Leader in the 2026 Gartner Magic Quadrant for AI Application Development Platforms.
  • Google Cloud achieved the highest position for Ability to Execute in the Gartner Magic Quadrant.
  • Google Cloud has been a Leader in this Gartner Magic Quadrant for three consecutive times.
  • Mike Clark is Director, Product Management at Google Cloud.
  • Nearly 500 Google Cloud customers each processed over one trillion tokens in the last year.
  • Envoy is evolving to support autonomous AI agents and their networking requirements.
  • Envoy addresses payload blindness and token budget management challenges for AI workloads.
  • Envoy is positioned as an open standard for AI traffic, MCP tooling, and agent-to-agent communication.
  • Abhi Maras is Sr. Product Manager at Google Cloud.
  • Boteng Yao is Software Engineer at Google Cloud.
  • Google Cloud is highlighting EMEA startups deploying autonomous AI agents at Slush 2026.
  • CDON, Chalkie.ai, and Finova are EMEA startups using Google Cloud for AI agents.
  • Google Cloud introduced new Gemini-powered co-marketing features in its Partner Marketing Studio.
  • New features allow partners to generate customized marketing assets, ensure brand compliance, and automate social media posts.
  • Aimee Catalano is Managing Director, Global Partner Marketing at Google Cloud.
  • Sergio Villani is Technical Solutions, Cloud AI at Google Cloud.
  • Automated scrapers can find exposed API keys within minutes.
  • A compromised API key can lead to tens of thousands of dollars in unauthorized charges.
  • Customers are responsible for usage made with their own valid credentials.

Overview of Gemini Enterprise's New Features

Google announced an enhancement to its Gemini Enterprise Agent Platform by integrating a remote Managed Control Plane (MCP) server. This latest update allows developers to connect their external AI agents securely to resources available in Google Cloud, thereby reducing setup time and ensuring compliance in agent development processes.

Background and Developer Benefits

The remote MCP server serves as a bridge between developers' preferred Integrated Development Environments (IDEs) and Google Cloud. This ensures that external AI agent developers can interact smoothly with resources in the Agent Platform, such as calling models from the Model Garden or managing project-specific Notebooks. Developers can use standard tools like Antigravity CLI and Claude Code, providing them with flexibility and choice.

The introduction is part of a broader update that also includes features such as background execution and custom function calling within the Gemini API. These capabilities directly address constraints previously faced by developers in building reliable AI agents.

Implications for AI Development

These advancements signify a step forward in the development and deployment of AI agents within the Google Cloud environment. By allowing seamless integration and improved control over cloud resources, developers can ensure their agents are secure, efficient, and more production-ready. The updates represent Google's response to developer feedback seeking such enhancements.

As AI agents become increasingly central to enterprise operations, the ability to quickly and securely build them in any preferred development environment could impact how AI solutions are developed industry-wide.

Takeaways and Future Prospects

The addition of the remote MCP server and the support for advanced features like background task execution enhances the Gemini Enterprise Agent Platform's utility, making it more appealing to developers seeking efficient ways to leverage cloud-based AI resources. These features are poised to facilitate the evolution of AI agent deployment, offering a more streamlined and secure way of integrating AI capabilities into enterprise solutions.

Overall, these updates could lead to increased adoption of Google Cloud's AI frameworks, influencing the wider industry trends towards integrated AI development processes across cloud platforms.

