Thomas Kurian opened Google Cloud Next ’26 in Las Vegas with a message that is going to ripple through boardrooms from Raffles Place to one-north over the next quarter. The “Agentic Enterprise” is no longer a slide in a roadmap deck. It is shipping, at scale, and Google Cloud has just repositioned its entire stack around it.
If you are a Singapore CIO, CTO, or business leader wondering what this actually means for your operations across Singapore, this post breaks it down. We cover the major announcements, translate the hype into practical relevance for the Singapore market, and share Digigen’s honest take as a Google Cloud Partner working with enterprises across Singapore every day.
Let’s dig in.
Headline
Gemini Enterprise Is Becoming the Operating Layer for Business AI
The big framing from Kurian’s keynote is that Gemini Enterprise is now the front door to AI for every employee and every customer interaction. It is no longer positioned as a chat tool with enterprise controls. Google is calling it the connective tissue between your data, your people, your apps, and your agents.
The proof is in the numbers Google shared. Paid monthly active users grew 40% quarter-over-quarter in Q1. Nearly 75% of Google Cloud customers are now using AI products in production. And 330 customers each processed more than one trillion tokens over the past 12 months.
For Singapore, this matters because the adoption curve is already moving from experimentation to operationalization. A year ago, many enterprises were still focused on proofs of concept. Today, the question is how to deploy agents into production without creating governance, security, or compliance risk. Google’s answer is a single governed platform rather than a stitched-together stack.
1. Gemini Enterprise Agent Platform: The Real Story
This is the announcement that will affect the most Singapore businesses in the next 12 months. The new Gemini Enterprise Agent Platform is built around four pillars: build, scale, govern, and optimize.
Build
- ADK with graph-based multi-agent logic
- Agent Studio for low-code creation
- Agent Registry to catalog agents and tools
- Marketplace integrations across major SaaS tools
- MCP support across Google Cloud services
Scale
- Agent-to-agent orchestration
- Sub-second cold starts
- Long-running agents
- Secure sandboxes
- Memory Bank and Memory Profiles
Govern
- Agent Identity with cryptographic IDs
- Agent Gateway with Model Armor
- Anomaly Detection
- Security dashboard via Security Command Center
Optimize
- Observability with OTel telemetry
- Simulation using synthetic interactions
- Evaluation on live production traffic
Digigen’s take for Singapore
This is a major shift. Many Singapore organizations are caught between lightweight agent experiments with limited governance and heavyweight enterprise implementations that move too slowly.
The Agent Platform creates a more credible middle path: a governed foundation with real security controls, clearer observability, and a better chance of passing internal risk and compliance review.
For regulated sectors in Singapore such as financial services, healthcare, and the public sector, the governance layer is one of the most important parts of this announcement.
2. AI Hypercomputer: The 8th Generation TPU Arrives
Google announced the eighth generation of its TPUs, split into two specialized chips.
TPU 8t
Built for training. Scales to 9,600 TPUs and 2 PB of shared memory in a single superpod, with major gains in processing power and efficiency.
TPU 8i
Built for inference. A new topology and more on-chip SRAM deliver materially better price-performance for production AI workloads.
- NVIDIA Vera Rubin NVL72 joins Blackwell and Hopper on Google Cloud
- N4A Axion instances offer 100% better price-performance than comparable x86
- C4N and M4N boost network bandwidth per vCPU
- Managed Lustre now reaches 10 TB per second
- Rapid Storage jumps from 6 TB/sec to 15 TB/sec
- Virgo Network introduces a new AI-optimized fabric
Digigen’s take for Singapore
Most Singapore enterprises are not training frontier models themselves. The real business story is inference cost, latency, and scalability.
If you are running customer support agents, document workflows, or internal copilots at scale, improvements in inference efficiency can materially improve the operating model.
For enterprises balancing performance, governance, and regional deployment requirements, this makes Google Cloud a stronger production option.
3. The Agentic Data Cloud: Where Data Becomes Agent-Ready
Google is restructuring its data platform around a new assumption: agents, not just humans, are becoming primary consumers of enterprise data.
