The Cloud Hub by Digigen | Google Cloud & Microsoft Insights for Thailand

Google Cloud Next '26 Recap: Thailand & Singapore Lens

Written by Danai Boonsri | Apr 28, 2026 3:45:17 AM

Digigen Insights

Google Cloud Next '26: The Digigen Lens

All 260 announcements, organized and translated for businesses in Thailand and Singapore.

Google Cloud Next '26 wrapped in Las Vegas this week with 32,000 attendees and 260 announcements. This piece does both: a Digigen lens at the top and bottom for the strategic shifts, and the full numbered catalogue of all 260 announcements in the middle so you can scan or search for what matters in your stack.

260 Announcements
32K Attendees
$750M Partner Innovation Fund
1,302 AI Customer Stories

The Real Story: Five Shifts That Matter

Strip away the marketing, and Next '26 is really about six structural shifts. Every conversation we will have with a Thai or Singapore client over the next twelve months traces back to one of these.

1

Agents Become the Operating Layer

Vertex AI is now the Gemini Enterprise Agent Platform. The pitch is no longer "use AI." It is "delegate business outcomes to a governed fleet of agents." Identity, registry, gateway, simulation, evaluation, optimizer. This is enterprise software, not a chatbot.

2

The App Layer Goes Horizontal

Gemini Enterprise is now a horizontal app for every employee, with Canvas, Projects, Skills, Inbox for agents, and direct interoperability with Microsoft 365. This is the moment Google Workspace genuinely competes with Microsoft Copilot at the platform level.

3

Compute Stops Being a Bottleneck

The 8th gen TPUs (8t and 8i), Axion CPUs, A5X with NVIDIA Vera Rubin, and Confidential VMs reset the price-performance ceiling. For Thai and Singapore enterprises that have parked AI projects on cost concerns, the math just changed.

4

Data Becomes Agent-Ready

The Agentic Data Cloud, Knowledge Catalog, and zero-copy federation across Salesforce, SAP, ServiceNow, and Workday solve the dirty truth of enterprise AI: your agents are only as smart as your data plumbing. This is the unglamorous work that finally has tooling.

5

Security Catches Up to Agents

Wiz is now Google Cloud's. Combined with Agent Identity, Model Armor, Agent Anomaly Detection, and the Wiz AI Bill of Materials, there is finally a defensible security story for enterprises that need to put agents in production. CISOs in Singapore, especially, will notice.

6

The Partner Economy Just Tripled

The new Agent Marketplace in Gemini Enterprise lets partners sell agents directly to enterprise users. The $750M innovation fund and forward-deployed engineer partnerships with Accenture, Deloitte, and McKinsey signal that Google is betting on partners to deliver, not just resell.

How to read the catalogue: Each numbered item is a direct mapping to one of Google's 260 announcements, grouped by the same product domains Google used. Skim by section header, or use Ctrl/Cmd+F for a specific product name.

AI Models, Platforms & Agents

Announcements 1 to 28

Gemini Enterprise Agent Platform

  1. Gemini Enterprise Agent Platform. Evolution of Vertex AI with model selection, model building, agent building, plus new integration, DevOps, orchestration and security features.
  2. Agent Development Kit (ADK). Graph-based framework that organises agents into a network of sub-agents for complex reasoning.
  3. Agent Studio. Move from simple prompts to deployed agents, with one-click export to ADK for full-code customisation.
  4. Agent Runtime. Sub-second cold starts and seconds-fast agent provisioning.
  5. Agent Sandbox. Hardened environment for safely executing model-generated code and browser-based automation, now broadly available.
  6. Agent Memory Bank. Long-term memory with Memory Profiles for high-accuracy, low-latency recall.
  7. Agent Sessions. Custom Session IDs that map directly to your internal database and CRM records.
  8. Agent Identity. Verifiable cryptographic ID per agent with full auditable trail mapped to authorization policies.
  9. Agent Registry. Central library that indexes every approved internal agent, tool and skill.
  10. Agent Gateway. Single control point for the agent fleet with Model Armor protections against prompt injection and data leakage.
  11. Agent Anomaly Detection. Real-time detection of suspicious agent behaviour using statistical models and LLM-as-judge.
  12. Agent Security dashboard. Powered by Security Command Center for unified threat detection, asset discovery and vulnerability scans.
  13. Agent Simulation. Simulates realistic conversations against synthetic users and virtualized tools, scoring on success and safety.
  14. Agent Evaluation. Continuous live-traffic scoring with multi-turn autoraters and turnkey Agent Observability dashboards.
  15. Agent Optimizer. Clusters real-world failures and suggests refined system instructions automatically.
  16. Long-running agents. Autonomous execution of complex business processes inside secure cloud sandboxes.

