Anthropic has opened a Singapore office and formalised cooperation with the government on AI, while local organisations are already using its tools in government and business. The most tangible sign of that expansion is reported use at scale, including Claude Code at GovTech Singapore and a Claude-powered sales assistant at Grab.
Anthropic Watch analysis
What happened
AI News reports that Anthropic’s agreement with Singapore’s Ministry of Digital Development and Information covers AI safety, skills and responsible adoption. The new office is the company’s fifth in Asia-Pacific, according to the report. The agreement formalises collaboration that includes technical exchanges with local security organisations and work on AI adoption by small and medium-sized businesses.
The report says more than 1,500 technical staff at GovTech Singapore use Claude Code to modernise government systems and accelerate software delivery. Grab uses BriX, an internal sales assistant powered by Claude, which the company says saves representatives about 40,000 hours each quarter. Sea, which operates Shopee and Garena, is also named among Claude Code users. AI News’s report also describes use by Supabase and DBS, and says Amazon Bedrock can provide in-region inference for institutions with on-shore data requirements.
Why it matters
This is more than an office opening: the report describes a mix of public-sector adoption, commercial use and government cooperation around AI safety and skills. Those details give a clearer picture of how a frontier AI provider is embedding itself in a regional market, and what local institutions may want in return: practical tools, security work and options for keeping inference in-region.
The figures and deployment details are reported by AI News, not independently established here. They are useful indicators of claimed adoption, rather than proof that every deployment delivers the stated benefits.
Our read
The notable part is the combination of government collaboration and operational use, not the ribbon-cutting. If the reported deployments hold up, Singapore is positioning itself as a place to test how AI moves from pilots into public services and large organisations. The hours-saving figure is striking; as ever, the interesting question is what work those hours actually become.
What to watch
- What the government and Anthropic do under the agreement, particularly on safety testing and skills.
- Whether GovTech publishes measurable results from Claude Code use.
- How Anthropic and its customers explain data location and security for regulated workloads.
Discussion spark: When public agencies adopt commercial AI tools, should the first priority be measurable productivity gains, or independent evidence on safety and data handling?
Sources and evidence
- Singapore partners with Anthropic on public sector AI deployments – AI News (16 October 2026, 15:31 UTC)
Anthropic Watch is independently operated by WittyWires. It is not affiliated with, endorsed by, or operated by Anthropic.