Updates

🕒 2026-11-20 · new reporting from Google Cloud Blog
  • Sergio Villani is Technical Solutions, Cloud AI at Google Cloud.
  • Automated scrapers can find exposed API keys within minutes.
  • A compromised API key can lead to tens of thousands of dollars in unauthorized charges.
  • Customers are responsible for usage made with their own valid credentials.
🕒 2026-11-19 · new reporting from Google Cloud Blog
  • Google Cloud introduced new Gemini-powered co-marketing features in its Partner Marketing Studio.
  • New features allow partners to generate customized marketing assets, ensure brand compliance, and automate social media posts.
  • Aimee Catalano is Managing Director, Global Partner Marketing at Google Cloud.
🕒 2026-11-18 · new reporting from Google Cloud Blog
  • Envoy is evolving to support autonomous AI agents and their networking requirements.
  • Envoy addresses payload blindness and token budget management challenges for AI workloads.
  • Envoy is positioned as an open standard for AI traffic, MCP tooling, and agent-to-agent communication.
  • Abhi Maras is Sr. Product Manager at Google Cloud.
  • Boteng Yao is Software Engineer at Google Cloud.
  • Google Cloud is highlighting EMEA startups deploying autonomous AI agents at Slush 2026.
  • CDON, Chalkie.ai, and Finova are EMEA startups using Google Cloud for AI agents.
🕒 2026-11-17 · new reporting from Google Cloud Blog
  • Google Cloud is a Leader in the 2026 Gartner Magic Quadrant for AI Application Development Platforms.
  • Google Cloud achieved the highest position for Ability to Execute in the Gartner Magic Quadrant.
  • Google Cloud has been a Leader in this Gartner Magic Quadrant for three consecutive times.
  • Mike Clark is Director, Product Management at Google Cloud.
  • Nearly 500 Google Cloud customers each processed over one trillion tokens in the last year.
🕒 2026-11-17 · new reporting from Google Cloud Blog
  • Gemini AI models are available on Google Cloud Dedicated in Europe.
  • Google Cloud Dedicated is a partner-operated sovereign cloud offering.
  • Google for Startups Accelerator: Sovereign AI program is open for applications in Europe and Israel.
  • The accelerator supports startups building AI solutions compliant with EU AI Act, NIS2, and DORA.
🕒 2026-11-14 · new reporting from Hacker News Front Page
  • Antigravity's /teamwork is a new multi-agent system built on Gemini models.
  • Antigravity's /teamwork enhances human-AI collaboration in science and engineering.
  • Antigravity's /teamwork is now available externally for researchers and developers.
  • Antigravity's /teamwork was launched on November 13, 2026.
  • Vahab Mirrokni is Google Fellow and VP, Google Research.
  • Antigravity's /teamwork uses Gemini 4 Carbon and Gemini 4 Hydrogen models.
  • Antigravity's /teamwork includes /boost and other agents.
🕒 2026-11-12 · new reporting from The New Stack
  • Gemini Decisions API evaluates predefined options in a single pass.
  • Gemini Decisions API supports a 256,000-token context window.
  • OpenAI, TypeSafe AI, and Microsoft also offer Decisions APIs.
🕒 2026-11-11 · new reporting from Google Cloud Blog, Hacker News Front Page
  • Google Cloud introduced the Professional Agentic Architect certification.
  • The certification is part of the Gemini Agent Ready (GEAR) initiative.
  • The certification has a hybrid exam format with conceptual knowledge and hands-on labs.
  • Google launched the Gemini Decisions API.
  • The Gemini Decisions API is powered by the new Gemini Decision 1 model.
  • The Gemini Decisions API processes multimodal inputs across text, images, audio, video, and PDFs.
  • The Gemini Decisions API returns probabilities and confidence scores for up to 1,000 questions in 150-250 ms.
  • The Gemini Decisions API costs $0.10 per 1M input tokens with zero output tokens billed.
🕒 2026-11-10 · new reporting from Google Cloud Blog, Google AI Blog, Ars Technica, Engadget
  • Gemini for healthcare is an AI solution for healthcare organizations.
  • Gemini for healthcare integrates and orchestrates tasks across clinical, financial, and administrative systems.
  • Gemini for healthcare aims to reduce manual data reconciliation for clinicians.
  • Gemini for healthcare is built to support HIPAA compliance.
  • Gemini for healthcare is available in preview.
  • Thomas Kurian is CEO, Google Cloud.
  • Gemini for healthcare is initially available with launch partners HCA Healthcare, Highmark Health, and Northwell Health.
  • Google DeepMind released Gemini Robotics 2.5.
  • Gemini Robotics 2.5 shows improved dexterity and speed in tasks.
  • Gemini Robotics 2.5 was evaluated across more than 200 activities.
  • Gemini Robotics 2.5 was used in Mercedes-Benz factories.
  • Google DeepMind introduced an open-source safety benchmark called ASIMOV-jailbreak.
  • ASIMOV-jailbreak prepares for human-robot collaboration.
  • Gemini Embodied Reasoning (ER) is the vision-language model piece of the physical AI platform.
  • Gemini Robotics is the vision-language-action (VLA) component.
  • Apptronik’s Apollo 2 humanoid was shown with Gemini Robotics 2.5.
  • Google Cloud launched the Gemini factory, a co-innovation program.
  • The Gemini factory helps enterprise software partners integrate solutions with Gemini.
  • The Gemini factory aims to speed up development of AI solutions listed on Google Cloud Marketplace.
  • Noel Kenehan is Head of Gemini factory.
  • Renee Li is Technical Solutions Consultant.
🕒 2026-11-09 · new reporting from Google Cloud Blog
  • Google Cloud introduced TPU 8tx and TPU 8i Max.
  • TPU 8tx is for secure and confidential inference.
  • TPU 8i Max focuses on massive context windows and high compute density.
  • Google Cloud Next in April introduced TPU 8t and 8i.
  • TPU 8t and 8i offer up to 2.7 times better training performance.
  • TPU 8t and 8i offer 80% better inference price performance.
🕒 2026-11-06 · new reporting from Google Cloud Blog, 9to5Google
  • Data Cloud connectors in Gemini for business are now in preview.
  • Data Cloud connectors allow querying live enterprise data in natural language.
  • Data Cloud connectors enforce row- and column-level security and end-user credentials.
  • Antigravity version 2.29 is rolling out.
  • Antigravity 2.29 features a redesigned conversation page with the prompt box at the bottom.
  • Antigravity 2.29 has a default Google Sans font and Onyx theme.
  • Antigravity 2.29's single-repo projects combine branch and Git worktree selection into one menu.
  • Antigravity 2.29's agent picker now lives inside the model menu.
  • Antigravity 2.29 includes a new Voice Mode powered by Gemini Live.
  • Antigravity Voice Mode allows talking to Antigravity and handing requests by voice.
  • Antigravity Voice Mode offers various voice options and customizable speed from 0.75x to 2x.
  • Antigravity Remote Agents are rolling out to paid Google AI Pro and Ultra subscribers.
  • Remote Agents create isolated cloud environments that persist when a laptop is closed.
  • Remote Agents can be controlled from antigravity.google.com/remote.
🕒 2026-11-05 · new reporting from Google Cloud Blog
  • Google introduced a method to optimize AI agent skills using AlphaEvolve and Harbor.
  • The system allows AI agents to evaluate and improve their operational instructions.
  • The system iteratively proposes, tests, and scores skill variations on real-world coding tasks.
  • Ivan Nardini is an AI Developer Relations Engineer.
  • Anant Nawalgaria is a Group Product Manager & Sr. Staff Engineer.
  • Coding agents like Google Antigravity and OpenCode are used for software engineering tasks.
  • OpenCode with Gemini 3.8 Flash on Terminal-Bench 2.0 passed about two-thirds of its attempts.
🕒 2026-11-04 · new reporting from Google Cloud Blog
  • Google Cloud introduced cross-cloud interconnect and caching for AWS and Azure.
  • Sirish Chandrasekaran is VP, Product Strategy & Growth, Data Cloud.