Gemini Enterprise App

  1. Agent Designer. No-code builder for schedule- or trigger-based agents using any enterprise connector, with a visual flowchart.
  2. Inbox in Gemini Enterprise. Central place to monitor agent activity with categorized notifications: Needs your input, Errors, Completed.
  3. Projects. Dedicated workspace where agent memory is scoped to project files and conversations, with Drive, NotebookLM and Group Chat connections.
  4. Skills. @-mention shortcuts for repetitive tasks like brand guidelines, report formatting, or data lookups.
  5. Canvas. Interactive editor inside Gemini Enterprise for Docs and Slides, with Microsoft 365 export.
  6. Enhanced Deep Research agent. Multi-step research synthesizing the open web with internal company data.
  7. Dynamic agent UI experiences. Agent-to-UI protocol (A2UI) support for rich, native UI generated inside the Gemini Enterprise app.
  8. High-code agents. Engineering teams build full-code agents and publish them to the Gemini Enterprise app.
  9. Data Insights agent. Surfaces deep patterns across the entire data landscape.
  10. Bring Your Own MCP integrations. Securely connect Gemini Enterprise to custom and third-party tools.
  11. Cross-platform unified intelligence. Unified search and reasoning across Workspace, Microsoft 365 and other apps.
  12. Agent Gallery. Vetted third-party agents from Accenture, Adobe, Atlassian, Deloitte, Lovable, Oracle, Palo Alto Networks, Replit, S&P Global, Salesforce, ServiceNow, Workday and more.

AI & Compute Infrastructure

Announcements 29 to 47

8th Generation TPUs & Distributed Cloud

  1. TPU 8t. Training powerhouse with nearly 3x compute performance over the previous generation.
  2. TPU 8i. Inference and reinforcement-learning chip with up to 80% better performance-per-dollar.
  3. Native PyTorch support for TPU (TorchTPU). Run PyTorch models on TPU with full Eager Mode support, in preview.
  4. NVIDIA Blackwell on Google Distributed Cloud. NVIDIA HGX B200 and B300 with fifth-gen NVLink available for sovereign deployments.
  5. New GDC machine families. A4 for inference, M2 and M3 for memory-heavy workloads, in addition to N2 and N3.
  6. Enhanced GDC storage. 6PB object storage per zone (6x previous) and 30 IOPS/GB per zone (10x boost).

Compute Engine & Storage

  1. A5X with NVIDIA Vera Rubin. NVL72 platform with concepts from open-source Falcon networking, later this year.
  2. Google Axion N4A. Custom Arm-based CPU with up to 2x better price-performance than comparable x86 VMs, now GA.
  3. Axion N4A supports GKE Agent Sandbox. Up to 30% better price-performance than the next leading hyperscaler, GA.
  4. Axion C4A.metal. First Axion bare-metal instance for Android, automotive simulation, CI/CD, security and custom hypervisors, in preview.
  5. C4 with Intel Xeon 6 (Granite Rapids). Intel AMX and native FP16 across all C4 shapes.
  6. Flexible CUDs for more VM families. Now covering M1 to M4, H3, H4D, plus Cloud Run.
  7. C4N VMs. Nearly 400 Gbps VM-to-VM bandwidth and 25 GiB/s block throughput, in preview.
  8. M4N instances. 26.57 GiB RAM per vCPU for memory-intensive databases, preview.
  9. Z4D instances. Up to 84 TiB local SSD for SQL, NoSQL and vector workloads, preview soon.
  10. Hyperdisk Balanced HA. 4x performance improvement for HA databases like SQL Server and PostgreSQL.
  11. Hyperdisk ML. 2 TiB/s aggregate throughput, more than 200x per-disk versus competitors.
  12. Hyperdisk Exapools. Highest aggregate block storage performance and capacity per AI cluster of any hyperscaler, now GA.
  13. Z4M. Up to 168 TiB local SSD, 400 Gbps networking and RDMA support, preview Q3 2026.

Agentic Data Cloud

Announcements 48 to 106

Knowledge Catalog & Governance

  1. Knowledge Catalog. Universal context engine that maps and infers business meaning across the data estate.
  2. Deep Research Agent. Synthesises internal data, docs and web research with the Knowledge Catalog.
  3. Zero-copy federation. Direct data access across Palantir, Salesforce Data360, SAP, ServiceNow, Workday.
  4. LookML Agent. Reads strategy docs, sheets and reports to generate consistent business-ready semantics.
  5. BigQuery measures. Programmatic business logic embedded directly in the SQL engine.
  6. Data Products. Self-contained data assets with intent, SLAs and governance baked in.
  7. Reusable data quality rules. Expert-defined test logic shared across multiple assets.
  8. Structured approval workflows. Domain-expert validation for new business terms and data products.
  9. Column-level lineage. End-to-end lineage now includes Looker, plus column-level for BigQuery and Spark.

