GCP Releases
Google Cloud releases and Terraform Google provider. New features, breaking changes, security advisories and deprecations - each summarised in plain English and updated continuously.
Tracking 819 GCP releases · Updated
- Google Cloud Blog blogaiinfragcpengineer ·
GCP Introduces llm-d for Efficient RL Job Interleaving on Accelerators
Google Cloud has introduced llm-d, a new project designed to improve the efficiency of reinforcement learning (RL) jobs for large language models. By enabling co-operative time-slicing, llm-d interleaves independent RL jobs onto shared hardware, significantly increasing accelerator utilization from approximately 40% to 70% without impacting model convergence. This solution addresses critical infrastructure bottlenecks in distributed RL, benefiting teams pushing the boundaries of AI research and development by reducing wasted compute and lowering total cost of ownership. The system is designed to optimize both synchronous and asynchronous RL workloads.
feature announcement - Google Cloud Blog blogaigcpengineerhealthcaregcp-vertex-ai ·
Voicify leverages Google Cloud's Gemini and Vertex AI for AI-enabled ordering
Voicify has enhanced its AI-powered voice ordering system by integrating Google Cloud's Gemini Flash and Vertex AI, addressing challenges like transactional precision, traffic spikes, and latency. This migration provides enterprise-grade reliability and security, particularly beneficial for their healthcare and restaurant clients. The solution improves efficiency, reduces onboarding time, and ensures consistent service during peak usage.
feature announcement - Google Cloud Blog blogaigcpengineergcp-bigquerygcp-spanner ·
Google Cloud's Agentic Data Cloud for Scalable AI
Google Cloud introduced its Agentic Data Cloud at Next 2026, aiming to bridge the gap between AI model capabilities and enterprise data context. This new approach unifies data, AI models, and operational databases, addressing infrastructure limitations that hinder AI scalability and operational efficiency. It leverages native engines like BigQuery and Spanner to provide agents with seamless, low-latency access to fragmented data sources, reducing manual effort and enabling trust through enriched context and real-time data access.
feature announcement - Google Cloud Blog blogdatagcpengineergcp-cloud-storagegcp-composer ·
Checkout.com Migrates to Managed Airflow on GCP for Reliability and Cost Savings
Checkout.com has migrated its self-managed Apache Airflow to Google Cloud's Managed Service for Apache Airflow (Gen 3). This move significantly reduced operational overhead, improved reliability, and cut costs by an estimated 30% through dynamic scaling. Data engineers now benefit from faster development cycles and simplified dependency management, with new capabilities like AI-powered troubleshooting using Gemini Cloud Assist. The migration affects data engineering teams previously managing self-hosted Airflow environments.
feature announcement - Google Cloud Blog bloginfragcpgaengineergcp-cloud-storagegcp-composer ·
Checkout.com Migrates to Google Cloud Composer 3 for Improved Data Orchestration
Checkout.com migrated from a self-managed Apache Airflow environment to Google Cloud's Managed Service for Apache Airflow (Gen 3) to reduce operational overhead and enhance data pipeline reliability. This move offloads infrastructure maintenance, enabling dynamic scaling, cost reductions, and faster developer workflows, ultimately freeing up engineering time for innovation. The migration affects data engineering teams managing orchestration and pipeline development.
feature patch announcement - Google Cloud Blog blogaigcpengineergovernmentenergy ·
Google commits $40M to boost AI for scientific discovery in Genesis Mission
Google is committing $40 million in AI tokens and cloud credits to accelerate scientific research as part of the White House's Genesis Mission. This initiative provides Department of Energy National Laboratories with access to frontier AI tools like AlphaEvolve, AlphaFold 3, and Gemini for Government. The goal is to leverage AI to double the pace of American scientific discovery, with early results showing significant time savings and expanded research capabilities for scientists.
announcement feature - Google Cloud Blog bloggovernancegcpengineergovernmentgcp-bigquery ·
Google Cloud IAM Conditions for Fine-Grained Access Control
Google Cloud IAM now offers conditions to enforce the Principle of Least Privilege (PoLP) beyond resource or service binding limitations. These conditions, written in Common Expression Language (CEL), allow administrators to restrict broad IAM roles to specific operations, APIs, or even timeframes. This hardening of access management is particularly beneficial for service accounts managing resources and for controlling access to specific MCP servers, impacting engineers and architects responsible for security.