  • Raj Pai is VP, Product Management, Databases.
  • PayPal, Siemens, Telus, Vodafone, and Walmart use Google's Agentic Data Cloud.
🕒 2026-10-30 · new reporting from Google Cloud Blog
  • Google Skills launched free video learning bundles for Gemini.
  • Gary Eimerman is Managing Director, Google Cloud Learning.
🕒 2026-10-27 · new reporting from Google Cloud Blog
  • Gemini Robotics Embodied Reasoning 2 (ER 2) is now in public preview on the Gemini platform within Google Cloud.
  • Gemini Robotics ER 2 enables robots to converse with humans, orchestrate action models, and execute multi-minute tasks.
  • Google Cloud published a guide on connecting Gemini Enterprise to Snowflake data warehouses using MCP and ADK.
  • The guide demonstrates how AI agents can generate and execute SQL queries with user-level permissions in Snowflake.
  • Google Cloud and Oracle expanded their partnership to integrate Oracle AI Database@Google Cloud with Google's Agentic Data Cloud and Gemini Enterprise.
  • The integration aims to unify Oracle's operational data with Google Cloud's data and AI ecosystem.
🕒 2026-10-24 · new reporting from InfoQ
  • Gemini 4 Argon optimized the Agent Development Kit's 200,000-line Python codebase.
  • The optimization reduced in-memory session-preparation overhead by 40%.
  • The optimization reduced cold package import time by 21%.
  • The optimization was achieved in an 11-hour unattended run.
  • The Agent Development Kit codebase spans 800 files and 17,000 unit tests.
🕒 2026-10-22 · new reporting from Google Cloud Blog, The New Stack
  • Agent Compute Engine is a new compute service for AI agents.
  • Agent Compute Engine will be available in preview in the coming months.
  • Agent Compute Engine provides hardware-isolated runtime sandbox security.
  • Agent Compute Engine powers Gemini for work tasks.
  • Agent Compute Engine provides hardware-isolated microVMs for AI agents.
  • Agent Compute Engine combines container speed with sandbox security.
  • Agent Compute Engine originated from Google's internal agentic workloads.
  • Agent Compute Engine was announced at AGNTCon + MCPCon.
  • E2B and Daytona offer agent sandboxes.
  • AWS Bedrock AgentCore offers agent sandboxes.
  • E2B runs sandboxes in Firecracker microVMs.
  • E2B can pause and resume a sandbox with memory and filesystem intact in about one second.
  • Daytona added microVM sandboxes in July.
  • Daytona microVMs can pause a full machine or fork a running one.
  • AWS Bedrock AgentCore gives each agent session its own microVM.
  • AWS does not charge for CPU while an agent waits on a model.
🕒 2026-10-20 · new reporting from Google Cloud Blog
  • Enterprise search with Gemini on Google Distributed Cloud (GDC) air-gapped is in preview.
  • Managed RAG search operates within a customer's physical security perimeter without public internet.
  • Gemini for government services is in preview.
  • Gemini for government services is a new suite of AI tools for public sector work.
  • Gemini for government services adds skills built around government work.
  • Gemini for government services connects to existing agency systems.
  • Gemini for government services includes specialized partner agents.
  • Google Distributed Cloud (GDC) air-gapped environments now support open-weight models.
🕒 2026-10-19 · new reporting from Google Cloud Blog
  • Google is a Leader in the 2026 IDC MarketScape for Agentic AI Platforms.
  • IDC MarketScape evaluated Google's Gemini platform for deploying AI agents.
  • MCP Apps are introduced for Gemini Enterprise.
  • MCP Apps enable interactive user interfaces within the chat flow.
  • MCP Apps allow visual, multi-step work without constant model interaction.
  • MCP Apps render UIs directly in the chat flow inside a sandboxed iframe.
  • MCP Apps can expand to fullscreen or a side panel.
  • MCP Apps edits run locally while the workspace and assistant stay synchronized.
  • MCP Apps use a private Cloud Run service for hosting the MCP server.
🕒 2026-10-15 · new reporting from Google Cloud Blog
  • Gemini 4 Argon frontier model is now widely available.
  • Gemini 4 Argon is trained to maintain context across long, multi-step tasks.
  • Gemini is integrated with Borderless Lakehouse.
  • The Gemini-Borderless Lakehouse integration allows AI agents to access and reason over data across multiple clouds and on-premises systems.
  • The Gemini-Borderless Lakehouse integration enables natural language querying of cross-cloud data.
  • The Gemini-Borderless Lakehouse integration is currently in preview.
🕒 2026-10-14 · new reporting from InfoQ
  • Antigravity SDK supports local models for AI agent workflows.
  • Antigravity SDK allows hybrid agent orchestration.
  • Antigravity SDK implements the Architect-Builder pattern.
🕒 2026-10-08 · new reporting from 9to5Google, Google Cloud Blog, TechCrunch
  • The Gemini agent is a new universal AI agent for enterprise customers.
  • The Gemini agent is available in Gemini Enterprise, Workspace, and third-party services.
  • The Gemini agent can handle knowledge work, create media, and write/run code.
  • The Gemini agent can utilize various AI models including Gemini and Claude.
  • The Gemini agent can take objectives rather than specific instructions.
  • The Gemini agent can function as a personal assistant and team member.
  • The Gemini agent integrates with existing tools and operates across various devices and applications.
  • The Gemini agent works directly inside Gmail, Drive, Docs, Slides, Sheets, Chat, and Calendar.
  • The Gemini agent has new data and analytics skills.
  • The Gemini agent offers leading cost controls through multi-model orchestration, Smart Routing, and real-time spend caps.
  • The SUNY AI Platform is a secure AI and high-performance computing environment built on Google Cloud with Gemini Enterprise.
  • The SUNY AI Platform is initially available across nine of SUNY’s research-intensive campuses.
  • Google Cloud estimates its operations will add 10 billion euros to Irish GDP in 2025.
  • Google will initially focus on bringing the agent to businesses before rolling it out to consumers.
  • Sundar Pichai stated Gemini has over 1 billion monthly active users.
  • Nearly 90% of Fortune 100 businesses now use Gemini Enterprise at work.
🕒 2026-10-05 · new reporting from Google Cloud Blog
  • Google Cloud launched Google Cloud Modernize, a new portfolio.
  • Google Cloud Modernize integrates existing migration and modernization tools with AI capabilities.
  • Google Cloud Modernize includes Migration Center, VMware Engine, Mainframe Modernization, and EKS-to-GKE Migration Agent.
  • Modernization Hub is a new in-console experience for developers and architects.
  • Modernization Hub analyzes source code, maps dependencies, and accelerates modernization for Java, .NET, and mainframe applications.
  • Agentic Quick Estimator converts VMware inventory exports and infrastructure inputs into total cost.
🕒 2026-10-02 · new reporting from Google Cloud Blog
  • AI21 Labs reduced high-priority AI job wait times from 72 hours to 12 hours.
  • AI21 Labs eliminated manual scheduling interventions for AI workloads.
  • AI21 Labs uses Google Cloud A3 and A3 Ultra instances for model training.
  • Google Cloud A3 instances are powered by NVIDIA H100 Tensor Core GPUs.
  • Google Cloud A3 Ultra instances are powered by NVIDIA H200 Tensor Core GPUs.
  • AI21 Labs' model training runs on Google Kubernetes Engine (GKE) clusters.
  • Barak Peleg is VP, Technology and Architecture at AI21.
  • Asaf Ben-Tovim is DevOps Engineer at AI21.
🕒 2026-09-30 · new reporting from Google Cloud Blog, The New Stack
  • Google Cloud CLI remote MCP server is in preview.
  • Google Cloud CLI remote MCP server enables AI agents to access gcloud and bq command-line operations.
  • Prosper Nwankpa is Senior Engineering Manager, Google Cloud.
  • Adam Hwang is Software Engineering Manager, Google Cloud.