Cross-Cloud Lakehouse

  1. Cross-cloud Lakehouse (formerly BigLake). Borderless lakehouse with Iceberg REST Catalog across AWS, Azure and partners.
  2. Lakehouse catalog federation. Discovery, analysis and zero-copy sharing across AWS Glue, Databricks, Snowflake and SAP.
  3. Lightning Engine for Apache Spark. Up to 2x price-performance versus proprietary alternatives.
  4. Lakehouse federation for AlloyDB. Live joins between transactional and historical analytical data without ETL.
  5. Managed Iceberg storage and REST Catalog. Automatic management, multi-table transactions, read/write across BigQuery, Spark and OSS engines.
  6. Cross-cloud caching. Slashes egress fees and accelerates AWS and Azure data queries.

Data Practitioner Tooling

  1. Google Cloud Data Agent Kit. Skills, MCP tools and plugins for VS Code and Gemini CLI.
  2. Data Science Agent. Plain-English plans for loading, cleaning and visualizing data in BigQuery notebooks.
  3. Data Engineering Agent. Autonomously builds Dataform pipelines with governance and testing.
  4. Database onboarding and observability agents. Provisioning, monitoring and automated troubleshooting across the database fleet.
  5. Conversational Analytics. Natural-language queries publishable as agents into Gemini Enterprise.

Database Portfolio

  1. Spanner Omni. Globally consistent, multi-model database that runs anywhere: multi-cloud, on-prem, even on a laptop.
  2. Spanner Columnar Engine. Up to 200x faster analytical queries on live operational data.
  3. Managed remote MCP servers for databases. Direct connection between AI models and operational data.
  4. Vibe coding with AI Studio. Generate live apps from text prompts that connect to trusted databases.
  5. Tools for Data Agents. Modular ready-to-use functions including high-accuracy text-to-SQL.
  6. Bigtable in-memory. Sub-millisecond reads via hybrid in-memory storage.
  7. Memorystore for Valkey 9.0. Managed migration from self-managed Redis with ACLs.
  8. AlloyDB AI search at scale. ScaNN index for 10 billion vectors with 6x faster queries than standard PostgreSQL.
  9. New AI functions for AlloyDB. ai.if, ai.analyze_sentiment and ai.summarize embed Gemini reasoning into SQL.
  10. Firestore full-text and geospatial search. Google search technology natively in Firestore.
  11. Compute Engine to managed migrations. Automated networking and replication to move PostgreSQL to Cloud SQL or AlloyDB.
  12. Database Center with BigQuery. Single management plane across transactional and analytical with Gemini fleet analytics.

BigQuery

  1. Reverse ETL for BigQuery. One-click sync from lakehouse to AlloyDB or Spanner for low-latency serving.
  2. BigQuery Graph. Maps entities, relationships and business logic for grounded agentic reasoning.
  3. BigQuery fluid scaling. Per-second billing for variable agentic workloads.
  4. ObjectRef. Process unstructured and structured data together via SQL and Python.
  5. AI.PARSE_DOCUMENT. Single SQL function for OCR, layout parsing and chunking with Gemini.
  6. TabularFM model. Zero-shot regression and classification natively in BigQuery.
  7. Optimized mode for SQL-first AI. Up to 230x reduction in token consumption for AI.CLASSIFY and AI.IF.
  8. BigQuery-native Gemma embeddings. High-quality vector embeddings on standard CPUs.
  9. Autonomous embedding generation. Fully managed pipelines that keep vector indexes in sync.
  10. BigQuery hybrid search. Single function for combined semantic and full-text retrieval.
  11. Python UDF. Managed Python scalar functions that autoscale to millions of rows.
  12. Connected Sheets with TimesFM. Zero-shot forecasting and anomaly detection in Google Sheets.
  13. Geospatial analytics datasets. Google Earth AI and roads data joined to enterprise data.
  14. Proactive agentic workflows. Metric shift detection, root-cause analysis and scheduled briefings.
  15. BigQuery Agent Analytics plugins. Logs from ADK and LangGraph for audit and optimization.
  16. BigQuery Studio productivity tools. Contextual assistant, SQL and Visualization Cells, Files Explorer, Git workflows.
  17. Runtime, small query and history-based optimizations. 35% YoY performance gain with lower processing costs.
  18. New BigQuery workload management. Reservation groups, dynamic assignments, project-level slot controls.
  19. Colab Data Apps. Notebook analyses turned into shareable Python apps in Data Studio.

Looker & Business Intelligence

  1. Data Studio (formerly Looker Studio). Free and Pro versions with a unified browse interface for sources, apps and agents.
  2. Looker Dashboard Agents. In-dashboard natural-language Q&A with governed answers.
  3. Agentic workflows in Looker. Proactive metric monitoring and correlation discovery.
  4. Embedded conversational experiences in Looker. Gemini-powered multi-turn experiences via API and SDK.
  5. Looker self-service explores. Personal files blended with enterprise data via auto-generated LookML.
  6. Looker AI assistants. Visualization, Expression and Insight assistants for natural-language BI.
  7. Open BI and MCP for Looker. Open-source MCP Toolbox plus a Google-managed MCP server native to Looker.
  8. Continuous Integration in Looker. Automated SQL and content testing in development workflows.