- Cross-Cloud Lakehouse on Apache Iceberg with zero-copy integration across major platforms
- Lightning Engine for Apache Spark with stronger speed and price-performance
- Data Agent Kit for Gemini-powered authoring in notebooks, IDEs, and terminals
- Knowledge Catalog to create a dynamic context graph of the business
- Deep Research Agent connecting structured and unstructured BigQuery data
Digigen’s take for Singapore
This is one of the biggest practical blockers we see. Enterprise data is still fragmented across CRMs, ERPs, cloud warehouses, SaaS tools, and legacy internal systems.
The Cross-Cloud Lakehouse and Knowledge Catalog make it more realistic to deploy useful agents without first running a massive migration programme.
For Singapore-headquartered regional businesses operating across Southeast Asia, that flexibility is especially valuable when data residency and governance requirements vary by market.
4. Agentic Defense: Security for an Era of AI Attacks
Google is pushing hard on AI-powered security, combining Google Threat Intelligence, Security Operations, and Wiz’s cloud and AI security platform.
- Dark Web Intelligence for exposure profiling
- Threat Hunting Agent
- Detection Engineering Agent
- Wiz AI Application Protection Platform
- Wiz Red, Blue, and Green Agents
- Google Cloud Fraud Defense
- Triage and Investigation Agent with faster alert analysis
Digigen’s take for Singapore
Security teams in Singapore are under pressure to do more with speed, precision, and stronger governance. AI-assisted detection and triage are moving from nice-to-have to operational necessity.
For finance, healthcare, and digital services businesses, the ability to compress investigation time and increase analyst productivity is highly relevant.
One practical note: Wiz remains a separate commercial consideration, so licensing should be included in any serious rollout planning.
5. Customer Experience and Google Workspace
Two areas received notable upgrades.
Gemini for Customer Experience
- Shopping and food ordering agents
- Support agents and agent assist
- Omnichannel Gateway
- Low-latency multilingual voice
Workspace Intelligence
- AI Inbox and AI Overviews in Gmail
- Ask Gemini in Google Chat
- More agentic Docs, Sheets, and Slides
- Google Drive Projects
- Workspace agent for cross-app tasks
- Faster Microsoft 365 migration
Digigen’s take for Singapore
For Singapore businesses, the customer experience opportunity is especially strong in service-heavy industries where multilingual support, operational efficiency, and customer responsiveness matter.
On Workspace, this is also highly relevant for businesses evaluating productivity stack direction. The comparison with Microsoft 365 is no longer static. AI is changing the decision criteria.
6. Regulated Industries and Openness
For banks, insurers, healthcare providers, and public sector organizations in Singapore, this is important: Google is supporting Gemini 3 models in both air-gapped and connected environments. Combined with sovereign cloud options, that reduces one of the major blockers for highly regulated deployments.
Google also doubled down on openness. Claude models are available on Vertex AI alongside Gemini, giving customers more flexibility to mix models without locking themselves into a single vendor path.
Recommended next steps
What Singapore Enterprises Should Do in the Next 90 Days
- Audit current agent experiments. If you already have prototypes on automation tools or raw LLM APIs, plan the migration path toward a more governed production model.
- Start your Knowledge Catalog. Structure the context around your data before you scale agents.
- Evaluate inference cost now. Benchmark your current workloads against newer infrastructure and model options.
- Revisit the Workspace decision. Whether you stay or migrate, the comparison with Microsoft 365 has changed.
- Treat agentic security as core infrastructure. Detection, response, and fraud defense should be piloted early, not later.
Final Thoughts
Google Cloud Next ’26 is the clearest signal yet that the enterprise AI conversation has moved past “should we use this” to “how do we govern this at scale.” For Singapore, the timing is significant. Enterprises are now at the point where experimentation needs to convert into real operating models.
At Digigen, we help Singapore businesses design, deploy, and govern Google Cloud, AI, and productivity solutions. If you want a pragmatic view of what matters for your business, we are happy to talk through it.
The Agentic Enterprise is real. The question is whether your organization will lead or follow.
Talk to Digigen
Digigen is a Google Cloud Partner and Microsoft Solutions Partner serving Singapore and Thailand. We help enterprises design, deploy, and govern AI, cloud, and productivity solutions.
Contact us at hello@digigen.io or visit digigen.io.
For AI agent consulting work, see The Agentiv.