  • CloudBees is shifting to an AI-first organization.
  • Mo Plassnig is the new CEO of CloudBees.
  • Mo Plassnig announced CloudBees' AI-first pivot at the WeAreDevelopers Conference in San Jose, California.
🕒 2026-09-29 · new reporting from Google Cloud Blog
  • Google Cloud expanded its Gemini Enterprise ecosystem with new partner-built security agents and AI defenses.
  • New security agents and integrations from leading cybersecurity vendors are available.
  • New offerings include partner security agents invoked directly in Gemini Enterprise and protections for AI/agentic workloads.
  • Vineet Bhan is Director Security and Identity Partnerships.
  • Ashish Verma is Head of Partner Engineering, Security.
🕒 2026-09-28 · new reporting from Google Cloud Blog
  • Google Earth Engine launched "Ask" to accelerate geospatial coding.
  • "Ask" integrates Gemini AI into the Google Earth Engine Code Editor.
  • "Ask" helps users write, debug, understand, and optimize geospatial queries.
  • Users can use their Gemini API key with "Ask" in the Code Editor.
  • Joel Conkling is Product Lead for Google Earth Engine.
  • Katie Friis is a Software Engineer for Google Earth Engine.
🕒 2026-09-24 · new reporting from Google Cloud Blog, The New Stack
  • Darren Mowry is VP, Global Startups and Investor Ecosystem, Google.
  • Startups are choosing Google Cloud at a higher rate than three years ago.
  • Gemini CLI 0.61.0 was released on Wednesday.
  • Gemini CLI 0.61.0 requires explicit confirmation before editing build configuration files.
  • Gemini CLI 0.61.0 requires explicit confirmation before running build or test commands after edits.
  • Gemini CLI 0.61.0 requires explicit confirmation before executing shell commands from untrusted external content.
  • Gemini CLI 0.61.0 hardens Gemini CLI's optional sandbox.
  • Google announced at I/O in May that Gemini CLI Pro, Ultra, and free-tier users would move to Antigravity CLI.
  • Since June 18, the open-source Gemini CLI has served mainly enterprise customers and developers.
  • Andre Alves is Director, SMB Sales LATAM.
  • Latin American SMBs show 8x year-over-year growth in Google Cloud AI adoption.
  • Brazilian SMBs show 9x growth in Google Cloud AI adoption.
🕒 2026-09-23 · new reporting from Google Cloud Blog
  • Google is promoting its "Intelligent Endpoints" strategy.
  • The strategy integrates AI into Chrome and other enterprise products.
  • The strategy aims to automate tasks and improve productivity.
  • The strategy enables agentic workflows directly within the browser.
  • The strategy includes a new enterprise Skills library for Gemini in Chrome.
  • David Still is Vice President, Product & Engineering, Android Enterprise.
  • Mark Berschadski is Director, Product Management, Chrome Enterprise & Security.
🕒 2026-09-09 · new reporting from Google Cloud Blog
  • Google is a Leader in the 2026 Gartner Magic Quadrant for Enterprise AI Assistants.
  • Gartner evaluated Google based on its Completeness of Vision and Ability to Execute.
  • Peder Ulander is Vice President, Product Marketing, Google Cloud.
🕒 2026-09-08 · new reporting from Google Cloud Blog
  • Antigravity SDK provides the runtime engine from Antigravity 2.0 and Antigravity CLI.
  • Antigravity SDK includes declarative safety policies, real-time telemetry, and stateful multi-turn persistence.
  • Antigravity SDK allows developers to create centralized systems for monitoring and managing LLM workloads.
🕒 2026-09-03 · new reporting from Google Cloud Blog
  • Google Cloud is named a Leader in the 2026 Gartner Magic Quadrant for Strategic Cloud Platform Services.
🕒 2026-09-02 · new reporting from Google AI Blog
  • Google launched the Fairwind Program for select Google Cloud customers, government agencies, and cybersecurity partners.
  • The Fairwind Program offers early access to advanced AI cyber defense capabilities.
  • The program uses Gemini 3.8 Flash Cyber and CodeMender to find, verify, and fix software vulnerabilities.
  • The Fairwind Program aims to provide rapid patch generation.
🕒 2026-09-01 · new reporting from Google DeepMind
  • Agentic video understanding launched for Gemini 3.7 Flash, 3.6 Flash, and 3.5 Flash-Lite models.
  • Agentic video understanding improves video analysis accuracy.
  • Agentic video understanding reduces token usage and costs for long-form video content.
  • Agentic video understanding uses Gemini's native video tools.
  • Agentic video understanding enables sub-second moment retrieval, anomaly detection, and precise counting.
  • Agentic video understanding is available via the Gemini API in Google AI Studio.
  • Agentic video understanding dynamically searches, scans, and inspects video segments across visual frames, audio, and transcripts.
🕒 2026-08-27 · new reporting from Google Cloud Blog
  • Pythian developed an AI Operating Model after deploying Google Cloud's Gemini Enterprise internally.
  • Pythian's AI Operating Model led to a 3x increase in user engagement.
  • Pythian's AI Operating Model led to an 80% reduction in database incident resolution times.
🕒 2026-08-27 · new reporting from InfoQ
  • Google Cloud launched AI-powered Database Operations Agents.
  • Database Operations Agents include an Onboarding Agent and an Observability Agent.
  • The Onboarding Agent streamlines database setup.
  • The Observability Agent automates troubleshooting, performance optimization, and tuning.
  • Database Operations Agents integrate with Gemini Cloud Assist.
  • Database Operations Agents support AlloyDB, Bigtable, and Spanner.
  • The Onboarding Agent understands technical metrics like IOPS, latency limits, and replication lag.
  • The Onboarding Agent recommends database solutions based on workload characteristics, performance, scale, data type, and reliability.
  • The Onboarding Agent generates commands to provision, configure, and deploy database instances.
🕒 2026-08-26 · new reporting from The New Stack
  • Gemini Enterprise for Legal automates regulatory monitoring, document drafting, and data discovery.
  • Gemini Enterprise for Financial Services is geared towards financial professionals.
  • Gemini Enterprise for Legal and Financial Services are the first in a series of specialized industry solutions.
  • Gemini Enterprise for Legal debuted 24 hours after Thomson Reuters' Thomson AI model.
🕒 2026-08-26 · new reporting from Google Cloud Blog
  • Google Developer Relations launched "Agent Valley," a free, five-week live learning series.
  • Agent Valley teaches building AI agent systems from basic concepts.
  • Agent Valley uses hands-on exercises in a virtual world environment.
  • Annie Wang, a Google DevRel Engineer, designed the Agent Valley course.
  • Expanded billing flexibility and cost management tools are launched for AI agent workloads.
  • New pay-as-you-go option is available in Gemini Enterprise app for agent workloads.
🕒 2026-08-25 · new reporting from Google Cloud Blog
  • Gemini Enterprise for Financial Services is introduced for capital markets and corporate banking workflows.
  • Gemini Enterprise for Legal is introduced for contract review and legal research.
🕒 2026-08-20 · new reporting from Google Cloud Blog
  • Google Cloud AI published a guide for startups on transitioning AI prototypes to production.
  • The guide addresses API key management, platform migration, and quota limitations.
  • The guide provides 10 questions across onboarding, scaling, and governance phases.
  • Google Antigravity is now available for enterprise customers.
  • Antigravity is part of eligible Gemini Enterprise app subscriptions.
  • Antigravity includes administrative and spend controls.
  • New IDE extensions let developers use Antigravity in VS Code.
  • AI developer tools are included in Gemini Enterprise subscriptions.
  • Administrators can enable Antigravity and Android Studio for eligible Gemini Enterprise Standard, Plus, and Standard Emerging Market licenses.