Application Development & Runtimes

Announcements 107 to 139

Cloud Run

  1. Full-stack apps from AI Studio. Vibe-coded apps with Firestore and authentication, GA on Cloud Run.
  2. Cloud Run integration with Agent Platform. Move agents from experiment to managed production, preview.
  3. Fully managed remote MCP server. Easy deployments for developers and agents, GA.
  4. NVIDIA RTX PRO 6000 Blackwell GPUs on Cloud Run. 70B+ parameter inference without managing GPUs, scales to zero.
  5. Cloud Run billing caps. Maximum monthly spend before resource deactivation, coming soon.
  6. Cloud Run instances. Long-running background jobs, preview.
  7. Cloud Run sandboxes. Ephemeral sandbox tool isolated from agent code, coming soon.
  8. SSH for Cloud Run. Secure shell into running Cloud Run containers, preview.
  9. Cloud Run service bindings. Service-to-service comms for microservices, coming soon.
  10. Ephemeral disk for Cloud Run. Disk created at start, deleted at stop, preview.

Firebase

  1. Firebase SQL Connect. Data Connect renamed and connecting clients to Cloud SQL for PostgreSQL with realtime sync and offline cache.
  2. Firebase integration with Application Design Center. Unified deployment for Firestore, App Check, Authentication and Firebase AI Logic.
  3. Firestore Enterprise edition. Native full-text search, geospatial, JOIN via subqueries and DML.
  4. Firebase AI Logic. Generative AI in mobile and web apps with auto function calling and JSON schema mapping.
  5. Firebase Phone Number Verification. Carrier-based number verification across 10+ carriers in 6 regions.

Google Kubernetes Engine

  1. GKE Agent Sandbox. Add-on for isolated, stateful, single-replica AI agent runtimes.
  2. GKE hypercluster. Single GKE control plane managing millions of accelerators across regions, private GA.
  3. Predictive latency boost in GKE Inference Gateway. Up to 70% lower time-to-first-token with capacity-aware routing.
  4. Automatic KV Cache storage tiering. Across RAM, Local SSD and Cloud Storage/Lustre for long-context workloads.
  5. RL Scheduler. Removes the straggler effect and inter-batch tail latency.
  6. RL Sandbox. Kernel-level isolation for tool calls and reward evaluation with millisecond provisioning.
  7. RL Observability and Reliability dashboards. Out-of-the-box visibility for the full RL loop.
  8. Scaling on custom metrics. Intent-based autoscaling beyond CPU and memory.
  9. Accelerated node and pod startup. 4x faster nodes, 80% faster pods.
  10. Rapid model loading. 5x faster model loads via run:AI Model Streamer and Rapid Cache.

Gemini Cloud Assist & Operations

  1. Redesigned Application Design Center. Natural-language design to deployment for multi-resource environments.
  2. gcloud, kubectl and Terraform support. Multi-turn agents for infra automation and incident resolution.
  3. Proactive cost anomaly detection. FinOps agent for spending spikes and granular cost reports.
  4. Gemini Cloud Assist MCP servers. Help inside IDEs, CLI and third-party tools.
  5. App-centric Unified Maintenance. Consolidated maintenance view at the application level.
  6. ALM feature flags. Decouple feature releases from deployments with safe rollouts.
  7. FinOps Explainability agent. Investigates AI-related cloud cost drivers.
  8. Spend Caps. Project-level budget enforcement for AI Studio, Agent Platform, Cloud Run, Cloud Run Functions and Maps.

Customer Stories

Announcements 140 to 154

  1. Capcom. AI agents for visual inspection, predictive testing and institutional knowledge in playtesting.
  2. Citi Wealth. Citi Sky, an always-on multilingual agent for market insights and advisor engagement.
  3. Citadel Securities. TPU-based research with AI workloads up to 4x faster at 30% lower cost.
  4. Deutsche Telekom. MINDR multi-agent system on Gemini for self-healing networks, cutting event management by 95%.
  5. GE Appliances. 800+ Gemini Enterprise agents across manufacturing, logistics and supply chain.
  6. Highmark Health. Sidekick AI assistant delivered $27.9M in 2025 value.
  7. Home Depot. Magic Apron customer-experience agent and a phone agent that identifies needs in under 10 seconds.
  8. Macy's. Ask Macy's, a multimodal shopping concierge with virtual try-on.
  9. Mars. Gemini Enterprise as the primary AI operating system for the global workforce.
  10. Merck. Agentic platform across R&D, manufacturing, commercial and corporate ops for 75,000 employees.
  11. Papa John's. First user of the Food Ordering agent in Gemini Enterprise for Customer Experience.
  12. Tata Steel. 300+ specialized AI agents deployed in nine months across global operations.
  13. Unilever. Multi-agentic procurement system for faster, smarter buying decisions.
  14. Virgin Voyages. Rovey concierge agent on Gemini Enterprise and Google Distributed Cloud, even at sea.
  15. Vodafone Business. Managed detection and response on Google Security Operations and AI Concierge with Gemini.