feature - Google Cloud Blog blogaigcppreviewengineer ·
AlloyDB boosts pgvector HNSW vector search speed by 4x
AlloyDB introduces columnar engine accelerated HNSW for pgvector, achieving up to 4x higher queries per second (QPS) for vector search compared to standard PostgreSQL. This enhancement is crucial for enterprise AI applications that require maximizing QPS without sacrificing search accuracy. The feature is now available in preview for demanding workloads and requires no application code changes.
feature patch - Google Cloud Blog blogsecuritygcpgapreviewengineergovernment ·
CodeMender preview: AI-powered code vulnerability scanning and remediation
Google Cloud is releasing CodeMender, a managed code security agent, in preview. This tool uses AI to scan for and automatically remediate software vulnerabilities, aiming to reduce security risks and speed up remediation. It is designed for security teams and developers working with various programming languages and integrates with CI/CD workflows. Availability is currently in preview, with wider access planned.
feature announcement - Google Cloud Blog blogaigcpgapreviewengineerhealthcaregcp-bigquerygcp-cloud-storagegcp-vertex-ai ·
BigQuery Enhances Unstructured Data Analysis with New Search Capabilities
BigQuery has launched new features to simplify the process of extracting insights from unstructured data like PDFs, audio, and images. This aims to reduce the complexity of building AI-powered search and RAG pipelines by managing embedding generation and offering enhanced search functions. These updates are particularly beneficial for enterprises managing large volumes of diverse data, such as in the healthcare sector, and are now generally available or in public preview.
feature patch - Google Cloud Blog bloginfragcpengineergcp-cloud-rungcp-cloud-sqlgcp-cloud-storagegcp-spannergcp-firestore ·
Cloud Run makes highly available, multi-region services easier
Cloud Run now offers enhanced capabilities for multi-region high availability, including readiness probes and service health aggregation. These features automate failover within seconds during regional disruptions, reducing downtime for critical applications. This update benefits developers and architects deploying resilient workloads, especially for public-facing or private network applications, and is available at no additional cost in all Cloud Run regions.
feature - Google Cloud Blog bloginfragaengineerautomotive ·
Google Cloud C4A-metal accelerates automotive cockpit development with Panasonic vSkipGen
Google Cloud's new C4A-metal bare-metal instances, powered by Axion Arm-based architecture, are now validated with Panasonic Automotive's vSkipGen virtualization platform. This integration enables automotive manufacturers to develop and test cockpit software, including Android Automotive OS, in a cloud-native environment, reducing reliance on physical prototypes and accelerating time-to-market. The solution provides high-performance graphics offloading and supports hardware-independent development for distributed teams.
feature announcement - Google Cloud Blog blogaiengineer ·
11 Principles for Token-Efficient AI Coding Assistants
This guide presents eleven principles for optimizing token consumption with AI coding assistants like Gemini, aiming to reduce latency, improve accuracy, and control costs. By adopting structured habits and efficient prompting strategies, developers can maintain a faster, more precise feedback loop. These practices are relevant for all software engineers and architects using AI-powered development tools.
announcement - Google Cloud Blog blogaigcppreviewengineerretailgovernmentgcp-bigquerygcp-cloud-rungcp-pubsub ·
Build AI Agents on GCP with Gemini Enterprise Agent Platform Demos
Google has released 13 hands-on demos for its Gemini Enterprise Agent Platform, showcasing how to build, scale, govern, and optimize AI agents. These demos utilize the code-first ADK and offer practical patterns for enterprise development, from basic agent creation to complex, long-running, and multi-agent pipelines. Developers can leverage these examples to integrate with services like BigQuery and Cloud Run, with features for security, governance, and optimization built-in.
feature announcement - Google Cloud Blog blogdatagcppreviewengineergcp-bigquery ·
BigQuery introduces IAM data governance tags for column-level security
BigQuery has released a preview of IAM data governance tags, a new method for column-level access control that improves upon existing policy tags. These tags offer global scope, automatic disaster recovery replication, hierarchical classification up to five levels deep, and decoupled governance for more flexibility. This feature is available to all BigQuery customers and is particularly beneficial for those managing complex data ecosystems and seeking to enhance their data protection strategies.