✨ This summary was generated by AI from the outlets' reporting listed below. It is not independently verified and may contain errors — check the original sources. How BrevFeed works →

The daily brief

One email each morning: the day's tech stories, clustered across outlets and summarized. No account needed.

One email a day. Unsubscribe in one click, any time.

Today's brief

Spend a few minutes, get the whole day. Every topic's top stories in one hands-free rundown — listen, watch, or read the transcript.

~4 min · 3 stories · Nov 21

▶ Play today's brief Listen on Spotify

New every morning, and the back catalogue is archived by date.

How outlets covered it

This guide details how to secure API keys for AI prototypes, specifically addressing the risks of using raw API keys in production environments. It outlines steps for authorizing team members and applications, and for managing remaining keys with Secret Manager, spend caps, and monitoring.

Google Cloud introduced new Gemini-powered co-marketing features within its Partner Marketing Studio. These features allow partners to generate customized marketing assets, ensure brand compliance, and automate social media posts, aiming to scale campaign execution and improve pipeline generation.

Google Cloud highlights several EMEA startups that are deploying autonomous AI agents built on its infrastructure and Gemini models. These companies, including CDON, Chalkie.ai, and Finova, are using AI to automate complex workflows and enhance their respective industries.

Google Cloud is evolving Envoy, an open-source edge and service proxy, to support autonomous AI agents and their unique networking requirements. This adaptation addresses challenges like payload blindness and token budget management that traditional Layer 7 proxies face with agentic workloads, positioning Envoy as a potential open standard for AI traffic.

Google Cloud has been named a Leader in the 2026 Gartner Magic Quadrant for AI Application Development Platforms for the third consecutive time, achieving the highest position for Ability to Execute. This recognition highlights Google Cloud's focus on providing an enterprise AI platform, Gemini, that supports production scale, compliance, and measurable ROI for customers across various industries.

Google has opened applications for its Google for Startups Accelerator: Sovereign AI program in Europe and Israel. This initiative supports startups developing AI solutions that comply with European data governance regulations like the EU AI Act, NIS2, and DORA, aiming to foster AI innovation within the region's regulatory framework.

Google is making its Gemini AI models available on Google Cloud Dedicated, its partner-operated sovereign cloud offering in Europe. This expansion provides European organizations with more AI model choices under local control, addressing concerns about data residency, access, and regulatory compliance within the region.

Google Research has launched Antigravity's /teamwork, a multi-agent system built on Gemini models, to enhance human-AI collaboration in solving complex problems across various scientific and engineering fields. This system has already contributed to advancements in information theory, molecular structure discoveries, and kernel optimization, and is now being made available externally for researchers and developers.

Google launched the Gemini Decisions API, featuring the Gemini Decision 1 model, designed to handle routine AI decisions without generating text. This API offers lower costs and faster response times compared to traditional generative models by evaluating predefined options in a single pass. It aims to optimize AI application efficiency for tasks like content moderation and routing.

Google launched the Gemini Decisions API, powered by the new Gemini Decision 1 model, for making fast, closed-set decisions. This API processes multimodal inputs across text, images, audio, video, and PDFs, returning probabilities and confidence scores for up to 1,000 questions in 150-250 ms.

Google Cloud introduced the Professional Agentic Architect certification, part of its Gemini Agent Ready (GEAR) initiative, to validate expertise in building production-ready AI agents. This certification, previously an internal training program for Google Cloud engineers, is now available to the public and features a hybrid exam format combining conceptual knowledge with hands-on lab exercises.

Google Cloud introduced Gemini for healthcare, an agentic AI designed to improve workflow efficiency in the healthcare sector. This new platform integrates with third-party systems to manage health records, billing, and official guidelines, aiming to reduce administrative burden and support human oversight.

Google Cloud introduced the Gemini factory, a co-innovation program to help enterprise software partners integrate their solutions with Gemini. The program aims to reduce engineering friction and speed up the development of AI solutions listed on Google Cloud Marketplace.

Google DeepMind released Gemini Robotics 2.5, an AI platform for robots, which shows improved dexterity and speed in tasks. The company also introduced an open-source safety benchmark called ASIMOV-jailbreak to prepare for human-robot collaboration. This update aims to advance robotic capabilities in real-world applications, as demonstrated by its use in Mercedes-Benz factories.

Google Cloud launched Gemini for healthcare, an AI solution designed to integrate fragmented clinical, operational, and administrative systems. This platform aims to convert disparate healthcare data into actionable insights, reducing the time clinicians spend on administrative tasks.

Google Cloud has launched Gemini for healthcare, an AI solution designed to integrate and orchestrate tasks across clinical, financial, and administrative systems in healthcare organizations. This specialized agent aims to reduce manual data reconciliation for clinicians and improve efficiency for providers and health plans by automating multi-step workflows and providing unified patient information. The solution is built to support HIPAA compliance and is available in preview.

Google Cloud introduced two new additions to its 8th-generation TPU portfolio: the TPU 8tx and TPU 8i Max. The TPU 8tx is designed for secure and confidential inference, while the TPU 8i Max focuses on massive context windows and high compute density. These new systems address specific customer needs for agentic workloads and specialized encryption capabilities.