Gemini Enterprise for Customer Experience

  1. Shopping agent and Food Ordering agent. Conversational sales and ordering across web, mobile and third-party channels.
  2. Omnichannel Gateway. Cross-channel context across web, mobile and voice.
  3. Agent Assist. Real-time coaching for service reps with grounded answers.

Networking

Announcements 158 to 195

Cross-Cloud Network

  1. Virgo Network. Scale-out fabric for AI workloads with deterministic low latency.
  2. Agent Gateway partnerships. Broadcom, Check Point, Cisco, CrowdStrike, Exabeam, F5, Netskope, Okta, Palo Alto Networks, Ping Identity, Saviynt, Silverfort, Thales, Zscaler.
  3. Ambient networking. Sidecar-free service discovery, zero-trust access and traffic management for GKE and Cloud Run.
  4. Service bindings. Automatic service-to-service connectivity for agentic apps.
  5. Network Services Monitoring. Bridges application and network observability.
  6. Multi-region GKE Inference Gateway. Cross-regional failover and reduced global latency, preview.
  7. GKE Inference Gateway disaggregated serving. vLLM and SGLang via llm-d, GA.
  8. Accelerator network profiles. Up to 60% bandwidth uplift for distributed AI/ML, GA.
  9. Cloud Interconnect 400 Gbps circuits. Up to 3.2 Tbps in a single connection, GA.
  10. Partner Cross-Cloud Interconnect. AWS GA; CoreWeave and Lumen preview soon.
  11. Ultra Low Latency Solution for financial exchanges. CME Group migration to Google Cloud, preview.
  12. Cloud Network Insights. NPM and DEM via Broadcom AppNeta, preview.
  13. Private Service Connect published services. 40+ Google and third-party services with secure global access.
  14. Private Service Connect endpoint security. Granular authorization between producer and consumer, preview.
  15. Cloud Number Registry. Agentic IPAM with Infoblox Universal DDI integration, preview.
  16. Hybrid Subnets. Migrate workloads to VPC without changing IP addresses, GA.
  17. Cloud NAT IPv6 to IPv4. DNS64 plus private NAT64 for hybrid stacks, preview soon.
  18. Granular networking charge visibility. Cost Explorer and App Optimize API attribution, preview.
  19. Gemini Cloud Assist network security agent. Policy generation, recommendations and impact analysis.
  20. Gemini Cloud Assist network agent. Workload placement, PSC troubleshooting and observability cost estimation.

Global Front End, Cloud WAN & Network Security

  1. Network observability MCP tools and agent skills. Partners build agents on connectivity tests and natural-language VPC Flow Logs.
  2. Global Front End Enterprise. Global Cloud Load Balancing, Cloud Armor, Cloud CDN and Service Extensions with up to 15% lower TCO.
  3. Post-quantum cryptography (PQC). Industry-standard PQC algorithms protecting workloads.
  4. Google tag gateway. Serve advertising tags from advertiser domains for better measurement, GA soon.
  5. Image optimization in Cloud CDN. Best-fit content delivery with bandwidth savings, preview.
  6. Cloud CDN GKE Gateway support. Caching managed via GKE APIs, GA.
  7. Cloud WAN expanded reach. Site-to-site data transfer in over 25 countries.
  8. NCC Gateway. Third-party SSE integrations with Palo Alto Networks (GA soon) and Symantec (preview).
  9. Verified Peering Provider expansion. 175+ providers worldwide.
  10. Last mile connectivity. Site-to-cloud private connectivity in minutes from the console, preview soon.
  11. Cloud Armor managed rules. Powered by Thales Imperva for L7 attack and zero-day detection.
  12. Cloud Armor adaptive protection for NLB and VMs. ML-based volumetric DDoS protection at L3/L4.
  13. Cloud Armor visual rule builder. Simplified custom rule UX, preview soon.
  14. Cloud NGFW Internal Application support. Service-centric security for GKE, Cloud Run and PSC, preview.
  15. Cloud NGFW project-level policies. Endpoints, security profiles and groups at project level.
  16. Cloud NGFW Advanced Malware Sandbox. Powered by Palo Alto Networks Advanced Wildfire, blocking 99% of known and unknown malware.
  17. NCC static routes via internal load balancer. Integration with Secure Web Proxy and third-party security appliances, GA.
  18. NCC support for privately used public IPs (PUPI). PUPI IPv4 exchange with VPC and producer VPC spokes, GA.