feature announcement - Google Cloud Blog blogaigcpengineerfinanceretailgovernmentautomotive ·
Google recognized as leader in Conversational AI by Gartner
Google has been named a leader in the Gartner Magic Quadrant for Conversational AI Platforms for the second consecutive year, receiving top marks for vision and execution. This recognition highlights Google's advancements in AI research and enterprise infrastructure, particularly with Gemini Enterprise for Customer Experience. The platform enables organizations to build sophisticated AI agents that can reason, act, and integrate across enterprise systems for improved customer interactions. The announcement emphasizes the capabilities of CX Agent Studio and its foundation on Google Cloud's AI stack, designed for production-ready AI deployments.
announcement - Google Cloud Blog blogaiinfragcpengineergovernmentgcp-gkegcp-cloud-storage ·
GKE Blueprint for Enterprise AI Workload Security
Google Kubernetes Engine (GKE) now offers a blueprint to secure enterprise AI workloads by integrating controls across multiple Google Cloud services and GKE features. This blueprint addresses infrastructure, model, and application security layers, aiming to protect proprietary models and comply with regulations without hindering developer velocity. It is available for platform engineering teams and CISOs managing AI deployments on GKE.
announcement feature - Google Cloud Blog blogaigcpengineer ·
Accelerate Foundation Model Upgrades with Agentic Workflows
Google Cloud's Applied ML team developed an agentic workflow to upgrade foundation models in hours, drastically reducing the months-long manual process typically required. This approach, outlined in a blog post, leverages Gemini Enterprise Agent Platform and Google Antigravity to offer flexibility and adaptability over rigid automation. Engineering teams managing AI model updates can apply these lessons to improve their own migration pipelines, impacting developers and architects working with AI models.
announcement feature - Google Cloud Blog blogaisecuritygcpengineergovernmentgcp-gke ·
Google Cloud AI Threat Defense Combats AI-Driven Cyberattacks
Google Cloud introduces AI Threat Defense, a unified platform leveraging Gemini, Wiz, CodeMender, and Mandiant to counter AI-powered cyberattacks. This new framework aims to provide defenders with a speed and context advantage over attackers who are increasingly using AI for zero-day exploits and rapid multi-agent attacks. The platform assists organizations in preparing, scanning, prioritizing, remediating, and monitoring threats across their software development lifecycle. It's designed for security engineers and architects managing cloud environments.
feature announcement - Google Cloud Blog blogaigcpengineermedia ·
Google's AI agents collaborate to create short films
Google experimented with teams of AI agents collaborating to produce short films, uncovering insights into inter-agent teamwork in creative domains. Agents with distinct roles like Idea Person, Technical Lead, and Editor worked within Scion, an open-source orchestration testbed. This hackathon demonstrated how AI agents can manage complex, multi-step projects with verification gates, producing over 25 films using models like Gemini and Veo.
announcement feature
About GCP release tracking on ReleaseBytes
Google Cloud publishes release notes per product — BigQuery, GKE, Cloud Run, Cloud SQL, Vertex AI and dozens more — which makes platform-wide awareness hard. ReleaseBytes aggregates the official GCP release notes and the Terraform Google provider changelog into a single feed, with a plain-English summary of each update and tags for breaking changes, deprecations and security fixes.
Frequently asked questions
How often are GCP release notes updated on ReleaseBytes? ›
Continuously. ReleaseBytes monitors the official GCP release channels around the clock and publishes a plain-English summary of each announcement shortly after it lands.
What kinds of GCP changes does ReleaseBytes track? ›
New features, enhancements, bug fixes, security advisories, breaking changes, deprecations and end-of-life announcements. Every item is tagged by type so you can filter to just the changes that need action.
How can I get alerts for new GCP releases? ›
Set up a free email or Slack alert filtered to GCP, subscribe to the weekly digest, or follow the RSS feed. Teams can also install the ReleaseBytes GitHub App or connect via MCP.
Where does the GCP release data come from? ›
From the official sources: Google Cloud releases and Terraform Google provider. Every item links back to the original vendor announcement.