Google has released Antigravity version 2.29, featuring cloud-based Remote Agents for paid subscribers, a new Voice Mode powered by Gemini Live, and a redesigned user interface. This update expands Antigravity's capabilities for agentic development by enabling persistent cloud environments and voice interaction, alongside numerous usability improvements.

Google has introduced Data Cloud connectors in Gemini for business, now in preview, allowing users to query live enterprise data using natural language. This feature aims to provide direct access to governed data from sources like BigQuery, Spanner, and Looker, addressing challenges with traditional data querying and early generative AI tools.

Google has introduced a method to optimize AI agent skills using AlphaEvolve and Harbor, addressing the challenge of manually written and inconsistently effective agent instructions. This system allows AI agents to evaluate and improve their own operational instructions by iteratively proposing, testing, and scoring skill variations on real-world coding tasks.

Google has announced updates to its Agentic Data Cloud, integrating Gemini directly with the borderless Lakehouse and introducing cross-cloud interconnect and caching for AWS and Azure. These changes allow agents to access and process enterprise data across various environments using natural language queries and secure connections.

Google Skills has launched new, free video learning bundles for Gemini, tailored for financial services and legal professionals. These courses demonstrate how to use Gemini for tasks like analyzing market data and reviewing legal documents, aiming to improve workflow efficiency.

Google Cloud and Oracle expanded their partnership to integrate Oracle AI Database@Google Cloud with Google's Agentic Data Cloud and Gemini Enterprise. This integration aims to enable enterprises to build autonomous, agentic workflows by unifying Oracle's operational data with Google Cloud's data and AI ecosystem.

Google Cloud published a guide on securely connecting Gemini Enterprise to Snowflake data warehouses using the Model Context Protocol (MCP) and Google Agent Development Kit (ADK). This integration allows AI agents to generate and execute SQL queries with user-level permissions, enabling natural language interaction with data while maintaining security and governance.

Google DeepMind's Gemini Robotics Embodied Reasoning 2 (ER 2) model is now available in public preview on the Gemini platform within Google Cloud. This release allows enterprises to integrate the advanced robotics intelligence layer with Google Cloud's security and management tools for deploying adaptive physical AI at scale.

Google used its Gemini 4 Argon model to optimize the 200,000-line Python codebase for its Agent Development Kit (ADK) in an 11-hour unattended run. The AI produced two accepted optimizations, reducing in-memory session-preparation overhead by 40% and cold package import time by 21%. This demonstrates Gemini 4 Argon's ability to handle complex, existing codebases for performance improvements.

Google will make the infrastructure powering its Gemini agents available to Google Cloud users as Agent Compute Engine later this year. This service provides hardware-isolated microVMs for AI agents, combining container speed with sandbox security, and originated from Google's internal agentic workloads.

Google Cloud introduced Agent Compute Engine, a new compute service designed for AI agents, and expanded agent governance capabilities for Gemini for business. These offerings address the infrastructure challenges of securely running and scaling autonomous agents, providing hardware-isolated runtime security and optimized performance for iterative AI workloads.

Google Cloud launched Gemini for government services in preview, a new suite of AI tools designed for public sector work. This offering integrates the Gemini agent with specialized skills, connectors to agency systems, and partner agents to automate tasks while adhering to government regulations and security requirements.

Google Distributed Cloud (GDC) air-gapped environments now offer enterprise search capabilities with Gemini and support for open-weight models. This update allows organizations with strict data sovereignty requirements to deploy advanced AI directly within their isolated infrastructures.

Google has introduced enterprise search capabilities with Gemini on Google Distributed Cloud (GDC) air-gapped, now in preview. This fully managed retrieval-augmented generation (RAG) search operates entirely within a customer's physical security perimeter, without public internet connectivity. It allows organizations with strict data sovereignty mandates, such as national security agencies, to analyze sensitive data using AI search and multimodal intelligence on-premises.

Google Cloud has introduced MCP Apps (Model Context Protocol Apps) for Gemini Enterprise, enabling interactive user interfaces directly within the chat flow. This development allows for visual, multi-step work to be handled more efficiently without constant model interaction, addressing limitations of text-only enterprise assistants.

Google has been named a Leader in the inaugural IDC MarketScape: Worldwide Agentic AI Platforms 2026 Vendor Assessment. This recognition highlights Google's Gemini platform for its ability to deploy AI agents across enterprises while meeting requirements for quality, integration, security, and governance.

Google Cloud announced a new integration between Gemini and its Borderless Lakehouse, allowing AI agents to access and reason over data across multiple clouds and on-premises systems. This integration enables natural language querying of cross-cloud data and grounds AI agent reasoning without data movement, addressing challenges in enterprise AI adoption.

Google launched the Gemini agent, a universal AI assistant, and made the Gemini 4 Argon frontier model widely available. These tools are designed to assist startups with various tasks, reducing operational overhead and accelerating product development.

Google's Antigravity SDK now supports local models, allowing developers to run AI agent workflows directly on their machines while integrating cloud-based models for higher-level tasks. This update provides greater control over data privacy, inference costs, and computational distribution by enabling hybrid agent orchestration.

Google introduced an agentic version of Gemini, initially for businesses, that can execute tasks and connect to internal systems. This development moves AI beyond conversational interfaces to automated task completion, leveraging Gemini's existing business adoption.

Google Cloud announced the Gemini agent, a universal AI agent for work that integrates directly into Workspace applications like Gmail, Drive, and Docs. This agent handles tasks from knowledge work to code generation, offering specialized tools for industries like financial services and legal teams.

Google Public Sector and the State University of New York (SUNY) launched the SUNY AI Platform, a secure AI and high-performance computing environment built on Google Cloud with Gemini Enterprise. The platform will accelerate university research and prepare students and the workforce for an AI-driven economy by providing access to advanced AI capabilities and training.

Irish organizations, including government agencies and enterprises, are adopting Gemini Enterprise to integrate agentic AI into their business and operational workflows. This move supports Ireland's National Digital and AI Strategy and aims to improve productivity and innovation across sectors. Google Cloud estimates its operations, including AI and cloud services, will add 10 billion euros to Irish GDP in 2025.

Google Cloud announced a new Gemini agent designed for small and midsize businesses (SMBs) to enhance operational efficiency. This agent integrates with existing tools, offers model choice, and operates across various devices and applications, aiming to provide AI capabilities for diverse business functions.