Partners

Announcements 196 to 205

  1. $750M innovation fund. New models, tools, engineering support and incentives for partner-led agent development.
  2. Sell partner agents in Gemini Enterprise. New Agent Marketplace inside Gemini Enterprise.
  3. 70+ partner-built agents at launch. Including Accenture, Adobe, Atlassian, Deloitte and more.
  4. Forward deployed engineer partnerships. Google Cloud engineers available via Accenture, Deloitte and McKinsey.
  5. Deeper model integration with Palantir. Joins Box, DocuSign, Oracle, Salesforce and SAP with Gemini.
  6. AI-powered Google Cloud Partner Network. Partner agent and agentic Earnings Hub and Partner Finder.
  7. SAP Unified Data Foundation. SAP and non-SAP data unified with bi-directional zero-copy sharing.
  8. Additional SAP news. Gemini in SAP Engagement Cloud, Cortex Framework accelerators, SAP Sovereign support.
  9. Oracle AI Database@Google Cloud expansion. 20 regions, GoldenGate near-real-time replication and Knowledge Catalog integration.
  10. Partners of the Year. Hundreds celebrated across specialties, industries and geographies.

Security

Announcements 206 to 225

Google Security Operations & Platform Security

  1. Threat hunting agent. Proactive hunt for novel attack patterns and stealthy adversary behaviour.
  2. Detection Engineering agent. Identifies coverage gaps and creates new detections automatically.
  3. Third-Party Context agent. Enriches workflows with external content.
  4. Google Security Operations partner integrations. Including Darktrace, Gigamon and SAP.
  5. Model Armor. Integrated with Agent Gateway, Agent Runtime, Langchain and Firebase for inline enforcement and sanitization.
  6. Google Cloud Fraud Defense. reCAPTCHA evolves into a platform for securing the agentic web, GA.
  7. Chrome Enterprise AI-aware extension threat detections. Surface advanced extension telemetry, preview.
  8. Chrome Enterprise shadow AI reporting. Visibility into unsanctioned AI and SaaS use, GA soon.
  9. Identity and Access Management updates. Streamlined predefined roles, IAM role picker and re-authentication for sensitive actions.
  10. Confidential Computing for G4 VMs. NVIDIA RTX PRO 6000 Blackwell on Confidential G4 VMs, preview globally.
  11. C4 Confidential VMs. Intel TDX on 6th Gen Xeon for AI and analytics, preview.
  12. Confidential External Key Manager (cEKM). Host external keys in any region with verifiable control, preview.
  13. KMS Quantum Safe Key Imports. Bring your own quantum-safe keys, preview.
  14. Secret Manager with ADK. Native integration to prevent password leaks and prompt injection, GA.

Wiz (now Google Cloud)

  1. Agent studio support. Databricks, AWS Agentcore, Gemini Enterprise Agent Platform, Azure Copilot Studio and Salesforce Agentforce.
  2. Wiz Defend with Google Security Operations. Easier automatic threat-info forwarding to Mandiant Threat Defense.
  3. Wiz vibe coding integration. Security scanning inside Lovable with vulnerabilities surfaced in the built-in security view, GA in May.
  4. Inline AI security hooks. Wiz inside IDEs and agent workflows for prompt and output evaluation.
  5. Wiz Skills. Code-to-cloud context and validated findings for agent-based remediation.
  6. AI-Bill of Materials. Inventories all AI frameworks, models and IDE extensions across the environment.

Storage

Announcements 226 to 240

  1. Rapid Bucket. 15+ TB/s, 20M requests per second, sub-millisecond latency in a zonal bucket, GA.
  2. Rapid Cache. Anywhere Cache renamed, 2.5 TB/s aggregate read throughput, no code changes.
  3. Rapid Cache ingest-on-write. Up to 2.2x faster checkpoint restores.
  4. Managed Lustre throughput. Up to 10 TB/s, 2.6x faster checkpoint writes and restores.
  5. Managed Lustre Dynamic tier. Low-latency AI training storage at $0.06/GB-month.
  6. Managed Lustre with TPUDirect and RDMA. Data bypasses host directly to accelerators.
  7. Smart storage object context. Mutable IAM-governed metadata per object, GA.
  8. Smart storage automated annotations. Auto-metadata for unstructured data.
  9. Cloud Storage MCP server. Read, write and analyze Cloud Storage via MCP.
  10. Storage Intelligence dashboards. Cost anomalies and DSPM integration with Security Command Center, preview.
  11. Storage Intelligence object events and bucket activity. New Insights Datasets tables for placement and 429 troubleshooting.
  12. Enhanced batch operations. Change ACL and storage class operations, multi-bucket support.
  13. Google Cloud NetApp Volumes. Flex Unified for block and file on the same pool, ONTAP-mode for existing automation.
  14. Filestore for GKE. Shares from 100 GiB with independent capacity and IOPS scaling.
  15. Backup and DR agentic AI. Auto-audit backup estate and remediate gaps, with AlloyDB and Filestore integrations.

