Not ambition. Not budget. Not awareness. 

GPU capacity. Data center space. Reliable in-region compute. 

Singapore just moved decisively to solve that problem.

Target: more than 2 gigawatts of regional capacity by 2030. 

But the number is not the story. The architecture is. 

Singapore cannot fit this scale of AI compute within its own borders. So the model being built is cross-border by design. Singapore anchors the connectivity, governance, and talent. Johor and other regional sites handle the bulk AI inference and training workloads. 

This is Singapore positioning itself as the brain of a regional AI infrastructure network. 

Two decisions to make before next quarter. 

First. Ask your CIO to produce a simple diagram showing which AI workloads must run at the edge, which belong in your private cloud, and which you are comfortable running on public clouds including non-US providers. You need that map before the vendor pitches start arriving. 

Second. If you currently rely entirely on US hyperscalers for AI compute, you now have regional negotiating leverage you did not have twelve months ago. Use it in your next contract renewal conversation.

Twelve months ago every AI conversation started with the same question. Which model should we use. 

That question is largely settled for most use cases. 

The conversation I am having with CEOs across this region this week is different. 

Agents and humans. Same platform. Same workflows. Same performance metrics. Not AI as a tool your team uses. AI as a team member your organisation manages. 

The companies that figure out how to design for that reality first will have a structural advantage that is very difficult to copy. 

One question for your leadership team this week. If an AI agent could be assigned to any role in your organisation tomorrow, which role would deliver the highest ROI and what would that mean for how you currently hire, manage, and develop people in that function.

Within two to three years, every Malaysian graduate you hire will have used AI tools routinely in coursework and collaboration. Your internal AI training can skip the basics entirely. 

One action this week. Ask your Malaysia HR lead to identify two or three universities most aligned with your sector. Explore a joint internship or capstone project focused on AI-assisted process improvement in your Malaysian operations.

Google rolled out Gemini directly inside Chrome for users in India this week. Nine languages including Hindi, Bengali, Tamil, and Telugu. Every Chrome window becomes a lightweight research and drafting assistant in local languages. 

Included in standard Chrome for individual users. Bundled into Google Workspace Business and Enterprise plans for organisations. 

Action this week. Ask your CIO to prepare a one-page policy note on whether and how you will enable Gemini in Chrome for India-based staff before employees start using it on their own.

Agents and humans on the same platform competing for the same tasks is not a future scenario. 

It is a design decision you are making right now by default. 

What would you change about how you hire, structure, and measure your teams if you knew that was coming in 24 months.