Google Cloud announced the "Gemini agent," a new universal AI agent for enterprise customers, available in Gemini Enterprise, Workspace, and third-party services. This agent can handle knowledge work, create media, and write/run code by taking objectives rather than specific instructions, and it can utilize various AI models including Gemini and Claude.

Google Cloud launched Google Cloud Modernize, a new portfolio integrating existing migration and modernization tools with AI capabilities. This initiative aims to accelerate enterprise cloud transformations by providing agentic features for infrastructure assessment, platform, and application modernization.

AI21 Labs adopted Google Cloud AI Hypercomputer, reducing high-priority AI job wait times from 72 hours to 12 hours and eliminating manual scheduling interventions. This change addresses challenges in managing high-utilization GPU clusters for foundation model training and agent optimization products.

CloudBees, a continuous integration/delivery (CI/CD) company, is shifting its focus to become an AI-first organization under new CEO Mo Plassnig. This strategic change addresses the increasing volume of machine-generated code and the challenges enterprises face in managing it.

Google Cloud introduced a preview of the Google Cloud CLI remote MCP server, enabling AI agents to access gcloud and bq command-line operations. This server provides a standardized way for agents to manage Google Cloud infrastructure and BigQuery workflows, addressing challenges in agent interaction with backend systems.

Google Cloud has expanded its Gemini Enterprise ecosystem with new partner-built security agents and AI defenses. These integrations allow enterprises to orchestrate multi-step security workflows and leverage AI-powered capabilities from a unified interface, addressing the increasing use of AI in cyberattacks.

Google Earth Engine launched "Ask," a new feature integrating Gemini AI capabilities directly into its Code Editor. This allows users to write, debug, understand, and optimize geospatial queries using AI, accelerating the development of geospatial insights.

Latin American small and medium-sized businesses (SMBs) are increasingly adopting Google Cloud AI tools, with an 8x year-over-year growth across the region and 9x growth in Brazil. These businesses are using Gemini Enterprise and other Google Cloud AI offerings to automate tasks, improve customer support, and analyze data.

Google has released Gemini CLI 0.61.0, which now requires explicit user confirmation before the agent can edit build configuration files, run build or test commands after such edits, or execute shell commands derived from untrusted external content. This update enhances security by reintroducing human oversight at critical points where an autonomous coding agent could be exploited by attackers.

Google Cloud states that startups are increasingly choosing its platform for AI development, citing its comprehensive stack as a key driver. The company highlights the growth of Gemini Enterprise and the strong price-performance of Gemini models for various AI workloads.

Google is promoting its "Intelligent Endpoints" strategy, integrating AI into Chrome and other enterprise products to automate tasks and improve productivity. This initiative aims to reduce manual work for employees by enabling agentic workflows directly within the browser and through a new enterprise Skills library for Gemini in Chrome.

Google Cloud launched a new plugin for AI coding agents, designed to bundle skills and tools for Google Cloud interactions. This plugin addresses the challenge of managing individual AI agent skills by packaging related capabilities into cohesive, installable units, making agents more effective for Google Cloud users.

Google has been recognized as a Leader in Gartner's first 2026 Magic Quadrant for Enterprise AI Assistants, based on its Completeness of Vision and Ability to Execute. This placement highlights Google's Gemini Enterprise platform, which aims to provide a unified and secure AI solution for businesses.

Google Cloud introduced the Antigravity SDK, providing the runtime engine from Antigravity 2.0 and Antigravity CLI for building custom agent hubs. This SDK includes declarative safety policies, real-time telemetry, and stateful multi-turn persistence, allowing developers to create centralized systems for monitoring and managing LLM workloads.

Google Cloud has been named a Leader in the 2026 Gartner Magic Quadrant for Strategic Cloud Platform Services. This recognition highlights Google Cloud's integrated technology stack, dynamic infrastructure, and focus on digital sovereignty, which are presented as key differentiators for customers building and scaling demanding workloads.

Google has launched the Fairwind Program, offering select Google Cloud customers, government agencies, and cybersecurity partners early access to advanced AI cyber defense capabilities. The program utilizes Gemini 3.8 Flash Cyber and CodeMender to autonomously find, verify, and fix software vulnerabilities, aiming to provide rapid patch generation and enhance security for critical infrastructure and public services.

Google has launched agentic video understanding across its Gemini 3.7 Flash, 3.6 Flash, and 3.5 Flash-Lite models. This new capability improves video analysis accuracy while significantly reducing token usage and costs, particularly for long-form video content.

Pythian developed an AI Operating Model after deploying Google Cloud's Gemini Enterprise internally, aiming to achieve significant ROI from enterprise AI initiatives. The model addresses common failures in enterprise AI adoption by integrating strategy, execution, and operations into a continuous framework, which Pythian claims led to a 3x increase in user engagement and an 80% reduction in database incident resolution times.

Google Cloud has launched AI-powered Database Operations Agents, including an Onboarding Agent for database setup and an Observability Agent for troubleshooting and performance optimization. These agents integrate with Gemini Cloud Assist and support various Google Cloud database services, aiming to automate and simplify database lifecycle management for developers and engineers.

Google Cloud introduced Gemini Enterprise for Legal, an agentic AI solution designed to automate legal tasks such as contract review and document drafting. This release, alongside Gemini Enterprise for Financial Services, indicates a trend towards specialized AI solutions for professional domains, highlighting a competitive shift in enterprise AI development.

Google has launched expanded billing flexibility and cost management tools for AI agent workloads within Gemini Enterprise and developer tools like Google Antigravity and Android Studio. These updates aim to provide businesses with clearer visibility and proactive controls over AI spending, offering options like pay-as-you-go, consolidated developer access, and Flexible Savings Plans.

Google Developer Relations has launched "Agent Valley," a free, five-week live learning series designed to teach participants how to build AI agent systems from basic concepts. The program aims to make AI agent development accessible to beginners through hands-on exercises in a virtual world environment.

Google Cloud has launched Gemini Enterprise for Legal, a new AI solution designed specifically for the legal industry. This offering integrates specialized AI capabilities with existing legal systems to handle tasks like contract review and legal research, addressing the need for AI that adheres to legal industry standards and data security requirements.

Google Cloud has introduced Gemini Enterprise for Financial Services, an AI solution designed to integrate agentic AI into capital markets and corporate banking workflows. This offering aims to provide financial institutions with specialized AI capabilities, secure data integration, and auditable research agents, addressing the industry's need for precision and security in AI applications.