Startups

Announcements 241 to 245

  1. Notable startups using Google Cloud. Lovable, Notion, Proximal Health, Gamma, Mavvrik, Insilica, Parrallel and more.
  2. NVIDIA hardware via Google Cloud. Thinking Machines, Interfact, Nucleus AI, Reflection, Neuraldefend.
  3. Global Gemini Startup Forum. Two-day June gathering for seed-to-Series A founders.
  4. Google for Startups AI Agents Challenge. Six-week global competition with $500 in credits and a $90,000 prize pool.
  5. Future of AI generative media report. Interviews with eight startup founders, investors and Googlers.

Workspace

Announcements 246 to 260

  1. Workspace Intelligence. Real-time semantic understanding across Workspace apps, projects and collaborators.
  2. Ask Gemini in Chat. Unified command line for tasks, documents and file lookup.
  3. Gemini in Sheets, Docs and Slides. Build and edit spreadsheets, infographics and full editable decks via natural language.
  4. AI Inbox and AI Overviews in Gmail. Streamlined inbox triage and search-driven answers.
  5. Gemini in Drive. AI Overviews, Ask Gemini and Drive Projects for organising files, emails and context.
  6. Third-party data in Sheets. Import HubSpot and Salesforce, build dashboards and kanban boards in Sheets canvas.
  7. Workspace Skills. Build, share and invoke agentic skills via Workspace Studio.
  8. Custom Avatars in Vids. Branded avatars and presentations-to-video conversion.
  9. Take Notes for Me everywhere. Automated meeting summaries and action items including Zoom and Teams.
  10. Workspace in Gemini Enterprise app. Calendar scheduling, Docs and Slides creation directly in canvas mode, private preview.
  11. Auto browse with Gemini in Chrome. US Workspace customers can enable in Chrome Enterprise.
  12. Workspace MCP Server. Synthesize Drive docs and draft Gmail responses via developer MCP, preview.
  13. Simplified agent governance. AI control center, agent management and Workspace Studio controls for monitoring and audit.
  14. New sovereign controls and client-side encryption. Lock data processing to US or EU, with Germany and India coming.
  15. Improved migration and interoperability. New cloud service to migrate full organizations from Microsoft 365 to Workspace, plus Office app collaboration.

What Digigen Is Watching Closely

Of the 260 announcements, these are the ones that change how we advise clients and structure deals over the next two quarters.

Workspace The M365 Migration Path Just Got Easier

Announcement #260 is quiet but consequential. Google is launching a managed cloud service for migrating entire organizations from Microsoft 365 to Workspace, including emails, files, and conversations, plus improved interoperability with Office. For Digigen, which sits dual-certified across both stacks, this changes the conversation with clients evaluating cost, data sovereignty, or vendor consolidation. The Microsoft-to-Google migration was always technically possible. Now it is operationally serviceable.

Workspace Gemini in Workspace Closes the Copilot Gap

The new Gemini capabilities in Sheets, Docs, Slides, and the AI Inbox in Gmail materially close the gap with Microsoft Copilot. Combined with the new Workspace Skills framework, Workspace becomes a genuine agent platform, not just a productivity suite with a chatbot bolted on.

Agents Agent Platform Is the New Vertex AI

For our Agentiv consulting work, the Agent Platform stack is the new reference architecture: ADK for orchestration, Agent Studio for prototyping, Agent Runtime for production, Agent Sandbox for safe execution, Memory Bank for long-term context, Identity for governance, and Anomaly Detection for SecOps. This is what we will build on, full stop.

Data The Knowledge Catalog Is the Quiet Game Changer

Most clients in Thailand and Singapore have a data estate that is fragmented across BigQuery, on-prem databases, SAP, Salesforce, and a dozen SaaS tools. The Knowledge Catalog plus zero-copy federation means we can finally build agents that reason across this estate without a six-month data engineering project upfront. Combined with Spanner Omni, the architecture options for regulated industries open up significantly.

Infra Cloud Run Becomes a Real Application Platform

Cloud Run now supports full-stack apps from Google AI Studio, NVIDIA RTX PRO 6000 Blackwell GPUs for 70B+ model inference, billing caps, instances for long-running jobs, sandboxes, SSH, and ephemeral disk. For SMB and mid-market clients in Bangkok who do not want to manage Kubernetes, Cloud Run just became the default landing zone for production AI workloads.

