[SINGAPORE] Anthropic is working to bring its most advanced artificial intelligence models “on soil” in Singapore, as the US AI company deepens its presence in the Republic through a new agreement with the government on AI safety and adoption.

“We are working as quickly as we can to bring the frontier models on soil in Singapore, but no specific dates to share on that,” Anthropic co-founder Benjamin Mann said on Wednesday (Oct 14) at the launch of the company’s Singapore office.

The Claude developer officially opened the office on Wednesday, its fifth in Asia-Pacific after Bengaluru, Seoul, Sydney and Tokyo. The opening had been announced in September.

It also signed a memorandum of understanding with Singapore’s Ministry of Digital Development and Information (MDDI), covering AI safety and evaluation, research into the technology’s economic impact, responsible adoption by businesses and its use in the public sector.

The agreement was signed by MDDI permanent secretary Chng Kai Fong and Mann.

Dale Finlay, Anthropic’s general manager of Asean, said the company was also committed to sharing information about AI-related incidents with governments “as quickly as we responsibly can” when it identifies incidents of concern.

The MOU does not give Singapore priority access to Anthropic’s frontier models ahead of their commercial release, Finlay clarified. It does, however, cover testing and evaluation work.

Anthropic has worked with Singapore institutions on AI safety since 2023, including on a “red-teaming” project that tested its models for potential weaknesses, harmful outputs and other safety risks across English, Mandarin, Malay and Tamil.

It also held a session with Singapore’s AI Safety Institute this week to train government and civil society experts in how to evaluate AI models and assess their risks, Finlay said.

Building out Singapore team

The company is expanding its Singapore team, although it declined to disclose its current headcount or hiring target.

At least 12 Singapore-based positions were advertised on Anthropic’s careers page as at Wednesday, spanning sales, finance, marketing, recruitment and research.

These include Asia-Pacific roles in tax, payroll and accounting, as well as Asean-focused sales positions and a regional research economist.

Finlay said the near-term focus was on building out the company’s go-to-market capabilities, given its existing customer base across the public sector, enterprises and startups.

Chris Ciauri, Anthropic’s managing director of international, said its typical market footprint includes sales, technical pre and post-sales, customer success, partner management and public policy.

But the company has no immediate plans to establish a core research operation in Singapore.

“Right now, we don’t have plans to necessarily bring research here,” he said.

That could change over time as Anthropic’s research teams assess the local talent pool, he added.

Anthropic also does not currently plan to replicate its Singapore office footprint elsewhere in South-east Asia, Ciauri said, with the company using its sales and customer success organisation to extend its reach across the region.

Strong local adoption

Singapore is already one of Anthropic’s most active markets.

Its latest Anthropic Economic Index ranked the Republic second among 121 countries for Claude usage relative to population, with usage nearly six times the level that Singapore’s population alone would predict.

The company counts several major Singapore organisations among its customers.

More than 1,500 GovTech engineers and technical staff have adopted Claude Code, while DBS uses Claude in areas including software development and productivity.

GIC is rolling out Claude across its investment and corporate services teams, while Grab uses the model to power a sales assistant that Anthropic said saves employees about 40,000 hours every quarter.

Sea Ltd also uses Claude in areas including customer service and software development.

Wednesday’s launch comes as Anthropic is preparing for an initial public offering that could value the company at more than US$2 trillion.

Mann said Anthropic viewed safety as an important competitive differentiator, particularly as more capable AI agents are deployed in regulated industries.

Large banks, for example, would be reluctant to deploy such models unless they could trust how the technology handles issues ranging from user data to bias, he said.