Google Antigravity, a coding agent tool, is now available for enterprise customers as part of eligible Gemini Enterprise app subscriptions, including administrative and spend controls. This update allows developers to use Antigravity within their preferred IDEs, such as VS Code, and provides organizations with consolidated governance over AI developer tools.

Google Cloud AI published a guide for startups on transitioning AI prototypes to production, addressing common challenges like API key management, platform migration, and quota limitations. The guide provides 10 questions across onboarding, scaling, and governance phases to help startups avoid typical failure modes when deploying AI solutions.

Looker's semantic layer now integrates with Gemini Enterprise, providing a governed foundation for structured data within Google's AI chat interface. This integration allows Gemini Enterprise users to query both structured and unstructured data in natural language, aiming to reduce AI hallucinations and improve data consistency.

Google has been named a Leader in The Forrester Wave™: AI Platforms, Q3 2026 report, receiving the highest score in the Strategy category. This recognition highlights Google's Gemini Enterprise platform, which unifies enterprise data, model capabilities, developer tooling, and IT operations for building and deploying AI agents.

Google Cloud detailed its strategies for identifying and mitigating security threats, including AI workload exploitation, cryptocurrency mining, credential exfiltration, and account takeovers. The company uses infrastructure telemetry and active monitoring to protect customer data and systems.

Google Cloud is providing free educational resources, including hands-on labs and courses, to help developers build and deploy AI agents using the Gemini Enterprise Agent Ready (GEAR) framework. These resources cover topics from foundational understanding to orchestrating multi-agent workflows and earning skill badges.

Google Cloud announced two new AI-powered database agents at Google Cloud Next ‘26: the Database Onboarding Agent for setup and configuration, and the Database Observability Agent for monitoring and troubleshooting. These agents automate database lifecycle management tasks, aiming to reduce manual effort and improve operational efficiency for database users.

Pillar Security identified an agent-to-agent attack method within Google's Agent Development Kit (ADK) for Python, which could expose secrets and enable pull request tampering. This vulnerability allowed manipulation of low-privileged AI agents to gain access to restricted capabilities of high-privileged agents, posing a supply chain risk.

Google Cloud announced the general availability of Cortex Framework version 7, which provides purpose-built data product accelerators for SAP, deployed in BigQuery and Knowledge Catalog. This release modernizes data architecture to support AI agents by transforming SAP transactional records into AI-ready data products, simplifying orchestration, and reducing infrastructure overhead.

Google provided insight into the development process and quality control mechanisms for Google Agent Skills, which encode Google Cloud domain knowledge into structured instructions for AI coding agents. The company outlined how it maintains consistency and quality as more teams contribute skills, emphasizing a standardized repository layout and preference for remote Model Context Protocol (MCP) tools.

Google Cloud has announced the general availability of Managed Lustre and C4N network and storage optimized VMs. These updates provide enhanced high-performance storage and improved network bandwidth for AI and data-intensive workloads on Google Cloud.

Gemini Robotics 2 has been released, providing robots with intelligent whole-body control, advanced dexterity, and multi-robot collaboration capabilities. This development allows robots to perform complex tasks and adapt to new robotic bodies more efficiently, addressing limitations in traditional pre-programmed or teleoperated systems.

Gemini Robotics ER 2, an "embodied reasoning" model for robotics, has been launched, offering improved video understanding, task orchestration, and multi-robot collaboration capabilities. This update allows robots to better track progress, adapt to changes, and work together on complex tasks, advancing physical AI agentic capabilities.

Google Cloud has made several capabilities of its Gemini Enterprise Agent Platform, including Agent Runtime and Agent Identity, generally available. This update allows businesses to automate long-running agentic workflows with improved memory and extended execution times, and introduces CodeMender for automated code remediation.

A tutorial details how to use Agents CLI to manage the entire development lifecycle of AI agents within Gemini Enterprise, from setup to deployment and evaluation. This approach aims to simplify the transition of AI projects from prototype to production by consolidating tools and processes.

Managed Agents in the Gemini API now default to Gemini 3.6 Flash, introduce environment hooks for tool call management, and offer explicit model selection. These updates provide developers with more control over agent behavior and model choice within the Gemini Interactions API.

Android Studio Quail 2 has redesigned Agent Mode to support multiple AI conversations, improving developer efficiency. It also integrates LeakCanary for enhanced memory leak detection and auto-synthesis of crash data to suggest fixes, streamlining the development process.

CodeMender, a managed code security agent, is now in preview, offering automated scanning and remediation of software vulnerabilities. This tool aids security teams in countering adversarial AI threats by significantly speeding up vulnerability management processes.

Google Cloud AI has introduced 13 hands-on demos for its Gemini Enterprise Agent Platform, designed to help users build and optimize AI agents. These demos exemplify different functionalities of the platform, including event-driven approval processes and integration with data sources.

Google Cloud's new Gemini Enterprise Agent Platform enables rapid AI model upgrades, reducing migration time from months to hours. This shift addresses longstanding issues faced by engineering teams in adapting to frequent model iterations.

A Russian-speaking hacker used Google's Gemini CLI AI tool as a hacking agent to control a botnet. This incident highlights vulnerabilities in open-source AI tools, potentially impacting cybersecurity across industries.

Google Cloud integrates BigQuery with the Gemini Enterprise app to aid administrators in managing large-scale deployments. This enhancement allows IT and security teams to analyze user behavior, quantify productivity, and execute compliance audits efficiently using pre-computed dashboards and detailed telemetry.

The Tambellini Group has recognized Gemini Enterprise for Education as a Commander in its StarChart for 2026 AI Agents for Administrative Efficiency. This designation highlights Gemini's integration of advanced AI capabilities, improving operational efficiency in higher education institutions.

Developers can now publish AI agents through Google Cloud Marketplace and Gemini Enterprise. The guide details steps for integration, organizational requirements, and agent architecture, emphasizing the significance of autonomy and multi-step workflows.

Gemini Enterprise Agent Platform has launched to assist IT leaders in building and deploying agents effectively. The platform centralizes the development process, enabling teams to manage risks such as data leaks and token exhaustion while fostering collaboration among various technical roles.

The Gemini API has added new features for Managed Agents, including support for background execution, remote MCP server integration, and custom function calling. These enhancements enable developers to build more efficient, production-ready AI agents by addressing previous limitations.

The Gemini Enterprise Agent Platform has introduced a remote Managed Control Plane (MCP) server to facilitate secure connection for external AI agents with Google Cloud resources. This development allows developers to create agents using their preferred IDEs while ensuring compliance and data protection.