Security Wiz Integration Closes the AI Security Gap

The Wiz acquisition is now complete and the integrations with Gemini Enterprise Agent Platform, AWS Agentcore, Azure Copilot Studio, and Salesforce Agentforce mean Wiz is positioned as the multi-cloud AI security graph. The new AI Bill of Materials is exactly the artifact every Singapore CISO has been asking for under MAS TRM and the upcoming EU AI Act spillover requirements.

Practical note from Digigen: If you are already on Vertex AI, your migration to Agent Platform is largely a rename plus opt-in to new features. If you have been waiting to start, this is the moment to start with Agent Studio rather than building from scratch.

What We Are Telling Clients This Week

  • If you are evaluating Microsoft 365 versus Google Workspace, the new migration service and Gemini capabilities materially change the comparison. Re-run the analysis.
  • If you have a stalled AI pilot, the bottleneck was probably governance, identity, or data plumbing. Agent Platform plus Knowledge Catalog now address all three.
  • If you are running production agents anywhere, including on AWS or Azure, the Wiz AI Bill of Materials should be on your security roadmap by Q3.
  • If you are cost-sensitive, TPU 8i delivers up to 80% better performance-per-dollar for inference and Cloud Run billing caps finally let you cap exposure on agentic workloads.
  • If you are a regulated business, Spanner Omni, Confidential VMs, the Confidential External Key Manager, and post-quantum cryptography in KMS open architectures that were not previously viable.
🇹🇭

What's In It for Thailand

Thailand's enterprise AI conversation has been dominated by two questions: cost and capability. Next '26 answers both.

  • Microsoft Thailand cloud region momentum meets a real Google response. With TPU 8i pricing, Axion N4A on GCP, and Cloud Run billing caps, the unit economics for Thai SMB and mid-market AI workloads finally make sense without enterprise budgets. Expect deal cycles to shorten.
  • Thai-language workflows benefit immediately. Gemini's continued strength in Southeast Asian languages, combined with Workspace Skills and the Agent Designer no-code builder, means our Thai clients can deploy genuinely useful agents without waiting for English-first features to be localized.
  • Manufacturing, hospitality, and retail get reference architectures. The Tata Steel, GE Appliances, Home Depot, and Macy's case studies map cleanly onto Thailand's industrial estates, hotel groups, and retail chains. We now have credible playbooks to point at.
  • Data residency conversations get easier. Spanner Omni and Google Distributed Cloud with NVIDIA Blackwell mean Thai banks, hospitals, and government-linked entities can run modern agentic workloads on infrastructure they control, while still using the latest models.
  • Workspace migration paths open up for Thai conglomerates. The new M365-to-Workspace managed migration service is exactly the lever we need for the largest Thai enterprises that have been quietly unhappy with Microsoft pricing and licensing complexity.
🇸🇬

What's In It for Singapore

Singapore's market is governance-led. Every CIO and CISO is reading these announcements through the lens of MAS TRM, PDPA, the AI Verify framework, and increasingly, EU AI Act compatibility. Next '26 was unusually well-aligned with that posture.

  • The Wiz acquisition closing matters most here. The AI Bill of Materials, agent studio coverage across AWS, Azure, and Salesforce, and the integration with Google Security Operations give Singapore CISOs the auditable governance artifacts they need to put agents into regulated production.
  • Sovereign and confidential controls finally line up. Confidential VMs on Intel TDX, Confidential External Key Manager, post-quantum cryptography in KMS, and Workspace sovereign controls close the gap with what financial services and government in Singapore have been asking for.
  • Agent Identity changes the audit story. Every agent now gets a verifiable cryptographic ID with a clear authorization trail. This is a direct answer to MAS guidance on AI accountability and the upcoming Model AI Governance Framework 2.0 expectations.
  • The Agent Marketplace creates an export vector for Singapore startups. Local AI startups can now sell agents directly into Gemini Enterprise customers globally. Combined with the $750M partner innovation fund and the Google for Startups AI Agents Challenge, Singapore's startup ecosystem has a credible distribution channel.
  • Cross-cloud is finally serviceable, not aspirational. Cross-cloud Lakehouse with Iceberg REST Catalog, lakehouse federation, and zero-copy federation match how Singapore enterprises actually run their stack: heterogeneous, multi-vendor, and reluctant to repatriate data.
  • Citi Sky is the reference for Singapore wealth and banking. An always-on, multilingual, agentic wealth advisor built on Google's stack is exactly the pattern DBS, UOB, OCBC, and the private banks will be evaluated against by their boards over the next 12 months.

Want to talk about what this means for your stack?

Digigen is a dual-certified Google Cloud and Microsoft Solutions Partner with offices in Bangkok and Singapore. We help enterprises move from AI pilots to governed production deployments.

Get in touch with Digigen

Digigen | Bangkok & Singapore | digigen.io

Source coverage based on the Google Cloud blog post "260 things we announced at Google Cloud Next '26" published April 25, 2026.
View full post